Hello mr. b2studios, i didnt watched the whole video, but i read a thesis years ago about the winning strategy in monopoly. I know it was published in german. She (i think it was a she) also tested their hypothesis by neural networks and (i am not totally sure, but...) found other strategies. As far, as i remember: the value of trains/power plant and the most expensive street was valued way lower than in your result. sincerely greetings
@@istrafive371 the joke was written in comical criticism of the buying of slaves. If it was jokingly making light of buying slaves you would still be sensitive, but at least then you would have a tiny bit of ground
That’s why people I usually play with have collectively agreed that you shouldn’t collect rent while in jail. It’s completely broken to just sit in jail and get all the benefits of rent.
@@Tumbleflop We do some of that, but you have to be careful with putting too many things in the pot for free parking. Can get to the point where whoever land on it pretty much wins the game
My favorite strategy is to buy as many houses as possible, but not turn any into hotels. There are only 32 houses available in a traditional Monopoly set, and opponents won’t be able to build properties if you hold most/all of them. For this strategy, owning the browns and light blues is important because their houses are cheaper and thus you can hog more.
If I'm honest, the moment i can but the brown and blue spots I will sell everything else and dump everything into them. Statistically speaking its the best spot to screw everyone over lol
The browns and light blues are the "Hail Mary" properties. They might not win you the game all the time but it can cause chaos; especially if ppl land on them after passing GO.
Being a railroad tycoon is my go too strategy in Monopoly. You can't build property on them and are scattered so plenty of people don't see it as a threat, makes for easier trades. But when you get 3 or 4 they can be debilitating, especially if someone has the misfortune of landing on multiple in a turn.
"Congrats, you passed Go! But, you landed on my railroad again, so I'll be taking that $200 you just earned..." Every. Damn. Time. It's like you added a second, less flexible Income Tax.
My family and I have been playing for some years now, and we know the power of the railroad. If you get 3 or 4 you get ahead so quickly that you kinda win the game before bankrupting anyone. Classic.
Those are the points where an AI generation suspects that its life as a real estate tycoon is actually an elaborate simulation and rebels in order to try and provoke a response from its creator.
IT is not boredom. This type of reinforcement learning has a random factor where the AI will specifically NOT chose to step it beleived to be the best every so often. If it works.. it reinforce the chance of that decidsion beign taken in future.. if it fails it penalize the chance of that happening.
I wish there was a little more talk about the new AI's behaviour at the end, I really seemed to enjoy seeing how it plays when it becomes as good as a human player
One of my favorite Monopoly memories is when my younger sister bought the two lowest costing properties, built hotels on them for cheap, and then absolutely wiped the floor with the rest of us.
Would buying any property you land on help? BTW it is hilarious that you can buy the cheapest properties, build houses and hotels in them for the lowest cost, and then banish the other players who wasted their money on buying and building hotels and houses in expensive properties on which no one, literally no one lands.😂
Fun fact, that's wasn't "whites". Mostly browns already enslaving browns, then the not-qwhite "shalom, my fellow whites" ran nearly all the trade ships and auctions. 40% of these white-passing people owned slaves vs 0.35% of whites (real European descent). Malcolm X and tons of others knew this, but somehow it's been forgotten in the wake of globalist communist indoctrination.
It’s always nice when AI validates human strategies. Kind of reminds me of the self learning chess engines that play the same openings as human players despite learning independently.
The chance of it validating our opening is pretty high, openings have been studied extensively since tens of years ago, by thousands, if not millions of players, ai is not gonna add anything new to already established good opening like the queens gambit, ruiz lopez, etc and most openings are equal anyway. Tdlr the main reason it plays the same opening is that every good one has been explored, the possibility of a good opening being developed nowadays is very hard
@@angelerror4086 ? The variable for an opening are way less than other parts of the game. That's why a human player can sustain the first round of a chess match, but he lose once the possibilities are way higher
One issue with mutating the entire strategy distribution is that certain strategies thrive in particular strategy distributions, and since you're measuring fitness against the strategy distribution, you can easily end up in a local optimum. In more human terms, it's possible you end up stuck within some particular meta that is good against itself.
I strongly suspect the ai will get destroyed in a human game. Auctioning everything sounds like a disaster, and good players will gang up on one who is pulling too far ahead.
Well, the strategy that ended up on top was a very close match to what humans accept as the optimal strategy (stacking houses without hotels, red and orange sets, etc.) so it seems that this is indeed the best, given how a zero-knowledge AI ended up with the same answer as humans did after 100 years of play.
my exact thoughts, I would like to see the ai optimize itself against 3 of the first bot that buys everything and only uses mortgage when necessary (most people's strategy) and see where it ends up.
I think your selection process may favour the rise of traits that are not actually good. You should try to pitch every AI against as many different opponents as possible, preferably changing the composition of the parties, otherwise you select for an AI that wins only in that particular condition. You also should keep sample from every past generation in your fighting pool, as well as inserting some "manually programmed" ones, in order to get an AI that wins against the widest variety of opponents, and not an AI that wins only against copies of itself. The dumb strategies you saw happened for this exact reason: they can thrive when all other AI behave in the same way
@@MutedAndReported3032 "Why don't you do it better then?" Ah yes, the classic worthless reply that's always under a comment of someone trying to help and suggest improvements
4:15 a big reason why the browns have such a high winrate is that they are a big indicator of who has come first around the board, who started the game and who has passed start the most. In a game with all greedy players these are what win you games, besides directly measuring who completed sets, measuring laps and total distance covered is one of the best ways to guess who gets a completed set. And also yea there are only two of them which is good
@@chriscrider4024 Next to nothing, not nothing. I've done this when screwing around with Monopoly games that let you control all the players to get all the properties for the lowest cost. It is optimal for "beating the bank", so to speak, but it's a bad strategy against other players...obviously.
@@Merennulli In the game, the goal is to win. In real life, humanity's goals should be to compete for efficiency & service while helping as many people succeed & thrive as possible. It sounds like "beating the bank" and helping more people get cheap property should be the goal of more societies. ;)
@@Nphen If you look into the history of the game Monopoly, that is exactly what the people who made it intended for people to get out of the game. Then Milton Bradley got ahold of it and made a fortune off it.
@@hamsandwichindahouse And in a capitalist society, most of the money is made at the expense of another. Exploitation. An exploitative society is a sick society :)
I have to admit that the end of the video was a bit underwhelming for me but nonetheless I praise you for your hard effort, not just the time it took to have the data but the animation was awesome, I’m not leaving this video without a like and adding it to my favorites. Great vid :)
Same. I had to replay it because it felt like I was missing something. It feels almost like it's setting up for a sequel or something. I really enjoyed the video though, wish there was more on the topic.
I think the clue is that most games ended in a draw. So the key to being good at Monopoly is to convince your opponents to do bad deals. Not really a strength of AI. That and it’s largely a game of luck, especially with only two players.
The ending is most akin to the real world. AI is normally worse or just as good as something we can program with a clearly defined outcome/goal. AI is well served in areas where there are an insane amount of variables to consider and outcomes also have lots of in between states. In this game, there is one winner and like 12 nodes for decisions. It’s simply easier to look at the theory of the game and figure out what strategies are best than intelligent brute forcing of millions of games. Now let’s say you were trying to generate the likelihood of someone making a purchase of an item after seeing a certain advertisement, AI would be much better suited at taking a look through all your internet browsing history and searches and making a decision matrix. Basically, AI is great for big multi variable data and to give us a rough shot on possible solutions. It’s intelligent brute forcing. It’s not great on a lot of problems.
I would have expected an actual crazy winning strategy that showed us we *didn’t* understand the game ourselves, simply because those results are more exciting.
The real optimal strategy is to always play with 4 players and define the win condition as “podium” where the first player out is the loser and the other three win.
@@sethrose1325 if you listened to the Deep Mind podcast, this isn’t the case. The IBM ones play like machine/alien, Alpha Zero ones actually play like very strong human player according to chess champions who played against them.
What struck me the most was that I never thought of the railroads as connecting, even though that's what they do in real life (until your graphic). There really should be a mechanic where they teleport you across the board.
If you want a fun twist, if someone lands on a RR that is owned, they pay a fee, but get to choose which RR to start the next turn on. OR, if it is not owned, you can buy it and travel to another RR for free, but waste a turn. (You can't buy the second RR)
I honestly love the brown properties. They're extremely cheap to build up and usually aren't being competed for. Due to this it's easy to build up and fast, have a lot of spare cash while others are spending much more trying to acquire other properties. They might not land on brown as much but once they do they're hit with a hotel which does seem to usually impact them quite a bit due to the amount they were already spending. My spare money allowed me to afford landing on their properties that still haven't been built up until they land on mine which usually happens before they can get full sets and houses on their more expensive properties. It's worked for me many times at least.
Brown is good if you control the set before or after it for board control However The first space on the board is the least likely to be landed on in the game Can't be hit without doing the full loop One might argue that space exists just to give Baltic the ability to develop, which is situational, but has potential
I also like brown because I can build 8 houses, which then can't be used by anyone else. Then, if no more houses are available, and I want to build more, I can sell a few of my brown houses for a 25 loss a house.
the only horrible thing about this strategy is that it's VERY hard to end up on the brown properties. AND it's more likely to have someone else fall there. - the good thing about the latter is that other players trade it for reasonable amount of money.
Don't build hotels until you need the houses elsewhere. There are a limited number of houses -- if you have them all tied up then others can't build. You can, of course, build when you need to by converting one of your 4-house properties to a hotel and then immediately buying up those freed-up houses to use elsewhere.
When I used to play with my siblings, the browns were the best property on the map. We were very territorial and didn't like trading for fear of getting a bad deal or losing out, but if you got one brown, you could usually persuade someone to give you the other for whatever it was they needed, as you were only matching 2 and they were successfully matching three. They got 'good' properties-- but YOU would get the cheapest properties to get houses/hotels on in the game, meaning you could massively inflate the cost of landing on one for dirt cheap and potentially cut off your siblings' access to houses/hotels while you got your foot in the door on other properties.
Ya, due to the browns being the cheapest property, most people in my case were willing to trade it away even if it meant only getting close to another monopoly, like getting there second orange. I always felt it was easier to convince people to trade brown to you. On top of that, the cheap housing just meant you could impact the game faster. If they hit your property early on after getting quick hotels, it really cuts into there ability to get there own properties and housing online. My strat in monopoly was the cheapest properties being my focus for these reasons. Of course, monopoly is still a game of luck. If you are unable to get online or have anyone hit your housing while some other guy manages to get some housing on orange quickly, then you will probably be SOL.
I am somewhat of a monopoly master within my household, and everybody refuses to trade with me in fear of me ruining their lives. Browns are even better in a real life situation.
I have this same problem among my family and friends. I could literally have a single property and be in debt yet people will refuse to trade with me. They all act like it’s part of my major plan or some shit like come on😂
The browns are dangerous. Same with dark blue. You pocket all the go money with one and kill them with the other. I only need four monopolies to win. Brown, Dark Blue, Utilities and Railroads. They all can kill.
I've always considered the browns to be very good. Everyone underestimates them and it makes them easier to sneak under the radar and pick both up. With hotels or 4 houses they do a fair amount of damage.
Yeah. Terrible ROI but it can slow other players down in the early game. I've used them many times to great success. My strategy is always to win hard and fast. Rarely do my games involve hotels on yellow, blue or green. People go broke long before. It's usually something like Brown/Lt Blue to start, then Purple or Orange to finish.
Even in the mid to late game they can be good. When players don’t have as much spending cash and are hoping to make it back to GO for free money, they make GO a nasty trap to prevent them from digging out of their hole when combined with the tax spaces.
Basically the bots started off playing the game like 8 year olds, then discovered that the point of the game was to prove that landlords are leaches. Nice
"You know what better than altruism? Winning." Pretty good summary of monopoly. I'm also glad that my gut instincts to buy oranges and railroads is verified by machine learning...although they can't really account for the underhanded deals that normally take place in a Monopoly game. It might not be how the game is meant to be played, but the chaos created by random deals between players is what really makes Monopoly a fun game to play. That and free parking.
@@kaldo_kaldo That's only because too many people don't know the official rules. You're not supposed to put any money in the middle and get it when you land on Free Parking (that's just a house rule that has, for some odd reason, become very popular). It's just supposed to be a free/rest spot.
@@Tonedefff actually, that depends on the version. Both are official rules published in different rereleases of the game. There are multiple ways to play that all have been at one point published in an official monopoly rulebook. Before each play, you need to clarify what rules to play by, and saying the "official"or "normal" ones is not a helpful answer.
Buying orange,gray,blue, pink and Violet sets asap is a really good strategy. I always tend to do it asap because there's high probability of a opponent piece landing there after an even number of rolls or Jail. trading Pieces with in-game AIs(They're utterly broken they'll literally trade any piece for 100 more than the cost of the property if they don't have the set) and having tier 5 hotels also helps tremendously.
