Remember kids, if you have a 90% accuracy rating, you're going to miss half of your shots. Also if your foe has a 5% critical chance, you are *going* to die.
I feel it’s worth noting that a large enough playerbase is *going* to find out that a system is pseudo random, so it has to still feel good knowing it’s pseudo.
@@kindacruiseThe important thing is: The average, casual player, cannot naturally discover a way to chease the game. If they need to search outside the game in order to find ways to exploit its mechanics then is all fine. (except for multiplayer/competitive games)
Looking at minecrafters who figured out how to use Redstone to manipulate the games random tick generators and make the odds in their favor for anything
One good old tip to fail at RNG is to add it to a speedrunning-focused game. Just ensure the following: 1. The RNG must not be something the player can react to within a reasonable amount of time in a way that no time loss would happen. 2. The favorable outcomes should be very small (but still existent enough for a speedrunner to get lucky once in a blue moon, so the strategy involving RNG wouldn't be banned) 3. The RNG cannot be manipulated or bypassed. Remember to remove any exploits that could do that. 4. If you think your RNG is in a short supply, save it for the end of the game.
I love how League players will tell that if you get 7 losses in a row with 50% WR the system is undoubtedly rigged when it's actually proof of the opposite
@@ultimaxkom8728 the funny thing is that even assuming the lowest possible winrate (20%) you only have a 0.59% chance to get a 23 games losestreak ((1-0.2)^23), bro is even lucky at inting
it's also very crucial to note that rng is always objectively bad game design and should never be implemented for any reason, no matter what. if a million people play your game, not one of them should have a single unique experience
Replayability is horrible, why give the player an opportunity to enjoy the game as much as possible for as long as they want when you can make another game for them to buy and enjoy again?
Although true output RNG is the worst type of all, it has 1 advantage over others. The emotions that generate when you get the crit you very well needed are absolutely priceless
@@Poke238s To be fair to Limbus, a single turn probably calls up the chance roll like 50 times more than a turn in XCOM, so that fail chance has way more opportunities to show up. That said, I have no excuse for whatever the hell XCOM is doing with point-blank misses...
You see how many people make randomizer mods for games, where random items or events happen at random and (typically) incorrect times? That means that's what players want in a game, so your game should be like that, too. Don't worry that players don't have context for when these events are meant to happen. It's all part of the fun, and it's not going to confuse players at all.
The more random your game is, the more personalised their play experience should be, so therefore you should make everything random to deliver an experience that can cater to everyone
If XCOM taught me anything, it's that a 90% chance to hit at point-blank range means jack shit. I guess they were employing true random. Didn't know about the distinction between pseudo and true random before this video, so I did learn something from this after all. Thanks!
XCOM uses true randomness, but on lower difficulties they have a constant multiplier to the displayed values (1.1 or 1.2), stacking aim bonuses for every miss in a row (but only if the base chance was at least 50%) and even MORE stacking aim bonuses if you have less than 4 characters alive. Plus equivalent debuffs for enemies.
@@tipoima The hidden bonuses for repeated misses are also capped at 95%. But also none of this is quite enough to deal with people not properly understanding just how often that 10% miss chance will come up.
Unironicslly all jokes and satire aside, I ended up discovering this on my own. I was making a superboss, and thr rng of their attacks compared to their cooldown phase sometimes made it really frustrating to work with, so eventually I implemented a counter that increments based on how many attacks were used before forcing a cooldown, where the counter resets. It made the boss go from annoyingly difficult to actually entertaining to replay while keeping its difficulty. I honestly wished this video came out sooner because it took me a month to realize on my own.
@@steluste Valve is notorious for their unbalanced and overused rng to the point that people started calling their ceo and founder, Gabe Newell, rnjesus, that’s why it’s the Valve method.
@@Raspingpython-10x can you give an example where they used rng? Cuz I don't remember anything random even though I played all hl portal left4dead cs games
I first learned of this kind of stuff from the Slay The Spire wiki. That game has a luck system where every non-rare card you get bumps up your chance of getting a rare card until you do get one!
This sounds similar to a chain system that's used in a pokemon-like game i play. You KO multiple creatures of the same type, and each one that is normal contributes a point to boost your chain. The higher the chain, the higher your likelihood of catching a 'shiny'. And it resets one you do get one.
True RNG sounds like an absolute nightmare, and that's because it can be - but that frustration can do something interesting. In Fear & Hunger, the main gimmick of every action is a coin flip. Looting a container? Flip a coin! Dodging an instant-kill attack? Flip a coin! Sleeping? You better have a quarter in every pocket, because you are FLIPPING A COIN! The game is ansolutely brutal and sounds like a nightmare to play, but the coin flips further the game's central themes - that life is cruel but ultimately fair. There is no force altering the world against you... the world is just built like that. It's a really interesting case study on RNG in games, and I would highly suggest anyone looking into RNG at least give the game a peak. Just, please don't copy it one-for-one, I don't want to feel more pain!
I felt less pain in fear and hunger than i felt trying to farm for all the dev sets in terraria and killing 300 fucking destroyers AND STILL not getting that last set because rng
Gotta be honest, I will never design an indie (not because I don’t think I can, it just doesn’t interest me), but starting your videos with “the best way to fail is to never try” is genuinely some inspiring shit. Really rings true. Keep up the great work!