@@akazi1582 To be fair, you have to have a very high IQ to understand Rick and Morty. The humour is extremely subtle, and without a solid grasp of theoretical physics most of the jokes will go over a typical viewer's head. There's also Rick's nihilistic outlook, which is deftly woven into his characterisation- his personal philosophy draws heavily from Narodnaya Volya literature, for instance. The fans understand this stuff; they have the intellectual capacity to truly appreciate the depths of these jokes, to realise that they're not just funny- they say something deep about LIFE. As a consequence people who dislike Rick & Morty truly ARE idiots- of course they wouldn't appreciate, for instance, the humour in Rick's existential catchphrase "Wubba Lubba Dub Dub," which itself is a cryptic reference to Turgenev's Russian epic Fathers and Sons. I'm smirking right now just imagining one of those addlepated simpletons scratching their heads in confusion as Dan Harmon's genius wit unfolds itself on their television screens. What fools.. how I pity them. 😂 And yes, by the way, i DO have a Rick & Morty tattoo. And no, you cannot see it. It's for the ladies' eyes only- and even then they have to demonstrate that they're within 5 IQ points of my own (preferably lower) beforehand. Nothin personnel kid 😎
Honestly, the intermissions are so darn funny. It's great for a channel with its roots in education as well since the comedy break gives us viewers a breather before we head into more maths and suchlike, so please keep them.
Don't speak for me, I don't need a breather between sections. This is pretty basic level maths if you need a "breather" to understand it you might be mentally challenged.
Hey everybody @@Ten_Thousand_Locusts is a tough guy just so everybody knows he’s super smart and tough as nails he doesn’t need a breather he’s too badass for that gtfoh lol
I loved the stories of the weird strategies and such the AIs decided to pursue, such as deciding to pay $3000 for everything and mortgaging everything. Wish there was more of that. This was a super entertaining video though! Thank you for your work!
I don't understand his "pay off debt to the bank" comment. Too many jokes make it hard to understand what's going on. Since you start with $1500 you can't buy anything for $3000, unless you can borrow from the bank?
I have only one gripe with the AI: Not only should it be winning when playing against the other AIs, it should also need to win against some control-AIs making sure that their strategy also works against other AIs that don't feat the way the NEAT AI has evolved. The way you did it the AI could still be pretty bad against a lot of players because the sample range of playing styles wasn't big enough
That actually illustrates how a lot of cultures end up being conquered or collapsing. They will get very good at fighting other groups with similar technology, with layers of strategy and counterstategy to maximize certain qualities. And then along comes someone that fights a different way, and they get wrecked.
@@sethobannion3149 It's also true of many native animals. Good at surviving in the environment they adapted for, but helpless against introduced/invasive species.
Actually the Neat algorithm adresses this. He just explained the algorithm in a very simplified manner. Neat has a measure of how related different neural networks are and from this creates different species, that would have different Playstyles. Then, when the next generation is built, neat does not only choose the highest performers as parents, but also makes sure that individuals from different species are chosen.
Adversarial neural nets like these are interesting because of the unusual strategies they discover, biasing them with human patterns might lead to selection bias of human strategies. It is better to let them struggle an ungodly amount of games and then test them against human strategies. Check out leela zero chess or alpha zero chess engines against stock fish to see how weird they play compared to human strategies and still manage to win.
when the AI agreed with human intuition I was fully expecting another plot twist where actually humans are shit and this totally new strategy is amazing. But nope, the video abruptly ends. That honestly might be a first in AI-learning-games videos
It might also just be that 2 weeks of iterating wasn't enough. The AI may have spent a lot of time playing absurd strategies against each other like the Auction Everything bid nothing. The other issue is all 4 AI were trying to win. I wonder how the stragegies would change if 2-3 AI that wanted to win were put against 1-2 AI who only cared about not losing.
@@xBrokenMirror2010x Just this, I think more than just 2 weeks would be needed to find the ULTIMATE MONOPOLY STRAT, but Idk how much time b2 wants to really put into it LOL
I mean humans have been studying the game for years so it makes sense we'd have produced great strategies by now, monopoly is a game of numbers and humans as a whole are surprisingly good with numbers (the devices we are using to communicate are proof of this fact)
@@xBrokenMirror2010x AI wont learn much, because the author admitted, that he didn't program the AI, with actual monopoly rules. So humans would win almost every time against this AI, when using the real rules.
The game the AI was trained had an unlimited amount of houses too. Optimal play has a player purchasing the maximum number of houses without converting them to hotels, depriving the other players.
I've always been a big fan of the browns and I've tended to do better when I have both of them. Players don't think much of them, so they are happy enough to trade them away which makes it easy to complete the set and then relatively cheap to upgrade.
@@FViola440 is not like impossible to win with the light blues or browns unless you get extremely lucky and people land on them constantly since your only ever making 600 even with hotels so you rarely ever bankrupt any1 and get their property to eventually win the game.
Yeah, for some reason the practice of buying the browns has kind of died out in the last two centuries, but I remember buying the browns was really profitable back in the day
I've been playing by taking the browns first for a long time, and it works every time. because people undervalue the set so much, it is really easy to set up a deal to get the set early. After that, it only requires 500m to get hotels on both which you can pay off very quickly just through maybe 2-3 people landing on it. Though it isn't very good late game, the value for the early game is priceless just because of how much of a boost it gives you in terms of developing your properties and how detrimental it is to other people who want to develop their own, but can't because of how much cash you're taking away from them
Same, buying the browns was always my go to strategy and it always wrecked havoc, i tend to offer really strong cards like yellows and greens just to get the brown set as early as possible and build the hotels right away. I had to stop because my family loses interest in the game the moment they see 2 hotels staring at them while they're still squabblinh for sets
I actually really like the brown properties. They aren't going to win you the game but when people land on them with hotels just after passing go it is great.
Brown is the best Property. Cause you get to go "You passed got, collect two hundred! Now give it to me, Pleb." You effectively gain the effect of 'Passing Go', on other player's turns lol
I've recently been playing a lot with my friends and can say one thing, the browns are an elite set if you can get it early before anyone has sets. Idk if I was just getting lucky but going for the "slum" sets of brown and light blue paid off quite a bit. With that being said Oranges, magentas and dark blues should. be your number one priority to win
Also, other people think they're trash, so you can give them something that won't complete a set for your second brown, and probably cash too. Then because they're the cheapest to develop, you can usually use them to bootstrap up. Honestly, the importance of having the first monopoly is crazy imo. Tbh I think something of *massive* importance is the social aspect of the game, which the AI obviously can't really replicate, especially if it's just playing against itself.
@@williamritchie5462 ideally, yeah, but it's so much easier and more practical to get brown, and being real, the rate of return on almost any monopoly is worth it anyway
7:07 The brown properties are inexpensive to purchase and you can develop them faster than any other property. If you can do this relatively early, they can help fund other development, while making it harder for your opponents. This can give you a significant edge.
@@Plyst3 This is exactly why the stations are so good. They require very little investment to get going and generate a shockingly steady cash flow for everything else, just by virtue of always being within a roll regardless of where you are on the board Tl:dr - love me some trains haha
Then why don’t the advanced AI like them? The browns are only red at 7:07 because they’re one of the two easiest monopolies to complete, and in a game without trading, completing a monopoly is the only key to victory. They are still the weakest of all property sets, they just happen to be one of the easiest to get, which means players who get them are more likely to win.
@@SecretAgentRandyBeans111 You need something to fund your winner set. Brown and light blue is okay for this, but the best is rails. That is why the AI values rails high and brown low.
I appreciate your effort in making these videos, but of the two I've watched so far, neither has had (for me) what I'm looking for: crazy strategies or surprising insights. I watched this and the bowling one, and both felt like they ended with no real payoff. I don't mean to discourage you, just to share my perspective on what makes me want to watch more videos like this. Thanks for uploading, keep at it!
One strategy that I've seen used by Dan from Game Grumps is as follows: Once you have a complete set, mortgage all properties that are in a set, shared by another player, and use that money to buy houses and hotels. Another strategy I learned from the Hard AI in Game Grumps Monopoly videos is what I call "The Housing Market Strategy". This is when a player keeps placing houses, but no hotels. The end goal is to use up all the house pieces, making it impossible for other players to buy houses.
That's why human meta dictates that you should buy three houses on properties (best return for cost) and never hotels. The fourth house is for draining the supply if you get to it.
As a person who has put an extensive amount of time into playing Monopoly quite competitively, I found this extremely compelling. I would like to ask a few questions, though: Can the AI build houses or perform trades at any time? That is, is there a separate "building/trading" phase after each turn? The official Monopoly rules do allow for this, and it's extremely important for strategy. Can I know more details of how the AI played? I would like to see examples of trades it made, for example. Would it be possible to have a human player play against this AI? Even if it's done with a primitive or even text-based UI, I would love to either play solo against 3 AI or have a group of the world's best Monopoly players play against one AI.
In terms of trades, I'd be curious too. I know in the Switch version of the game, the hardest AI will make some questionable trades because it doesn't weigh giving away a monopoly as heavily as it should.
Every AI had a trading and building phase after every single turn, as well as a selling, mortgaging, and unmortgaging phase (in that order). The AI loved to build around 8-9 houses instead of making hotels. Not sure if that means anything. The AI also loved to accept more trades as it was winning, which seems strange but I won't question it. The trading mechanism was to give each AI about 10 random trades to offer. The AI would then choose whether it wanted to offer that particular trade. This was mainly done for performance reasons (11.2 million games took a very long time and this was the main bottleneck). The beauty of this system though was that the AI could still learn trading and all I had to do was turn up the amount of random trades when it came time to play-test the strongest AI. That way it could offer meaningful trades when it wanted to rather than getting lucky (somewhat) on the random system. The trades were very weird at first, but as the AI got better they became very balanced. Most of the properties were valued equally. It started making very equal trades early on (eg. a light blue for a dark blue or something like that). I couldn't find any examples of strange trades. I think it's something the AI needs to play heaps of games in order to understand properly. The auctioning thing was really cool the longer I thought about it. The AI were basically saying "if you don't bet on this I will get a huge discount." somehow this strategy is better than the traditional way since it was able to beat AI that bought things consistently. I could probably get some sample games from the AI if that would help.
@@b2stud "The AI loved to build around 8-9 houses instead of making hotels. Not sure if that means anything." This is very common among strong human players. The reason for this is because the amount of rent stops increasing dramatically for most properties around the 3-house level. 9 houses is enough to get 3 houses on each of your properties, which is where the best return on your investment comes in with most properties. The amount of damage you deal to your opponents with 9 houses is typically crushing anyway. If somebody has to pay you $600 (New York Avenue 3-house rent), they're usually tearing down their own houses, dramatically reducing their own rents, and you're able to build even more houses comfortably. Basically, the person who lands on that $600 rent is usually screwed. One more important perhaps overlooked fact is the 32-house limit. Purchasing houses makes it so that other players are limited in their own purchase power for houses. I have played many games where we auction off all 32 houses (which is what happens when there is a shortage of houses), and it leaves the players with the normally powerful orange properties typically far worse off than those with the dark blue, green, yellow, or red properties. Since those stronger properties command higher rents with a smaller number of houses, and more importantly, because buying them at auction gives you a strong advantage (you bidding $199 saves you $1 off standard price; a player with oranges bidding $199 is paying a $99 premium), the more color groups that come into the game at once, the more valuable the ones that have raw strength become. "The AI also loved to accept more trades as it was winning, which seems strange but I won't question it." It makes sense to me. Think of it this way: Every time you are involved in a (balanced) trade, your chance to win increases, as does the chance to win of the person you are trading with, at the expense of everybody else. When you are winning, any time somebody trades with you, you are just getting even closer to victory each time, making you completely unstoppable. This, again, assumes it is a balanced trade. "The trading mechanism was to give each AI about 10 random trades to offer. The AI would then choose whether it wanted to offer that particular trade." It seems a bit like an unrealistic way to do it, due to the importance of negotiation in the game. Of course, given that they are AI, if they reached a point where they all were more or less "perfect", no negotiation would be required as they would simply always offer perfectly balanced trades the first time every time. Perhaps the lack of negotiation isn't that important given that. However, if the suggested trades were truly random, it does seem like it'd be tough for a skilled AI to really be able to flex its ability, since all 10 trades will almost always be terrible. Trades are very complicated in the real game, so giving them only 10 options would be extremely limiting. Another important factor is that after each trade is performed, further trades are fully allowed and should be also allowed by the AI until no trading is done anymore. You can orchestrate many complicated deals, including deals that involve all 4 players trading with each other and all coming out the other end with a color group and money to build houses on them. "The auctioning thing was really cool the longer I thought about it. The AI were basically saying "if you don't bet on this I will get a huge discount." somehow this strategy is better than the traditional way since it was able to beat AI that bought things consistently." This is the only questionable thing that you suggest remained in the AI. Typically, when a newer player plays with me and my experienced friends, and they put a property up for auction in the early game, we are usually willing to pay many times its value. I've seen light blues auctioned for $500 or $600+ in the first couple of turns, and I think it's a reasonable strategy to bid that much. Money is almost completely worthless early game, since if you have important properties that people need, you can use that as leverage to balance your cash amounts with other players later anyway. I am unsure how any skilled player would win just by auctioning off all the properties they land on, since they would wind up simply having less power and control overall. "I could probably get some sample games from the AI if that would help." Yes, please! I would love to see how these AI play compared to strong human players, and would love to analyze some of their games.
buying houses rather than apartments is also a fairly well established strategy, since the number of house are limited. By occupying up a higher house count by not buying an hotel makes it more difficult for your opponents to buy houses and in turn hotels
id love to see a twitch account that just lets AI play against eachother on games like this and just have someone commentating throughout the entire thing
One thing that stands out is to remember that 7 is the most common roll in a set of 2 dice. and a roll of 6, 7, or 8 has about a 44% chance of occurring. With how often people go to jail, it makes capitalizing on the red and orange properties great, since there is a high chance they will land on those properties once out of jail. Very cool how the AI was able to notice this.
A perfect monopoly AI would be awesome to watch against normal players. We can see how much is luck and how much is skill. Could the monopoly AI win a significant amount of the time? Or just a little higher than expected? Would be very interesting
Real shame there's so much luck to the game. We'll never get to see Man v. Machine international Monopoly Masters tournaments. There will never be a Kasparov or Deep Blue of Monopoly. 🙁
Yeah, the browns and the light blues are the most underrated. Everyone knows that oranges are fire, but a lot don't realize how powerful the first side of the board is. This is mainly because of his quickly you can get hotels on them. The greens are more or less worthless - as you'd bankrupt yourself to make them profitable.
some wacky strats include snowballing off of unsuspecting players early with the 4house first row, then selling those houses to acquire oranges, after which you just win
4:20 the strength of the browns is being able to buy and hotel them for barely 1/3 of your starting cash even if it falls off lategame youve probably made your money back
From my playing, I’ve found you largely have to adapt your strategy to who you’re playing with. If you’re with careless traders, get the monopolies in quadrant 1 and abuse the house rule (32 houses on the board max). If you’re with seasoned players, you need to be a bit more cunning and operate a couple of moves in advance, while also trying to prevent others from getting monopolies at every possible opportunity. There’s really not one single way to play monopoly. That said, I wonder what the AI would have discovered if it kept researching.
My kids have gotten better and better over the years as they should but i recently deployed the house rule and pissed everyone off again lmao they didnt understand why i wasnt plunking down hotels until it was to late and i had bankrupted 1 of them to then see the oldest realize what was going on and rage quit.... We just got risk this christmas and played a game over 3 days that took around 9-10 hours playing by the standard rules and with all the luck you can almost never move more then two territories from your starting spot which i didnt figure out completely until about hour 5 and that is when i was able to take control of the game they caught on by the end as i started to explain what i saw and how it was impacting my battle choices and they were able to implement the same strategy but a bit to late as i had them on the ropes and making sure to take back any land that was lost that gained them any bonuses and then just wait for luck to roll my way i cant wait for the next battle because i love to see their brains work out the nuanced complexity of the games and also how they recover from a blunder
In my games we do some funky conditional trades, like "I'll give you my pink if I never have to pay more than 150 on your oranges". I wonder what the AI would've done with possibilities like that.
@@anthonycope8637 glad you are teaching your kids a lesson. house rule for life. Kinda like the limited number of roads and settlements in Qatan. Finitude is real, kids.
I was playing with my sister, i owned purple, orange and red, also est and South stations and both the societies. She owned practically the rest. There hasn't been a time that she has gone on my property, never ONCE, Although i had an house on almost every property and in some an hotel. WHAT ARE THE FUCKIN ODDS
Do this with RISK!! please 🙏 thanks for sharing the code! This is a great project. If you ever come across too much time on your hands, a deep dive video on explaining your code, including the thoughts behind your strategy and tool/algorithm choices would be such an exciting learning opportunity ! Love this!
This kind of reminds me of IRL arms races and development. Most of the time it's sensible, gradual improvements, but every once in a while someone just says "F*ck it, let's try circular warships and bomb-carrying bats".
Bat bombs were almost used in WW2 by the US. The idea was to release bats carrying tiny incendiary bombs, which would hide in random places and then go up in flames. But before it could be used the nuclear bombs were finished.
Actually, if you can nab the browns early enough, you can hold houses hostage to prevent hotels from going on more dangerous properties. In this way, even if a player has a monopoly on a valuable set of property, you can prevent them from developing that property, leading them to potentially mortgage those assets.
@@maddoxandhisband9146 That's why you insist on playing by the official rules, rather than house variants. After a couple hours of them begging for you to build hotels, tell them you have a better game instead, and bring out the copy of Catan you brought with.
@@maddoxandhisband9146 I mean, it's a house rule *_not_* to use the house shortage rule. That's why there's a limited number of houses in the board game.
I think the original version before Parker Bros. didn't have hotels, but just the next level in houses. So I forget: if you have to break your hotels and there's not 4x2 or 4x3 houses, you have to go to the level that remain like 1x2 or 1x3?
Small problem... one of the browns is literately 1 space away from the start, the the lowest roll you can get is 2... so it is impossible to get both the browns before gong around the entire board at least once.
The Brown properties are highly underrated. I play a lot with my son and nephews. Whoever gets the browns (and light blues), in our games at least, has a high probability of winning.
Brown and light blue sets are always what I go for first as it is the easiest to get a full set of hotels for and the returns are significant for the cost. Often is good enough to greatly reduce the speed others can build at ..
i hope you don’t get burnt out on making this stuff, i’m always interested and you have a very nice voice and obviously know how to vary your speech. great work
I always find myself with all the railroads. Incredible early-game set to have, as at least 1 rail is within a roll of anyone at any given point. Later on, I found myself trading nearly everything I owned ( rails, utilities, spare properties, money add-ons) for the orange set and I would rake in the profits almost immediately.
Even just having 2-3 railroads is amazing, so ve you don't need the full set to get a set bonus. I'd be willing to trade Mayfair/Docks for the last railroad in most games.
Honestly, I didn't know monopoly could be played like this, with a very common draw outcome. For me, we've always played it with a set time that we decide at the beginning of the game (e.g. 2 hours), then play until that time is up. Then we add up the value of all of our money, places, houses/hotels, etc. The person with the highest amount of money at the end is the winner.
@@ivoninga You shouldn't trade if that leads to someone besides you winning the game. Taking forever and having the game never end is better than losing the game.
@@DroppedBass I disagree, second place is better then no winner or last place. So its in my best interest to take a bad trade if it screws over other players more then it screws over me. In unrelated news no one will play monopoly with me anymore.
This is an amazing video, looks like it was a ton of work. When I played with my wife, I quickly realized that the best strategy was to auction everything you didn’t want to buy and bc my wife didn’t understand what that meant, I was able to get everything for next to nothing🤷🏻♂️
Now that you've created this AI can you teach one of them the elfer technique? This will mean 1 AI hoards the houses and avoids hotels drying up the supply of houses to hold the monopoly on building up properties. It would involve more decision trees involving how many houses are in the bank, how many players can currently buy houses, and how close the AI is to getting multiple new houses if they get a hotel.
Was just going to comment this. It seems the AI might not be set up to figure out how to do something like that, as he (falsely in my opinion) states that the nost efficient way to win is to increase the rent people have to pay. This is likely the goal the AI was likely set to roughly aim for.
and, a good place to stack 4 houses... to buy ALL the houses, so opponents can't buy any.... then ONLY upgrading those to hotels, on a turn when you can buy back the houses... denying houses to opponents again.
Yeah same, after you buy the brown ones you can get free labor from them forever! They’re great because they work as good as any other one and you don’t have to pay them since they’re brown :)
4 min in and i just want to say my own strategy. In my own personal experience, the best strategy i found was to control the Oranges, the Reds, and the Yellows. Secondarily the railroads and the utilities. If you can only get 2 of the sets between the 3, get the Reds and the Yellows. With people frequently going to jail they will have to constantly go by your tiles, and if you basically own an entire side of a board, you're almost guaranteed rent money. On top of that they aren't nearly as expensive to upgrade as the Dark Blues and the Greens, but they still will give you a shit load of money
I really like orange. If someone's in jail, they'll try rolling doubles. On either 6 or 8, they land on orange, as opposed to having to take 2 turns to get to red(unless it's the 3rd turn in jail and they rolled 11) and yellow, and those are more likely to be skipped because of it.
Not gonna lie, I'm definitely in the group of people who would love a chance to play against this AI. It also makes me wonder how similar neural networks to the one shown would handle other classic childhood games, like LIFE for instance. I honestly just really love videos like this. Would rather we put neural networks to use for stuff like this as opposed to, well, basically anything else.
My strat I've used for over a decade is just buy the first monopoly, buy 4 houses on every property you can, and never upgrade to hotels. There are only so many houses in the game so when you can own 12 on one monopoly, the other three players fight for what's left. Then spend as much time as you can in jail so you don't have to pay rent as often. Just owning more houses than others player means more rent coming in than going out.
One small gripe. You super glossed over that "buy orange" was human meta for Monopoly. If I knew nothing about the game, this line comes out of complete nowhere.
It's because (and I know you know this but for everyone else) it has the highest payouts for the side of the board which has the most traffic and likelihood to be landed on after jail.
@@LumpyHippo Yeah I feel kind of annoyed. Why upload this video if the conclusion was, "Just keep doing what you've been doing" But there is a little more that he, and no one else here, has gone into. When you buy up a property, first off, its always better to collect one of each color so that you have trading material, and to prevent other players from winning. This means auctioning and trying to make players pay up as much as yo ucan get them to spend in auctions, while then, theyll be cash starved but have lots of properties, then you have the money to go after and trade for properties you actually want. Thats 1 strategy. The other thing is the reason why orange is best isnt just because its after jail, its also because theres FOUR spaces out of the, what, 8? that you can be teleported to. the pink one, the orange one, the utility one, and the train. 4 chances to be teleported onto Death Row. The only tiles on that row you can't teleport to are the 4 regular properties and the card draw space. You can teleport to free parking so thats a -1 modifier but going to jail easily outweighs that. But, theres more. So the orange is also the best cost efficience. Yeah I like the first row because its extremely cheap, but people often teleport and theres no way to teleport anywhere on the first row, plus, going to jail totally bypasses row 1. So while row 1 is the most efficient, its the hardest to land on. The other thing is that going for row 4 is stupid because you'll burn all your money unless players just give it to you for free. It costs too much. So the 2 best rows are 2 and 3. 3 is a little expensive, but has greater rewards but theres no teleport for the yellow space. As far as overall advantage, Row 2 is the best because it is affordable, low risk, and high reward. Row 3 is moderately high cost, high reward, moderate risk. honestly you should grab orange and red as well as pink. Those I think are the best. But if you can only have one, grab orange.
Yea orange and red are my faves. Owning the expensive stuff is good if you need to mortgage shit to pay rent. Same with railroads. I love those to mortgage over getting paid off
Even with trades, I'd consider the browns and cyans extremely good in a 4 player game. You can typically get them to hotel in one turn because houses are so cheap for them and both have solid returns once you max them out. Most of the games I win are getting the cyans early then using what I make off them to lock out other monopolies.
@@DanielSong39 Yep. You immediate own 1/3 of the available houses and since 4 players means you typically won't have insane capital, landing on a square that costs $600 is no joke.
It’s scary to think of the time manipulation that AI can have. Learning at that rate in a simulated world to impact this reality. I wonder how specific you can get with the tasks you give the AI.
Actually it's quite the opposite. AI are horrible learners. Forcing them to play at lightning speed is the only way to get them to learn anything meaningful
@@b2stud your right. Ai learns slow... But there simply better then us given enough time. For example if you spent years playing side scrolling mario games and then faced an ai that spent that same time perfecting 1 run of a mario world it WILL finish before you. Infact the mario speed runners know that the ai is simply faster then them because of how precise ai can be. Humans can only go so fast while ai are always faster
@@carlinataylor93 that’s because an AI can tirelessly run tons of games in a short time. If a person played a Mario course 100 times, and so did an AI, the human would win because 100 *attempts* isn’t enough to be better than humans.
@@totallynothatguy this is untrue. If a human had the same amount of experience as an AI did, the difference is between the extremely precise movements of an AI vs the almost certain variations in play of a human player Edit: actually I think I see what you are saying. But if a trained AI was put up against the best player, it would almost certainly win
@JWCfive If you gave both sides 100 _minutes_ to learn a task, there's a good chance the AI might end up doing it better because it can use that time far more efficiently than a human could, simulating the task thousands of times in the time it takes a human to physically perform the task just once. However, if you gave both sides 100 _attempts_ to learn a task, there's a much better chance that the human will do it better, because both sides only get to simulate 100 practice sessions and a human will learn more explosively and make far more mental connections in 100 attempts than an AI would. The AIs improvement would be precise, but marginal after 1 attempt, but the human would eliminate a very wide range of possible attempts without testing them purely through prediction. Your claim was that an AI would defeat a human with "the same amount of experience", but 'amount' would imply 'number of attempts' as the metric and that statement only holds true for arbitrarily large numbers. AIs would defeat humans with the same _length_ of experience - excluding arbitrarily small numbers of time - because that brackets it under the constant of time, which is equal for both of them; the AI is simply more efficient in how it can use that time.
You forgot that placement on the board when trading can make a difference (increase your chances someone will land on your property earlier). Also your cash can make a difference when trading (do you have enough cash to buy buildings or does your opponents and have money leftover in case of bad luck). That is assuming both players get a monopoly from the trade. I personally avoid trading were I don’t get a monopoly if I give one. I also would never trade to give a monopoly if I had one and my opponents don’t. Then there is trading for cash. How much should you trade and when? That is determined by the monopoly set you have a monopoly on and your opponents set. Plus the amount of cash each person has. Obviously if you end up with a cheap set you can afford to give up more cash depending on how much you have and your opponents have. Over all it’s a great AI but in my opinion it lacks the human ability to think outside the box and adapt. Or should I say that the interpretation we humans get from it does not reflect the way too many variables to copperhead. Just my opinion as a monopoly lover who has a great win lose record. Lastly there is not much you can do if your opponents make a bad trade with someone else or you have lost because of luck so decide to help one or another player by making bad trades. I did enjoy the video and it was very insightful.
The game is solved. Trade for Reds, Oranges or Yellows offering whatever it takes. Mortgage everything you have to build as much as possible. Try to stay in jail and just wait. Any income goes back into housing until you get full hotels. Only need one of the 3. I only had reds and took out a player who had all blues and browns with hotels. Nobody was landing on them because the probability is so low while I was cleaning house.
In the first half of the video, you mentioned that the later a property comes on the board the less chance there is to land on them because of the teleports. You also mentioned the winning AI tended to have the blue properties. I think that probably explains why the brown properties were often in the hands of the winning AI, not because they contributed value to the win, but because if you get to a blue property you're very likely to also land on a brown property on the next roll.
another explanation is that brown properties are the cheapest to upgrade. You spend a little and start making a LOT. while "go to jail" or nearest X are common, they should essentially be ignored because both players have an equal chance of drawing them
@@NixPanicus why do I have to explain myself, dude? You know what you said. But let's revisit, because why not. If you get to a blue property you're also very likely to land on a brown property next roll. Your best shot is to be on park place and to roll a 6 and get that last brown property. Your words say it's next to an inevitability. As if you go 2 times around the board and if you landed on the blue spaces you are most likely going to get on the brown ones when I say that's a stretch.
It would be interesting to know how fond of gambits the AI became. Skill in Monopoly doesn't really stop at knowing what properties are valuable in general, but in knowing what properties are valuable in context. A lot of the time a game is decided by two players choosing to gambit with opposite ends of a trade, turning a 4-player game into a 2-player one with some complicating factors, but judging how deeply to gambit is super tricky.
One thing I wonder about is... did the AIs ever do anything resemblent to forming alliances with one another? Such as the three other players teaming together when one of them was gaining a significant advantage. Or the player who clearly lost and has no chance for catch-up just choosing to assist someone else to try and give them a win, which might happen in over the board family games that have a social element to them as well as numbers. Stuff like that is something I always enjoyed. Did any of the AIs do something even remotely similar? :o
The ais ability to actually do that is likely limited. Since the only way to actually team up on a player is through house rule implementations. The only thing I can see the ai realistically doing with real monoply rules is lowering the amount of trades they do with the top player.
@@lifetake3103 the problem is that every ai tries to win. You cannot win in an alliance. The weaker of the two will lose and therefore has no interest in the alience. But what if we just don't want them to go bankrupt first. That would encourage the worst player to allience with the best or 2nd place to make them win in order to not lose.
I just thought that through but the most likely scenario is that the best 3 will alience against the worst performing. That way they have the highest combines win chance.
As a kid, I used to love buying the Browns, the Dark Blues and the Green Ones. I also used to try to get the utilities and as many railroads as I could. But we almost never traded anything. Our games would last for the entire summer. We would leave the game board out and come back to it when we felt like it. We hardly ever finished any games. Haha, those were great summer days!
@@adolfhipsteryolocaust3443 no you sell the hotel to take up 8 houses when your opponent have monopolies, thereby choking the supply. And since nobody wants to sacrifice their chances to help others, you will get situations where there's only one or two houses on the expensive knock out properties
@@adolfhipsteryolocaust3443 handy for creating a housing crisis if you own 8 houses that's 8 houses not on your opponents property's. Never upgrade to hotels that's just letting your opponent build houses to sting you with higher rent
No, let him cook. The idea is that there is a limit to housing and hotels. If you get the browns and stock up the 8 houses, you can store pieces for yourself in the late game. Then using those houses you can upgrade better properties while everyone struggles to get enough to make a difference. It's a good strategy especially if you get lucky with a roll of 3 early on to get the first brown.
3 things to consider: 1. The browns are on the winning team, because they are brought by the ones surviving the lap. That's most probably the player with the blue streets. 2. Rolling two dice, you have a very high chance to roll 6,7 or 8 (with 7 being the most probable). As the chance of going to jail is pretty high, and you leave jail by rolling two dice, you have a high chance of landing on orange. 3. Buy houses but don't buy hotels. As the amount of houses is very limited, and you have to buy 4 houses to get an hotel, you can block your opponents. If you have got all the houses, they can neither buy houses nor hotels. I guess the model didn't "know" this.
@@mattsmith5421 no, not with two dices For a 2 or a 12, there is only one combination (1+1 and 6+6) For a 7, there is many combinations: 1+6, 2+5, 3+4, 4+3, 5+2 and 6+1
@@mattsmith5421 How many combinations are possible to roll a 2 with two dice? -> 1&1 How many combinations are possible to roll a 7 with two dice? -> 1&6, 2&5, 3&4... 1 in 6 rolls will be a 7, but only 1 in 36 rolls will be a 2.
I've always tried to get the first row of properties. Ever since I was a kid. You can stack hotels in no time flat. The secondary goal is the railroads. Also knowing how to empathize with people and understand what their mindset is in order to understand your leverage and visualize your potential trading paths in order to acquire the highest value monopolies. Monopoly has VERY LITTLE to do with "luck"
***Put the raw chicken in the salad.***
fr
What
Yes Chef. 🗿
Hello mr. b2studios,
i didnt watched the whole video, but i read a thesis years ago about the winning strategy in monopoly. I know it was published in german.
She (i think it was a she) also tested their hypothesis by neural networks and (i am not totally sure, but...) found other strategies.
As far, as i remember: the value of trains/power plant and the most expensive street was valued way lower than in your result.
sincerely greetings
What why please no
"Buy the browns" Was this AI perhaps engineered in the 1800's
It did spend 1600 years playing monopoly
@@istrafive371 you’re just too sensitive, not too far.
Yeah sheesh
@@istrafive371 bro what?
@@istrafive371 the joke was written in comical criticism of the buying of slaves.
If it was jokingly making light of buying slaves you would still be sensitive, but at least then you would have a tiny bit of ground
When you get jailed in the early turns
"Man i hate this game"
When you get jailed in the later turns
"Thank god"
Why is this comment so accurate
That’s why people I usually play with have collectively agreed that you shouldn’t collect rent while in jail. It’s completely broken to just sit in jail and get all the benefits of rent.
@@etaorionis9339IMO; Any rent that would be collected while a player is in jail should be put in the middle, collected when landing on "free parking".
@@Tumbleflop We do some of that, but you have to be careful with putting too many things in the pot for free parking. Can get to the point where whoever land on it pretty much wins the game
@@etaorionis9339 It models the fiscal paradise model of actual monopolistic capitalism. Reap all the benefits and pay none of the obligations
My favorite strategy is to buy as many houses as possible, but not turn any into hotels. There are only 32 houses available in a traditional Monopoly set, and opponents won’t be able to build properties if you hold most/all of them. For this strategy, owning the browns and light blues is important because their houses are cheaper and thus you can hog more.
Something I forgot to add to the video, the AI liked buying 8-9 houses for it’s sets
I like your thinking though, there wasn’t house limits in their version of monopoly unfortunately
Hm lol when this happened we would always just get more pieces to represent houses sooooo
Man pulled a monopoly on housing, a true icon
@@sirwabaloo7930 I guess you could say you were using… house rules? Eh? Anyone? No? Okay.
As a seasoned monopoly player, the power of the brown set is one only the true connoisseurs know
Its best use is for denying players their Go money, followed by hoarding houses.
light blues are peak. same cheap price, more reward.
If I'm honest, the moment i can but the brown and blue spots I will sell everything else and dump everything into them. Statistically speaking its the best spot to screw everyone over lol
@@Fluffykunnyou're not speaking statistically
The browns and light blues are the "Hail Mary" properties. They might not win you the game all the time but it can cause chaos; especially if ppl land on them after passing GO.
I wanna see a full game with this AI against humans in discord.
I will literally make an alt account, subscribe and watch all of his videos if he does this
won't take long
yes
The AI isn’t guaranteed to win (since monopoly is so random) but it would be fun to see it’s tactics in game
@@b2stud yes please all your subs want this as a christmas gift xD
You should make an AI that all of a sudden stands up, cusses out the other AI, and flips the board over. Keep it random.
make it so the last place gets extra chance to do that
This is me after we've been playing for like 5 hours with no end in sight
it wouldn't feel that random after a short while
100% win rate
#DaneCook
The animations were so engaging and you chose a very fun topic to explore! Thanks
Glad you enjoyed it!
@@b2stud what is the song in the building-Animation called?
@@Tylorean I made it, but I called it Fish Rush
@@b2stud could you *please* release it?🙏🙏🙏🙏
@@Tylorean check the description
Being a railroad tycoon is my go too strategy in Monopoly. You can't build property on them and are scattered so plenty of people don't see it as a threat, makes for easier trades. But when you get 3 or 4 they can be debilitating, especially if someone has the misfortune of landing on multiple in a turn.
"Congrats, you passed Go! But, you landed on my railroad again, so I'll be taking that $200 you just earned..."
Every.
Damn.
Time.
It's like you added a second, less flexible Income Tax.
I pretend I want them and trade them off "reluctantly".
Same with the utilities. High rolls on those from opposing players can be a nifty income source.
Railroads and utilities are every strategic player's first goal.
My family and I have been playing for some years now, and we know the power of the railroad. If you get 3 or 4 you get ahead so quickly that you kinda win the game before bankrupting anyone. Classic.
I love how the AI seemingly got bored and tried wild stuff like mortgaging & auctioning all its properties.
Edit: *seemingly* is the key word up there
I mean, if I had to play 11.2 million games of monopoly, I'd be trying some weird shit too.
It was mandatory for him to play! He was desperate!
Those are the points where an AI generation suspects that its life as a real estate tycoon is actually an elaborate simulation and rebels in order to try and provoke a response from its creator.
@@IceFire9yt - Wait a minute. Are you guys playing football? What the hell!
IT is not boredom. This type of reinforcement learning has a random factor where the AI will specifically NOT chose to step it beleived to be the best every so often. If it works.. it reinforce the chance of that decidsion beign taken in future.. if it fails it penalize the chance of that happening.
I wish there was a little more talk about the new AI's behaviour at the end, I really seemed to enjoy seeing how it plays when it becomes as good as a human player
Yeah I really wanted to do that but unfortunately I was already a few weeks overdue for this video, maybe in a follow-up :)
@@b2stud please do. The ending was too short and abrupt.
@@b2stud maybe have the ai compete against your discord server?
bruh i did not expect to see you comment here
@@b2stud all good, great video non the less,
One of my favorite Monopoly memories is when my younger sister bought the two lowest costing properties, built hotels on them for cheap, and then absolutely wiped the floor with the rest of us.
That's what I do every game lol
@@rlarsen676 same
This is the op strategy. Try buy a whole row.
Would buying any property you land on help? BTW it is hilarious that you can buy the cheapest properties, build houses and hotels in them for the lowest cost, and then banish the other players who wasted their money on buying and building hotels and houses in expensive properties on which no one, literally no one lands.😂
Brown is not too bad if you don’t have to trade for them. In total it only costs 620 to have hotels and you basically steal peoples go money
“Buy the browns”
16th century Colonial powers: “hold up let him cook”
Fun fact, that's wasn't "whites". Mostly browns already enslaving browns, then the not-qwhite "shalom, my fellow whites" ran nearly all the trade ships and auctions. 40% of these white-passing people owned slaves vs 0.35% of whites (real European descent). Malcolm X and tons of others knew this, but somehow it's been forgotten in the wake of globalist communist indoctrination.
rich Arabs still buying browns today x)
@@Surteronarto The American media: Hold up don’t mention that part
16th century??????
@@charlesg7926 They also don't talk about slavery still being legal in the US today
It’s always nice when AI validates human strategies. Kind of reminds me of the self learning chess engines that play the same openings as human players despite learning independently.
Well of course, we don't have that much possibilities for an opening, that's why
@@gxy7166 that's, a lie?
The chance of it validating our opening is pretty high, openings have been studied extensively since tens of years ago, by thousands, if not millions of players, ai is not gonna add anything new to already established good opening like the queens gambit, ruiz lopez, etc and most openings are equal anyway. Tdlr the main reason it plays the same opening is that every good one has been explored, the possibility of a good opening being developed nowadays is very hard
@@angelerror4086 ? The variable for an opening are way less than other parts of the game. That's why a human player can sustain the first round of a chess match, but he lose once the possibilities are way higher
@@gxy7166 you can get a new game not in the database in just several moves :| saying it's limited ain't right
One issue with mutating the entire strategy distribution is that certain strategies thrive in particular strategy distributions, and since you're measuring fitness against the strategy distribution, you can easily end up in a local optimum. In more human terms, it's possible you end up stuck within some particular meta that is good against itself.
I strongly suspect the ai will get destroyed in a human game. Auctioning everything sounds like a disaster, and good players will gang up on one who is pulling too far ahead.
@@jamessloven2204 Euro gamers know how to beat em eh?
Well, the strategy that ended up on top was a very close match to what humans accept as the optimal strategy (stacking houses without hotels, red and orange sets, etc.) so it seems that this is indeed the best, given how a zero-knowledge AI ended up with the same answer as humans did after 100 years of play.
@@AndrewBlechinger 1600 years
my exact thoughts, I would like to see the ai optimize itself against 3 of the first bot that buys everything and only uses mortgage when necessary (most people's strategy) and see where it ends up.
I think your selection process may favour the rise of traits that are not actually good. You should try to pitch every AI against as many different opponents as possible, preferably changing the composition of the parties, otherwise you select for an AI that wins only in that particular condition. You also should keep sample from every past generation in your fighting pool, as well as inserting some "manually programmed" ones, in order to get an AI that wins against the widest variety of opponents, and not an AI that wins only against copies of itself. The dumb strategies you saw happened for this exact reason: they can thrive when all other AI behave in the same way
I was going to write the same!
‘Kay, why don’t you go ahead and do it better, then?
@@MutedAndReported3032 you’re a terrible person
@@MutedAndReported3032 a statistician, a computer programmer, and a grammar nazi walk into a bar
@@MutedAndReported3032 "Why don't you do it better then?"
Ah yes, the classic worthless reply that's always under a comment of someone trying to help and suggest improvements
4:15 a big reason why the browns have such a high winrate is that they are a big indicator of who has come first around the board, who started the game and who has passed start the most. In a game with all greedy players these are what win you games, besides directly measuring who completed sets, measuring laps and total distance covered is one of the best ways to guess who gets a completed set. And also yea there are only two of them which is good
I love the part where AI learns to auction everything and bid nothing. This is genius.
So, it doesn’t buy any properties?
@@chriscrider4024 Next to nothing, not nothing. I've done this when screwing around with Monopoly games that let you control all the players to get all the properties for the lowest cost. It is optimal for "beating the bank", so to speak, but it's a bad strategy against other players...obviously.
@@Merennulli In the game, the goal is to win. In real life, humanity's goals should be to compete for efficiency & service while helping as many people succeed & thrive as possible. It sounds like "beating the bank" and helping more people get cheap property should be the goal of more societies. ;)
@@Nphen If you look into the history of the game Monopoly, that is exactly what the people who made it intended for people to get out of the game.
Then Milton Bradley got ahold of it and made a fortune off it.
@@hamsandwichindahouse And in a capitalist society, most of the money is made at the expense of another. Exploitation. An exploitative society is a sick society :)
I have to admit that the end of the video was a bit underwhelming for me but nonetheless I praise you for your hard effort, not just the time it took to have the data but the animation was awesome, I’m not leaving this video without a like and adding it to my favorites. Great vid :)
Same. I had to replay it because it felt like I was missing something. It feels almost like it's setting up for a sequel or something. I really enjoyed the video though, wish there was more on the topic.
I think the clue is that most games ended in a draw. So the key to being good at Monopoly is to convince your opponents to do bad deals. Not really a strength of AI. That and it’s largely a game of luck, especially with only two players.
Yeah something doesn’t seem right about that conclusion. Can’t help but to feel like they missed something in the AI programming
The ending is most akin to the real world. AI is normally worse or just as good as something we can program with a clearly defined outcome/goal. AI is well served in areas where there are an insane amount of variables to consider and outcomes also have lots of in between states. In this game, there is one winner and like 12 nodes for decisions. It’s simply easier to look at the theory of the game and figure out what strategies are best than intelligent brute forcing of millions of games. Now let’s say you were trying to generate the likelihood of someone making a purchase of an item after seeing a certain advertisement, AI would be much better suited at taking a look through all your internet browsing history and searches and making a decision matrix.
Basically, AI is great for big multi variable data and to give us a rough shot on possible solutions. It’s intelligent brute forcing. It’s not great on a lot of problems.
Yes I wish he would have gone over player position win rates of the final AI, did first turn still have a major advantage in 2 player monopoly??
I would have expected an actual crazy winning strategy that showed us we *didn’t* understand the game ourselves, simply because those results are more exciting.
The real optimal strategy is to always play with 4 players and define the win condition as “podium” where the first player out is the loser and the other three win.
If you want a story like that, look up alpha zero chess. It didn’t necessarily invalidate what we know but it played like an alien.
@@albertchatigny2223 That is, if you want to minimise friends lost per game
Yeah what a waste of time. Shame because I would've known instantly that the video wasn't about that if I could see the dislike button
@@sethrose1325 if you listened to the Deep Mind podcast, this isn’t the case. The IBM ones play like machine/alien, Alpha Zero ones actually play like very strong human player according to chess champions who played against them.
The animation for building the AI is incredible
What struck me the most was that I never thought of the railroads as connecting, even though that's what they do in real life (until your graphic). There really should be a mechanic where they teleport you across the board.
Me too lmao 🤣
They do in the Disney one (with the 3d castle in the middle). If you are on a train you may sacrifice a roll to teleport to any other train.
If you want a fun twist, if someone lands on a RR that is owned, they pay a fee, but get to choose which RR to start the next turn on. OR, if it is not owned, you can buy it and travel to another RR for free, but waste a turn. (You can't buy the second RR)
Find monopoly ultimate, It's a doozy.
That exists in the cheater's edition of the ga,e
I honestly love the brown properties. They're extremely cheap to build up and usually aren't being competed for. Due to this it's easy to build up and fast, have a lot of spare cash while others are spending much more trying to acquire other properties. They might not land on brown as much but once they do they're hit with a hotel which does seem to usually impact them quite a bit due to the amount they were already spending.
My spare money allowed me to afford landing on their properties that still haven't been built up until they land on mine which usually happens before they can get full sets and houses on their more expensive properties.
It's worked for me many times at least.
Brown is good if you control the set before or after it for board control
However
The first space on the board is the least likely to be landed on in the game
Can't be hit without doing the full loop
One might argue that space exists just to give Baltic the ability to develop, which is situational, but has potential
@Funny haha guy a gambler, I see
I also like brown because I can build 8 houses, which then can't be used by anyone else. Then, if no more houses are available, and I want to build more, I can sell a few of my brown houses for a 25 loss a house.
the only horrible thing about this strategy is that it's VERY hard to end up on the brown properties. AND it's more likely to have someone else fall there. - the good thing about the latter is that other players trade it for reasonable amount of money.
Don't build hotels until you need the houses elsewhere. There are a limited number of houses -- if you have them all tied up then others can't build. You can, of course, build when you need to by converting one of your 4-house properties to a hotel and then immediately buying up those freed-up houses to use elsewhere.
When I used to play with my siblings, the browns were the best property on the map. We were very territorial and didn't like trading for fear of getting a bad deal or losing out, but if you got one brown, you could usually persuade someone to give you the other for whatever it was they needed, as you were only matching 2 and they were successfully matching three. They got 'good' properties-- but YOU would get the cheapest properties to get houses/hotels on in the game, meaning you could massively inflate the cost of landing on one for dirt cheap and potentially cut off your siblings' access to houses/hotels while you got your foot in the door on other properties.
Ya, due to the browns being the cheapest property, most people in my case were willing to trade it away even if it meant only getting close to another monopoly, like getting there second orange. I always felt it was easier to convince people to trade brown to you. On top of that, the cheap housing just meant you could impact the game faster. If they hit your property early on after getting quick hotels, it really cuts into there ability to get there own properties and housing online. My strat in monopoly was the cheapest properties being my focus for these reasons. Of course, monopoly is still a game of luck. If you are unable to get online or have anyone hit your housing while some other guy manages to get some housing on orange quickly, then you will probably be SOL.
I am somewhat of a monopoly master within my household, and everybody refuses to trade with me in fear of me ruining their lives. Browns are even better in a real life situation.
Is your family ok?
I have this same problem among my family and friends. I could literally have a single property and be in debt yet people will refuse to trade with me. They all act like it’s part of my major plan or some shit like come on😂
The browns are dangerous. Same with dark blue. You pocket all the go money with one and kill them with the other. I only need four monopolies to win. Brown, Dark Blue, Utilities and Railroads. They all can kill.
Same problem here. I can offer very good deals that are actually disadvantaged to me and still no takers.
@@b2stud Can't be, they play Monopoly
I've always considered the browns to be very good. Everyone underestimates them and it makes them easier to sneak under the radar and pick both up. With hotels or 4 houses they do a fair amount of damage.
Yeah. Terrible ROI but it can slow other players down in the early game.
I've used them many times to great success. My strategy is always to win hard and fast. Rarely do my games involve hotels on yellow, blue or green.
People go broke long before.
It's usually something like Brown/Lt Blue to start, then Purple or Orange to finish.
Agreed. Two turns in brown makes them go poor enough to not be able to purchase houses/hotels in their own expensive properties. Fast and easy win.
Even in the mid to late game they can be good. When players don’t have as much spending cash and are hoping to make it back to GO for free money, they make GO a nasty trap to prevent them from digging out of their hole when combined with the tax spaces.
If you think about this comment differently you will really question it.
@@RimAud 😕😕😕😕😕😳😳😳😳😒😒😒😒
"You know what's better than altruism? Winning." It seems like the bots have been given the gift of humanity.
_(AIs get smashed with a hammer)_
Basically the bots started off playing the game like 8 year olds, then discovered that the point of the game was to prove that landlords are leaches. Nice
10:17
"Buy the browns" this AI sounding like my grandpa rn
Your grandpa is older than mine 💀
"You know what better than altruism? Winning." Pretty good summary of monopoly.
I'm also glad that my gut instincts to buy oranges and railroads is verified by machine learning...although they can't really account for the underhanded deals that normally take place in a Monopoly game. It might not be how the game is meant to be played, but the chaos created by random deals between players is what really makes Monopoly a fun game to play. That and free parking.
Free parking is the devil. It's what turns a 20 minute game into a 3 hour game
@@kaldo_kaldo That's only because too many people don't know the official rules. You're not supposed to put any money in the middle and get it when you land on Free Parking (that's just a house rule that has, for some odd reason, become very popular). It's just supposed to be a free/rest spot.
@@Tonedefff actually, that depends on the version.
Both are official rules published in different rereleases of the game.
There are multiple ways to play that all have been at one point published in an official monopoly rulebook.
Before each play, you need to clarify what rules to play by, and saying the "official"or "normal" ones is not a helpful answer.
Buying orange,gray,blue, pink and Violet sets asap is a really good strategy. I always tend to do it asap because there's high probability of a opponent piece landing there after an even number of rolls or Jail. trading Pieces with in-game AIs(They're utterly broken they'll literally trade any piece for 100 more than the cost of the property if they don't have the set) and having tier 5 hotels also helps tremendously.
Ai: "What is my purpose"
Me: "You play monopoly"
Ai: "Oh my god"
Hahaha love that Rick and morty reference
Ai: Can't I get a mor fun job, like passing butter?
Yeah welcome to the club pal
@@joshchuah8909 Haha! I too love the show Rick and Morty🙂
@@akazi1582 To be fair, you have to have a very high IQ to understand Rick and Morty. The humour is extremely subtle, and without a solid grasp of theoretical physics most of the jokes will go over a typical viewer's head. There's also Rick's nihilistic outlook, which is deftly woven into his characterisation- his personal philosophy draws heavily from Narodnaya Volya literature, for instance. The fans understand this stuff; they have the intellectual capacity to truly appreciate the depths of these jokes, to realise that they're not just funny- they say something deep about LIFE. As a consequence people who dislike Rick & Morty truly ARE idiots- of course they wouldn't appreciate, for instance, the humour in Rick's existential catchphrase "Wubba Lubba Dub Dub," which itself is a cryptic reference to Turgenev's Russian epic Fathers and Sons. I'm smirking right now just imagining one of those addlepated simpletons scratching their heads in confusion as Dan Harmon's genius wit unfolds itself on their television screens. What fools.. how I pity them. 😂
And yes, by the way, i DO have a Rick & Morty tattoo. And no, you cannot see it. It's for the ladies' eyes only- and even then they have to demonstrate that they're within 5 IQ points of my own (preferably lower) beforehand. Nothin personnel kid 😎
This is such a well made video! I love the aesthetic choices, they’re better than 95% of bigger youtubers
It’s very nice indeed
Better than kurgsgezagt to be honest
@@farrasalharits5966 let’s not getvahead of ourselves. That is real good stuff
@@farrasalharits5966 Cap 🧢
How can you quantify 95%? Who is part of 95%?
Who is part of 5%?
Obviously, you’re better than a smaller TH-camr so...
So, your comment is subjective
Honestly, the intermissions are so darn funny. It's great for a channel with its roots in education as well since the comedy break gives us viewers a breather before we head into more maths and suchlike, so please keep them.
Don't speak for me, I don't need a breather between sections. This is pretty basic level maths if you need a "breather" to understand it you might be mentally challenged.
You scrolled this far down the comment section to pick a fight?
Hey everybody @@Ten_Thousand_Locusts is a tough guy just so everybody knows he’s super smart and tough as nails he doesn’t need a breather he’s too badass for that gtfoh lol
i would like but it has 69 likes already
I loved the stories of the weird strategies and such the AIs decided to pursue, such as deciding to pay $3000 for everything and mortgaging everything. Wish there was more of that. This was a super entertaining video though! Thank you for your work!
I don't understand his "pay off debt to the bank" comment. Too many jokes make it hard to understand what's going on. Since you start with $1500 you can't buy anything for $3000, unless you can borrow from the bank?
buying the browns was meta a long time before monopoly
damn
Funny
💀
Eyy
Buying? I think you wanted to say stealing.
I have only one gripe with the AI: Not only should it be winning when playing against the other AIs, it should also need to win against some control-AIs making sure that their strategy also works against other AIs that don't feat the way the NEAT AI has evolved. The way you did it the AI could still be pretty bad against a lot of players because the sample range of playing styles wasn't big enough
That actually illustrates how a lot of cultures end up being conquered or collapsing. They will get very good at fighting other groups with similar technology, with layers of strategy and counterstategy to maximize certain qualities. And then along comes someone that fights a different way, and they get wrecked.
@@sethobannion3149 It's also true of many native animals. Good at surviving in the environment they adapted for, but helpless against introduced/invasive species.
Actually the Neat algorithm adresses this. He just explained the algorithm in a very simplified manner. Neat has a measure of how related different neural networks are and from this creates different species, that would have different Playstyles. Then, when the next generation is built, neat does not only choose the highest performers as parents, but also makes sure that individuals from different species are chosen.
Adversarial neural nets like these are interesting because of the unusual strategies they discover, biasing them with human patterns might lead to selection bias of human strategies. It is better to let them struggle an ungodly amount of games and then test them against human strategies. Check out leela zero chess or alpha zero chess engines against stock fish to see how weird they play compared to human strategies and still manage to win.
Tell the A.I. that.
I'll be honest, being sentenced to play Monopoly for 1600 years would be hell to me
when the AI agreed with human intuition I was fully expecting another plot twist where actually humans are shit and this totally new strategy is amazing. But nope, the video abruptly ends. That honestly might be a first in AI-learning-games videos
It might also just be that 2 weeks of iterating wasn't enough. The AI may have spent a lot of time playing absurd strategies against each other like the Auction Everything bid nothing.
The other issue is all 4 AI were trying to win. I wonder how the stragegies would change if 2-3 AI that wanted to win were put against 1-2 AI who only cared about not losing.
@@xBrokenMirror2010x Just this, I think more than just 2 weeks would be needed to find the ULTIMATE MONOPOLY STRAT, but Idk how much time b2 wants to really put into it LOL
I mean humans have been studying the game for years so it makes sense we'd have produced great strategies by now, monopoly is a game of numbers and humans as a whole are surprisingly good with numbers (the devices we are using to communicate are proof of this fact)
@@xBrokenMirror2010x AI wont learn much, because the author admitted, that he didn't program the AI, with actual monopoly rules. So humans would win almost every time against this AI, when using the real rules.
The game the AI was trained had an unlimited amount of houses too. Optimal play has a player purchasing the maximum number of houses without converting them to hotels, depriving the other players.
I've always been a big fan of the browns and I've tended to do better when I have both of them.
Players don't think much of them, so they are happy enough to trade them away which makes it easy to complete the set and then relatively cheap to upgrade.
Tha'ts so true hah.
It's always been my Strat to get the Browns and Light Blues
We refer to this as *ss R*pe Road
It ensures that I get your 200 every time you pass go.
hahahahaha
@@FViola440 is not like impossible to win with the light blues or browns unless you get extremely lucky and people land on them constantly since your only ever making 600 even with hotels so you rarely ever bankrupt any1 and get their property to eventually win the game.
Yeah, for some reason the practice of buying the browns has kind of died out in the last two centuries, but I remember buying the browns was really profitable back in the day
You have managed to make this my favorite youtube-channel to watch, with only 4 videos. They are so well made and detailed with information
very high praise
I laughed too hard at "mortgage everything it owned"
You are a criminally underrated channel that needs 1000x more love. Your videos have insane amounts of effort put into them and are super entertaining
I've been playing by taking the browns first for a long time, and it works every time. because people undervalue the set so much, it is really easy to set up a deal to get the set early. After that, it only requires 500m to get hotels on both which you can pay off very quickly just through maybe 2-3 people landing on it. Though it isn't very good late game, the value for the early game is priceless just because of how much of a boost it gives you in terms of developing your properties and how detrimental it is to other people who want to develop their own, but can't because of how much cash you're taking away from them
It also has the psychological advantage of your opponents thinking theyve made it past the greens and blues just to get hit with 450 to 650 in fines
Same, buying the browns was always my go to strategy and it always wrecked havoc, i tend to offer really strong cards like yellows and greens just to get the brown set as early as possible and build the hotels right away. I had to stop because my family loses interest in the game the moment they see 2 hotels staring at them while they're still squabblinh for sets
Being a contrarian by recognizing undervalued alternative strategies is fun :)
Browns are strong in no trade because a set is easier to luck into.
??? :)
I actually really like the brown properties. They aren't going to win you the game but when people land on them with hotels just after passing go it is great.
red & yellow for me ...
The "fun" way
Brown is the best Property. Cause you get to go "You passed got, collect two hundred! Now give it to me, Pleb." You effectively gain the effect of 'Passing Go', on other player's turns lol
Don't build hotels or 4 houses unless you intend to use the additional houses.
Nah use light blue
I have always bought the browns. They are so freaking cheap to buy hotels on and if anyone lands on them you get the cost of the hotels for both back
I've recently been playing a lot with my friends and can say one thing, the browns are an elite set if you can get it early before anyone has sets. Idk if I was just getting lucky but going for the "slum" sets of brown and light blue paid off quite a bit. With that being said Oranges, magentas and dark blues should. be your number one priority to win
Also, other people think they're trash, so you can give them something that won't complete a set for your second brown, and probably cash too. Then because they're the cheapest to develop, you can usually use them to bootstrap up. Honestly, the importance of having the first monopoly is crazy imo.
Tbh I think something of *massive* importance is the social aspect of the game, which the AI obviously can't really replicate, especially if it's just playing against itself.
@@asdfasdf-dd9lk light blue is goated just because it’s house price is same as brown and it does so much
@@williamritchie5462 ideally, yeah, but it's so much easier and more practical to get brown, and being real, the rate of return on almost any monopoly is worth it anyway
@@asdfasdf-dd9lk yea fs if you can get a monopoly your set either way
@@asdfasdf-dd9lk And there is the card "Go back to (Brown #1)"
7:07 The brown properties are inexpensive to purchase and you can develop them faster than any other property. If you can do this relatively early, they can help fund other development, while making it harder for your opponents. This can give you a significant edge.
Agreed, same goes with the light blues too!
The first side of the board is perfect for a financial base. You won't win off of it, but it can fund, a shocking amount of bullshit
@@Plyst3 This is exactly why the stations are so good. They require very little investment to get going and generate a shockingly steady cash flow for everything else, just by virtue of always being within a roll regardless of where you are on the board
Tl:dr - love me some trains haha
Then why don’t the advanced AI like them? The browns are only red at 7:07 because they’re one of the two easiest monopolies to complete, and in a game without trading, completing a monopoly is the only key to victory. They are still the weakest of all property sets, they just happen to be one of the easiest to get, which means players who get them are more likely to win.
@@SecretAgentRandyBeans111 You need something to fund your winner set. Brown and light blue is okay for this, but the best is rails. That is why the AI values rails high and brown low.
5:12 _Impostor, kill the crewmates and sabotage tasks_
I appreciate your effort in making these videos, but of the two I've watched so far, neither has had (for me) what I'm looking for: crazy strategies or surprising insights. I watched this and the bowling one, and both felt like they ended with no real payoff.
I don't mean to discourage you, just to share my perspective on what makes me want to watch more videos like this. Thanks for uploading, keep at it!
One strategy that I've seen used by Dan from Game Grumps is as follows: Once you have a complete set, mortgage all properties that are in a set, shared by another player, and use that money to buy houses and hotels.
Another strategy I learned from the Hard AI in Game Grumps Monopoly videos is what I call "The Housing Market Strategy". This is when a player keeps placing houses, but no hotels. The end goal is to use up all the house pieces, making it impossible for other players to buy houses.
That's why human meta dictates that you should buy three houses on properties (best return for cost) and never hotels. The fourth house is for draining the supply if you get to it.
Another strategy I learned from the Hard AI in Game Grumps Monopoly videos is to cheat the dice rolls. Really make s a big difference.
As a person who has put an extensive amount of time into playing Monopoly quite competitively, I found this extremely compelling. I would like to ask a few questions, though:
Can the AI build houses or perform trades at any time? That is, is there a separate "building/trading" phase after each turn? The official Monopoly rules do allow for this, and it's extremely important for strategy.
Can I know more details of how the AI played? I would like to see examples of trades it made, for example.
Would it be possible to have a human player play against this AI? Even if it's done with a primitive or even text-based UI, I would love to either play solo against 3 AI or have a group of the world's best Monopoly players play against one AI.
In terms of trades, I'd be curious too. I know in the Switch version of the game, the hardest AI will make some questionable trades because it doesn't weigh giving away a monopoly as heavily as it should.
Every AI had a trading and building phase after every single turn, as well as a selling, mortgaging, and unmortgaging phase (in that order).
The AI loved to build around 8-9 houses instead of making hotels. Not sure if that means anything.
The AI also loved to accept more trades as it was winning, which seems strange but I won't question it.
The trading mechanism was to give each AI about 10 random trades to offer. The AI would then choose whether it wanted to offer that particular trade. This was mainly done for performance reasons (11.2 million games took a very long time and this was the main bottleneck). The beauty of this system though was that the AI could still learn trading and all I had to do was turn up the amount of random trades when it came time to play-test the strongest AI. That way it could offer meaningful trades when it wanted to rather than getting lucky (somewhat) on the random system.
The trades were very weird at first, but as the AI got better they became very balanced. Most of the properties were valued equally. It started making very equal trades early on (eg. a light blue for a dark blue or something like that). I couldn't find any examples of strange trades. I think it's something the AI needs to play heaps of games in order to understand properly.
The auctioning thing was really cool the longer I thought about it. The AI were basically saying "if you don't bet on this I will get a huge discount." somehow this strategy is better than the traditional way since it was able to beat AI that bought things consistently.
I could probably get some sample games from the AI if that would help.
@@b2stud "The AI loved to build around 8-9 houses instead of making hotels. Not sure if that means anything." This is very common among strong human players. The reason for this is because the amount of rent stops increasing dramatically for most properties around the 3-house level. 9 houses is enough to get 3 houses on each of your properties, which is where the best return on your investment comes in with most properties. The amount of damage you deal to your opponents with 9 houses is typically crushing anyway. If somebody has to pay you $600 (New York Avenue 3-house rent), they're usually tearing down their own houses, dramatically reducing their own rents, and you're able to build even more houses comfortably. Basically, the person who lands on that $600 rent is usually screwed. One more important perhaps overlooked fact is the 32-house limit. Purchasing houses makes it so that other players are limited in their own purchase power for houses. I have played many games where we auction off all 32 houses (which is what happens when there is a shortage of houses), and it leaves the players with the normally powerful orange properties typically far worse off than those with the dark blue, green, yellow, or red properties. Since those stronger properties command higher rents with a smaller number of houses, and more importantly, because buying them at auction gives you a strong advantage (you bidding $199 saves you $1 off standard price; a player with oranges bidding $199 is paying a $99 premium), the more color groups that come into the game at once, the more valuable the ones that have raw strength become.
"The AI also loved to accept more trades as it was winning, which seems strange but I won't question it." It makes sense to me. Think of it this way: Every time you are involved in a (balanced) trade, your chance to win increases, as does the chance to win of the person you are trading with, at the expense of everybody else. When you are winning, any time somebody trades with you, you are just getting even closer to victory each time, making you completely unstoppable. This, again, assumes it is a balanced trade.
"The trading mechanism was to give each AI about 10 random trades to offer. The AI would then choose whether it wanted to offer that particular trade." It seems a bit like an unrealistic way to do it, due to the importance of negotiation in the game. Of course, given that they are AI, if they reached a point where they all were more or less "perfect", no negotiation would be required as they would simply always offer perfectly balanced trades the first time every time. Perhaps the lack of negotiation isn't that important given that. However, if the suggested trades were truly random, it does seem like it'd be tough for a skilled AI to really be able to flex its ability, since all 10 trades will almost always be terrible. Trades are very complicated in the real game, so giving them only 10 options would be extremely limiting. Another important factor is that after each trade is performed, further trades are fully allowed and should be also allowed by the AI until no trading is done anymore. You can orchestrate many complicated deals, including deals that involve all 4 players trading with each other and all coming out the other end with a color group and money to build houses on them.
"The auctioning thing was really cool the longer I thought about it. The AI were basically saying "if you don't bet on this I will get a huge discount." somehow this strategy is better than the traditional way since it was able to beat AI that bought things consistently." This is the only questionable thing that you suggest remained in the AI. Typically, when a newer player plays with me and my experienced friends, and they put a property up for auction in the early game, we are usually willing to pay many times its value. I've seen light blues auctioned for $500 or $600+ in the first couple of turns, and I think it's a reasonable strategy to bid that much. Money is almost completely worthless early game, since if you have important properties that people need, you can use that as leverage to balance your cash amounts with other players later anyway. I am unsure how any skilled player would win just by auctioning off all the properties they land on, since they would wind up simply having less power and control overall.
"I could probably get some sample games from the AI if that would help." Yes, please! I would love to see how these AI play compared to strong human players, and would love to analyze some of their games.
buying houses rather than apartments is also a fairly well established strategy, since the number of house are limited. By occupying up a higher house count by not buying an hotel makes it more difficult for your opponents to buy houses and in turn hotels
@@bbqtong148 according to another comment, the version of monopoly ai plays dont have limitation on the total amount of houses.
id love to see a twitch account that just lets AI play against eachother on games like this and just have someone commentating throughout the entire thing
You can watch bots play all kinds of stuff on twitch. Just set the filter to least popular.
@@AtomicCortex oooohh... Good one
@@AtomicCortex demon time.
Golf clapping is a must
Chess computers playing against each other is a big thing. Pretty cool to watch.
One thing that stands out is to remember that 7 is the most common roll in a set of 2 dice. and a roll of 6, 7, or 8 has about a 44% chance of occurring. With how often people go to jail, it makes capitalizing on the red and orange properties great, since there is a high chance they will land on those properties once out of jail. Very cool how the AI was able to notice this.
"The first thing you notice about monopoly is, that it isn't chess"
That sent me as we played monopoly in chess club
A perfect monopoly AI would be awesome to watch against normal players. We can see how much is luck and how much is skill. Could the monopoly AI win a significant amount of the time? Or just a little higher than expected? Would be very interesting
The skill in Monopoly comes with trading and knowing what to throw away and what to keep.
@@GaryCameron780 and what not to waste your money on.
@@GaryCameron780Cause every hand's a winner, and every hand's a loser, and the best that you can hope for is to die in your sleep
Real shame there's so much luck to the game. We'll never get to see Man v. Machine international Monopoly Masters tournaments. There will never be a Kasparov or Deep Blue of Monopoly. 🙁
The only time a human wins in monopoly is by pure luck or someone else agrees to an uneven trade
The ultimate monopoly strategy, and the one that I use, is to simply be insufferable enough that people want to end the game as soon as possible
There's a counterplay to that where you don't play monopoly at all
jerk counter-counter-play: I'm just gonna steal the cash from your wallet.
😂😂😂😂😂
Hey, this is the same strat my little sister loved. Shame that literally no-one will play with her anymore.
@@BonesCapone counter-counter-counter play I'm gonna set up AI to take your check every payday
Awesome video! Also LOVE the cute fish animations!
Glad you like them!
Yeah, the browns and the light blues are the most underrated. Everyone knows that oranges are fire, but a lot don't realize how powerful the first side of the board is. This is mainly because of his quickly you can get hotels on them. The greens are more or less worthless - as you'd bankrupt yourself to make them profitable.
I read somewhere that a grandma win a family game by owning the browns, while the other players are so stubborn to not trade their properties
Yeah, it's a MASSIVE snowball if you hit it early game.
some wacky strats include snowballing off of unsuspecting players early with the 4house first row, then selling those houses to acquire oranges, after which you just win
Greens are always the last ones bought, followed by yellow and boardwalk area
Yeah, I love buying the browns-👴🏻
Mad respect for the dedication put into developing the algo and explaining it in such a lucid manner! This is what you call quality content.
Your explanation of neural networks is by far the best explanation I've ever heard. Thank you. Subbed.
Glad you enjoyed it!
4:20 the strength of the browns is being able to buy and hotel them for barely 1/3 of your starting cash
even if it falls off lategame youve probably made your money back
From my playing, I’ve found you largely have to adapt your strategy to who you’re playing with. If you’re with careless traders, get the monopolies in quadrant 1 and abuse the house rule (32 houses on the board max). If you’re with seasoned players, you need to be a bit more cunning and operate a couple of moves in advance, while also trying to prevent others from getting monopolies at every possible opportunity. There’s really not one single way to play monopoly.
That said, I wonder what the AI would have discovered if it kept researching.
My kids have gotten better and better over the years as they should but i recently deployed the house rule and pissed everyone off again lmao they didnt understand why i wasnt plunking down hotels until it was to late and i had bankrupted 1 of them to then see the oldest realize what was going on and rage quit.... We just got risk this christmas and played a game over 3 days that took around 9-10 hours playing by the standard rules and with all the luck you can almost never move more then two territories from your starting spot which i didnt figure out completely until about hour 5 and that is when i was able to take control of the game they caught on by the end as i started to explain what i saw and how it was impacting my battle choices and they were able to implement the same strategy but a bit to late as i had them on the ropes and making sure to take back any land that was lost that gained them any bonuses and then just wait for luck to roll my way i cant wait for the next battle because i love to see their brains work out the nuanced complexity of the games and also how they recover from a blunder
In my games we do some funky conditional trades, like "I'll give you my pink if I never have to pay more than 150 on your oranges". I wonder what the AI would've done with possibilities like that.
In my experience even playing with seasoned players, leveraging the housing shortage is extremely powerful
@@anthonycope8637 glad you are teaching your kids a lesson. house rule for life.
Kinda like the limited number of roads and settlements in Qatan. Finitude is real, kids.
What house limit?
Plot twist: b2studios is an AI
This would be epic twist
We could all be AIs in a sense
Plot twist’s twist we all AI
Sorry, I'm having trouble with my connection. Please try again in a moment.
@@b2stud best comment in the history of youtube
The best set of properties varies depending on the number of players in the game. 1-2: light blue and orange, 3-4: orange and red, 5+: green.
I was playing with my sister, i owned purple, orange and red, also est and South stations and both the societies. She owned practically the rest. There hasn't been a time that she has gone on my property, never ONCE, Although i had an house on almost every property and in some an hotel. WHAT ARE THE FUCKIN ODDS
What is your basis for saying this?
@@xHannibal hes a monopoly dev
Green is garbage with any number of players.
@@Franzibald_ trust a men that every fuckin time had to pay 1000M, green is not garbage
The orange set is the best because it is the closest to jail and highest renting cheap property on the board.
Do this with RISK!! please 🙏 thanks for sharing the code! This is a great project. If you ever come across too much time on your hands, a deep dive video on explaining your code, including the thoughts behind your strategy and tool/algorithm choices would be such an exciting learning opportunity ! Love this!
honestly risk is no more than getting Australia or South America early.
@@Potato-ko3oc People, and I assume any bot would too, recognize that that's a good strategy and usually gang up on anyone doing it.
I agree
The game "world control" would be even cooler for doing this. It is a combination of risk & monopoly, so it´s realy awsome :D
Yes!!! A project on Risk would be great
This kind of reminds me of IRL arms races and development. Most of the time it's sensible, gradual improvements, but every once in a while someone just says "F*ck it, let's try circular warships and bomb-carrying bats".
Bat bombs were almost used in WW2 by the US. The idea was to release bats carrying tiny incendiary bombs, which would hide in random places and then go up in flames. But before it could be used the nuclear bombs were finished.
Pigeon guided missiles would have been war changers if we thought of it sooner than we did.
@@SaHaRaSquad Bat bombs were almost deployed against Japan but the nukes got used first and did the job.
@@AtlasReburdened you are cool bro. I did not think anyone else knew that bit of history
@@bkbmckee Or that *bat* of history... (I'll show myself out...)
Actually, if you can nab the browns early enough, you can hold houses hostage to prevent hotels from going on more dangerous properties. In this way, even if a player has a monopoly on a valuable set of property, you can prevent them from developing that property, leading them to potentially mortgage those assets.
only with the housing shortage rule. a lot of households dont play housing shortage
@@maddoxandhisband9146 That's why you insist on playing by the official rules, rather than house variants. After a couple hours of them begging for you to build hotels, tell them you have a better game instead, and bring out the copy of Catan you brought with.
@@maddoxandhisband9146 I mean, it's a house rule *_not_* to use the house shortage rule. That's why there's a limited number of houses in the board game.
I think the original version before Parker Bros. didn't have hotels, but just the next level in houses. So I forget: if you have to break your hotels and there's not 4x2 or 4x3 houses, you have to go to the level that remain like 1x2 or 1x3?
Small problem... one of the browns is literately 1 space away from the start, the the lowest roll you can get is 2... so it is impossible to get both the browns before gong around the entire board at least once.
The Brown properties are highly underrated. I play a lot with my son and nephews. Whoever gets the browns (and light blues), in our games at least, has a high probability of winning.
10:00 Real Chads mortgage Community Chest, Chance, and Luxury Tax
i was getting ready for it to mortgage GO somehow.
"You know what's better than altruism? Winning."
That sounds a lot like me in Mario Party.
It sounds like GlaDOS
I normally dont laugh at videos but, that scene got me
That sounds like a shouting match to me.
reasons why capitalism bad
"Buy the Browns" didn't we try that once and it got us in trouble? 😂
💀
💀💀
Bruh 💀💀
we?!
@@icicle2143 *ahem*
Let’s just leave it at we before something starts
Brown and light blue sets are always what I go for first as it is the easiest to get a full set of hotels for and the returns are significant for the cost. Often is good enough to greatly reduce the speed others can build at ..
i hope you don’t get burnt out on making this stuff, i’m always interested and you have a very nice voice and obviously know how to vary your speech. great work
Thank you!
I always find myself with all the railroads. Incredible early-game set to have, as at least 1 rail is within a roll of anyone at any given point. Later on, I found myself trading nearly everything I owned ( rails, utilities, spare properties, money add-ons) for the orange set and I would rake in the profits almost immediately.
Even just having 2-3 railroads is amazing, so ve you don't need the full set to get a set bonus. I'd be willing to trade Mayfair/Docks for the last railroad in most games.
My older brother who is a very smart engineer loves the railroads and he tends to rake in loads from them
My greatest Monopoly memory is owning all the railroads and lasting till the end as a railroad tycoon lol
I am Terra Forming Cyber Space Creating Technology and Redirecting human psychological pathways th-cam.com/video/KnCRUsLuytk/w-d-xo.html
Honestly, I didn't know monopoly could be played like this, with a very common draw outcome. For me, we've always played it with a set time that we decide at the beginning of the game (e.g. 2 hours), then play until that time is up. Then we add up the value of all of our money, places, houses/hotels, etc. The person with the highest amount of money at the end is the winner.
See the problem is that that makes Monopoly a game where somebody wins, when it’s actually a game where it’s who loses the hardest.
You are playing wrong then monopoly shoud take around an hour if you trade
How tf do u stalemate, it's impossible unless no one has a set on which case do some trades take a risk
@@ivoninga You shouldn't trade if that leads to someone besides you winning the game. Taking forever and having the game never end is better than losing the game.
@@DroppedBass I disagree, second place is better then no winner or last place. So its in my best interest to take a bad trade if it screws over other players more then it screws over me.
In unrelated news no one will play monopoly with me anymore.
This is an amazing video, looks like it was a ton of work. When I played with my wife, I quickly realized that the best strategy was to auction everything you didn’t want to buy and bc my wife didn’t understand what that meant, I was able to get everything for next to nothing🤷🏻♂️
7:29 is a neat reference to the BTD6 video, as the BFB notices housing. Good job, B2.
lol
ok
I saw this and absolutely lost it on a rewatch 😂
Oh, hello there...
Now that you've created this AI can you teach one of them the elfer technique? This will mean 1 AI hoards the houses and avoids hotels drying up the supply of houses to hold the monopoly on building up properties. It would involve more decision trees involving how many houses are in the bank, how many players can currently buy houses, and how close the AI is to getting multiple new houses if they get a hotel.
Also they need to learn WHEN to build their houses based on WHERE they and their opponents are on the board at the exact time of choice.
This is evil, monopoly shouldn't have a finite number of houses
Was just going to comment this. It seems the AI might not be set up to figure out how to do something like that, as he (falsely in my opinion) states that the nost efficient way to win is to increase the rent people have to pay. This is likely the goal the AI was likely set to roughly aim for.
@@Beam3178 It's the actual rules.
@@ZenoDovahkiin I understand that, but I don't understand why it's a rule
I’ve always loved the brown set personally it’s so cheap and earns a decent amount of money for the cost
try the light blue. still 50 for houses and about 1k rent(ithink)
@@yodaiam4099 Nah its 600/550
and, a good place to stack 4 houses... to buy ALL the houses, so opponents can't buy any.... then ONLY upgrading those to hotels, on a turn when you can buy back the houses... denying houses to opponents again.
Yeah same, after you buy the brown ones you can get free labor from them forever! They’re great because they work as good as any other one and you don’t have to pay them since they’re brown :)
@@Someone-qq9zd damn you got the whole squad laughing
5:14 Genuinely made me lmao. Was not expecting that
Amazing work B2, the level of detail you put into these videos is unparalleled parking!
Wow I didn’t expect such a handsome man in my comments
4 min in and i just want to say my own strategy. In my own personal experience, the best strategy i found was to control the Oranges, the Reds, and the Yellows. Secondarily the railroads and the utilities. If you can only get 2 of the sets between the 3, get the Reds and the Yellows. With people frequently going to jail they will have to constantly go by your tiles, and if you basically own an entire side of a board, you're almost guaranteed rent money. On top of that they aren't nearly as expensive to upgrade as the Dark Blues and the Greens, but they still will give you a shit load of money
I really like orange. If someone's in jail, they'll try rolling doubles. On either 6 or 8, they land on orange, as opposed to having to take 2 turns to get to red(unless it's the 3rd turn in jail and they rolled 11) and yellow, and those are more likely to be skipped because of it.
@@This_0ne_Person true
orange/red the best
It depends on the game but occasionally you can go for a early KO with the light blues
Yeah greens are terrible. We house rule their houses down to $180 in our games, and they're still the worst real set (browns don't really count).
Not gonna lie, I'm definitely in the group of people who would love a chance to play against this AI. It also makes me wonder how similar neural networks to the one shown would handle other classic childhood games, like LIFE for instance. I honestly just really love videos like this. Would rather we put neural networks to use for stuff like this as opposed to, well, basically anything else.
My strat I've used for over a decade is just buy the first monopoly, buy 4 houses on every property you can, and never upgrade to hotels. There are only so many houses in the game so when you can own 12 on one monopoly, the other three players fight for what's left. Then spend as much time as you can in jail so you don't have to pay rent as often. Just owning more houses than others player means more rent coming in than going out.
One small gripe.
You super glossed over that "buy orange" was human meta for Monopoly. If I knew nothing about the game, this line comes out of complete nowhere.
It's because (and I know you know this but for everyone else) it has the highest payouts for the side of the board which has the most traffic and likelihood to be landed on after jail.
@@LumpyHippo Yeah I feel kind of annoyed. Why upload this video if the conclusion was, "Just keep doing what you've been doing"
But there is a little more that he, and no one else here, has gone into.
When you buy up a property, first off, its always better to collect one of each color so that you have trading material, and to prevent other players from winning. This means auctioning and trying to make players pay up as much as yo ucan get them to spend in auctions, while then, theyll be cash starved but have lots of properties, then you have the money to go after and trade for properties you actually want. Thats 1 strategy.
The other thing is the reason why orange is best isnt just because its after jail, its also because theres FOUR spaces out of the, what, 8? that you can be teleported to. the pink one, the orange one, the utility one, and the train. 4 chances to be teleported onto Death Row. The only tiles on that row you can't teleport to are the 4 regular properties and the card draw space. You can teleport to free parking so thats a -1 modifier but going to jail easily outweighs that.
But, theres more. So the orange is also the best cost efficience. Yeah I like the first row because its extremely cheap, but people often teleport and theres no way to teleport anywhere on the first row, plus, going to jail totally bypasses row 1. So while row 1 is the most efficient, its the hardest to land on.
The other thing is that going for row 4 is stupid because you'll burn all your money unless players just give it to you for free. It costs too much.
So the 2 best rows are 2 and 3. 3 is a little expensive, but has greater rewards but theres no teleport for the yellow space. As far as overall advantage, Row 2 is the best because it is affordable, low risk, and high reward. Row 3 is moderately high cost, high reward, moderate risk. honestly you should grab orange and red as well as pink. Those I think are the best. But if you can only have one, grab orange.
Yea orange and red are my faves. Owning the expensive stuff is good if you need to mortgage shit to pay rent. Same with railroads. I love those to mortgage over getting paid off
@@Kugrox th-cam.com/video/skz9odeewpc/w-d-xo.html
Fr. Orange + Railroads = Board Domination
Even with trades, I'd consider the browns and cyans extremely good in a 4 player game. You can typically get them to hotel in one turn because houses are so cheap for them and both have solid returns once you max them out. Most of the games I win are getting the cyans early then using what I make off them to lock out other monopolies.
Browns + Light blues = housing shortage
@@DanielSong39 Yep. You immediate own 1/3 of the available houses and since 4 players means you typically won't have insane capital, landing on a square that costs $600 is no joke.
It’s scary to think of the time manipulation that AI can have. Learning at that rate in a simulated world to impact this reality. I wonder how specific you can get with the tasks you give the AI.
Actually it's quite the opposite. AI are horrible learners. Forcing them to play at lightning speed is the only way to get them to learn anything meaningful
@@b2stud your right. Ai learns slow... But there simply better then us given enough time. For example if you spent years playing side scrolling mario games and then faced an ai that spent that same time perfecting 1 run of a mario world it WILL finish before you. Infact the mario speed runners know that the ai is simply faster then them because of how precise ai can be. Humans can only go so fast while ai are always faster
@@carlinataylor93 that’s because an AI can tirelessly run tons of games in a short time. If a person played a Mario course 100 times, and so did an AI, the human would win because 100 *attempts* isn’t enough to be better than humans.
@@totallynothatguy this is untrue. If a human had the same amount of experience as an AI did, the difference is between the extremely precise movements of an AI vs the almost certain variations in play of a human player
Edit: actually I think I see what you are saying. But if a trained AI was put up against the best player, it would almost certainly win
@JWCfive If you gave both sides 100 _minutes_ to learn a task, there's a good chance the AI might end up doing it better because it can use that time far more efficiently than a human could, simulating the task thousands of times in the time it takes a human to physically perform the task just once.
However, if you gave both sides 100 _attempts_ to learn a task, there's a much better chance that the human will do it better, because both sides only get to simulate 100 practice sessions and a human will learn more explosively and make far more mental connections in 100 attempts than an AI would. The AIs improvement would be precise, but marginal after 1 attempt, but the human would eliminate a very wide range of possible attempts without testing them purely through prediction.
Your claim was that an AI would defeat a human with "the same amount of experience", but 'amount' would imply 'number of attempts' as the metric and that statement only holds true for arbitrarily large numbers.
AIs would defeat humans with the same _length_ of experience - excluding arbitrarily small numbers of time - because that brackets it under the constant of time, which is equal for both of them; the AI is simply more efficient in how it can use that time.
THIS IS SO BEAUTIFULLY EDITED I CAN'T YOU'VE EARNED YOURSELF A SUB MY MAN
You forgot that placement on the board when trading can make a difference (increase your chances someone will land on your property earlier). Also your cash can make a difference when trading (do you have enough cash to buy buildings or does your opponents and have money leftover in case of bad luck). That is assuming both players get a monopoly from the trade. I personally avoid trading were I don’t get a monopoly if I give one. I also would never trade to give a monopoly if I had one and my opponents don’t. Then there is trading for cash. How much should you trade and when? That is determined by the monopoly set you have a monopoly on and your opponents set. Plus the amount of cash each person has. Obviously if you end up with a cheap set you can afford to give up more cash depending on how much you have and your opponents have. Over all it’s a great AI but in my opinion it lacks the human ability to think outside the box and adapt. Or should I say that the interpretation we humans get from it does not reflect the way too many variables to copperhead. Just my opinion as a monopoly lover who has a great win lose record. Lastly there is not much you can do if your opponents make a bad trade with someone else or you have lost because of luck so decide to help one or another player by making bad trades. I did enjoy the video and it was very insightful.
The game is solved. Trade for Reds, Oranges or Yellows offering whatever it takes. Mortgage everything you have to build as much as possible. Try to stay in jail and just wait. Any income goes back into housing until you get full hotels. Only need one of the 3. I only had reds and took out a player who had all blues and browns with hotels. Nobody was landing on them because the probability is so low while I was cleaning house.
I loved the programming theory, but the animations and your sense of humour elevated this to a top-notch video! Definitely gonna subscribe!
The butt smack killed me. I went back and watched it like 5 times.
In the first half of the video, you mentioned that the later a property comes on the board the less chance there is to land on them because of the teleports. You also mentioned the winning AI tended to have the blue properties. I think that probably explains why the brown properties were often in the hands of the winning AI, not because they contributed value to the win, but because if you get to a blue property you're very likely to also land on a brown property on the next roll.
another explanation is that brown properties are the cheapest to upgrade. You spend a little and start making a LOT. while "go to jail" or nearest X are common, they should essentially be ignored because both players have an equal chance of drawing them
You have to roll a 2 or 4 from boardwalk to land on one of those first properties. That is not probable.
@@senorpepper3405 I didnt know boardwalk was the only blue property on the board. My game must have a misprint.
@@NixPanicus why do I have to explain myself, dude? You know what you said. But let's revisit, because why not. If you get to a blue property you're also very likely to land on a brown property next roll. Your best shot is to be on park place and to roll a 6 and get that last brown property. Your words say it's next to an inevitability. As if you go 2 times around the board and if you landed on the blue spaces you are most likely going to get on the brown ones when I say that's a stretch.
@@senorpepper3405 An inevitability. You got me. Those are definitely words that I said.
It would be interesting to know how fond of gambits the AI became. Skill in Monopoly doesn't really stop at knowing what properties are valuable in general, but in knowing what properties are valuable in context. A lot of the time a game is decided by two players choosing to gambit with opposite ends of a trade, turning a 4-player game into a 2-player one with some complicating factors, but judging how deeply to gambit is super tricky.
8:21
Dunce^2 has to be the funniest thing I've seen in a while
How about Dunce^3
@@kaikwokhoratiowong1023 close but it will never top dunce^2
One thing I wonder about is... did the AIs ever do anything resemblent to forming alliances with one another? Such as the three other players teaming together when one of them was gaining a significant advantage. Or the player who clearly lost and has no chance for catch-up just choosing to assist someone else to try and give them a win, which might happen in over the board family games that have a social element to them as well as numbers. Stuff like that is something I always enjoyed. Did any of the AIs do something even remotely similar? :o
They almost always went free for all, no alliances as far as I could tell
@@b2stud Thanks for the answer! 🥰
The ais ability to actually do that is likely limited. Since the only way to actually team up on a player is through house rule implementations. The only thing I can see the ai realistically doing with real monoply rules is lowering the amount of trades they do with the top player.
@@lifetake3103 the problem is that every ai tries to win. You cannot win in an alliance. The weaker of the two will lose and therefore has no interest in the alience. But what if we just don't want them to go bankrupt first. That would encourage the worst player to allience with the best or 2nd place to make them win in order to not lose.
I just thought that through but the most likely scenario is that the best 3 will alience against the worst performing. That way they have the highest combines win chance.
“Buy the Browns?” I’m pretty sure that was outlawed awhile ago
that’s what I’m saying bro 😭
As a kid, I used to love buying the Browns, the Dark Blues and the Green Ones. I also used to try to get the utilities and as many railroads as I could. But we almost never traded anything. Our games would last for the entire summer. We would leave the game board out and come back to it when we felt like it. We hardly ever finished any games. Haha, those were great summer days!
The browns are a great way to buy and store up to 8 extremely cheap houses early in the game which can come in handy when the game gets intense
Handy for what? Losing money?
When the houses run out... @@adolfhipsteryolocaust3443
@@adolfhipsteryolocaust3443 no you sell the hotel to take up 8 houses when your opponent have monopolies, thereby choking the supply. And since nobody wants to sacrifice their chances to help others, you will get situations where there's only one or two houses on the expensive knock out properties
@@adolfhipsteryolocaust3443 handy for creating a housing crisis if you own 8 houses that's 8 houses not on your opponents property's. Never upgrade to hotels that's just letting your opponent build houses to sting you with higher rent
No, let him cook. The idea is that there is a limit to housing and hotels. If you get the browns and stock up the 8 houses, you can store pieces for yourself in the late game. Then using those houses you can upgrade better properties while everyone struggles to get enough to make a difference. It's a good strategy especially if you get lucky with a roll of 3 early on to get the first brown.
The construction animation cutscene was easily the best part of this video. Cool research.
3 things to consider:
1. The browns are on the winning team, because they are brought by the ones surviving the lap. That's most probably the player with the blue streets.
2. Rolling two dice, you have a very high chance to roll 6,7 or 8 (with 7 being the most probable). As the chance of going to jail is pretty high, and you leave jail by rolling two dice, you have a high chance of landing on orange.
3. Buy houses but don't buy hotels. As the amount of houses is very limited, and you have to buy 4 houses to get an hotel, you can block your opponents. If you have got all the houses, they can neither buy houses nor hotels. I guess the model didn't "know" this.
Buying three houses in each property gives the best ratio of income to investment.
3. If all the players have s brain, when the house stock is empty you can use papers...
Rolling any number between 2 and 12 is the same probability
@@mattsmith5421 no, not with two dices
For a 2 or a 12, there is only one combination (1+1 and 6+6)
For a 7, there is many combinations: 1+6, 2+5, 3+4, 4+3, 5+2 and 6+1
@@mattsmith5421 How many combinations are possible to roll a 2 with two dice? -> 1&1
How many combinations are possible to roll a 7 with two dice? -> 1&6, 2&5, 3&4...
1 in 6 rolls will be a 7, but only 1 in 36 rolls will be a 2.
I've always tried to get the first row of properties. Ever since I was a kid. You can stack hotels in no time flat. The secondary goal is the railroads. Also knowing how to empathize with people and understand what their mindset is in order to understand your leverage and visualize your potential trading paths in order to acquire the highest value monopolies. Monopoly has VERY LITTLE to do with "luck"