There was someone that made a challenge run of Pokemon where the RNG will always be against them. The opponent always crits and triggers side effects, while the player never does and will always miss if there's a chance to. Funnily enough, it's actually less frustrating because all of these become the expected results.
And when players complain about the RNG just point out that it's not technically random because computers aren't capable of actual random number generation
@@jirachido3997 Computers are purely logic based because what a computer boils down to is a insane amount of true and false statements (1 and 0). So we cannot do something like lets say a dice throw because in the computers world, unless we change the way the dice is thrown, the dice will always fall the same way on the same side. So how do we give random numbers? Most of the time its by using our clock. Either the program itself has a build in clock that counts the time since startup or they use your computers clock. Either way, time is a value that has a lot numbers (hours, minutes, seconds, miliseconds) and will almost always be different for any player engaging with a specific RNG event. So we might both start the game and immediatly throw a coin but I thrown the coin 5 miliseconds later than you did so our outcomes might be different.
@@diefontysstagiere5395 Or if you wanna have a blast, you hook the computer up to the random website which monitors atmospheric pressure changes for RNG
@@diefontysstagiere5395 adjust glasses Actually quantum computers use qubits so they aren't just 1 and 0. so they should in theory be able to give us a more true random.
Brings back Terraria memories, where the game has a 0.1% chance of dropping something, but whether you actually reach a moment where you see it dropping throughout your entire life is 50%.
Fun fact, pseudo randomness was used in completely wrong context in this video. All simulated randomness on the computer is using pseudorandom algorithms. Computer can not give you "true" randomness. So what video calls true randomness is actually pseudo randomness. And this computer pseudo randomness can have interesting effects on the games. Like reloading same save state giving same "random" results, due to keeping same seed for random numbers, etc...
Many modern computers can do true randomness these days. From using sensor data of fans and temperature, to using quantum effects. It's just much faster (and easier to debug) to use pseudorandom algorithms seeded with true random values.
@@tipoimaFalse. Computers cannot generate true randomness. In fact with the possible exception of quantum states, nothing in our universe is truly random
True, from a development POV, but everything here is from a game-design standpoint. In that sense, pseudorandom is more akin to curated, or rules-based random, vs just grabbing whatever Mathf.Random() gives you.
@@tipoima No, @david's right, that's just randomizing the seed, doesn't mean you're getting true-random output. But yeah, you're right that for game design & testing, working with fixed-seed (or at least saved) is the way to go, whether you're adjusting it afterwards or using it raw.
I actually expected you’d criticize pseudo randomness rather than true randomness. I thought you’d mention how a game might use a fixed list of "random" numbers, and that, over time, players could recognize patterns and even memorize the sequence if they played long enough. I vividly remember learning the RNG of some games such as the minigames of new super mario bros ds and the CPU's decisions in mario party games.
This is a good point actually, I considered talking about Stardew Valley's random system which can be controlled by players who understand it, so yes, pseudo-random can be damaging if not implemented well. Same thing with the example I gave, players will purposefully hit something weak until they've reached the max number of hits before a guaranteed crit, then hit the stronger enemy with that starting advantage. Kind silly how far players will go for min-max a game. :)
In older games like the original DOOM, this was mostly a hardware limitation, since generating random numbers on the go was apparently too expensive for early CPUs, at least while also trying to run a game at the same time. id Software sort of got around this by having the game make random number calls for nearly everything, including enemy idle animations, so that the number you were going to get from the lookup table was always changing. The final effect was something reasonably close to "true" randomness. On the other end of the spectrum is the Happy Lucky Lottery from Paper Mario: The Thousand Year Door, which is straight-up rigged. The winning numbers aren't randomly generated at all, the game simply checks how many IRL days have passed since you bought a ticket and awards prizes according. To add insult to injury, if you manually change the console's time and date, the NPC running this fake, rigged lottery will accuse YOU of cheating and have a big meltdown about it. The nerve of some people.
@@BarioIDL Unironically, yeah. In a single-player game, if people are dedicated enough to put together forums, compare notes, and write tracking spreadsheets, then congrats, you made a helluva game.
@@generalrubbish9513 I honestly had no idea DOOM used that method for its RNG. The date and time RNG feels like such a stupid move from the developers. And looking back, it’s funny how younger me could predict a game’s random number generator where it really showed just how flawed that method of pattern based RNG is. After a while, instead of feeling random, the game just becomes a predictable grind, where the developers expect you to be surprised by the randomness, but you just end up memorizing the pattern and going through the motions like a robot.
This first bit is effectively a critique of how Pokemon handles randomness with accuracy. Focus Blast is an essential metagame move. It is 70% accurate but it does not feel like it is. It feels closer to 50%. It seems that alot of testing is necessary to find good ratios to make randomness feel fair.
I'm not sure if the coin flip bit was because it actually fit into the rng bit, or bc RNG related stuff bothers Artindi so much that they found a way to give actual advice to try and stem the RNGesus bullshit in indie games.
I think a good way of makijg a balanced pseudo-random mechanic would be implementing the Gambler's Falacy in your game's RNG. Basically, you make the chance of a event happening depends on the results from before, increassing in 1 the Power of the Denominator of the fraction representing the chance of it happenong in 1 for each time it didn't happened in a row. Like, in this system, if the chance of sucess/failure is one half, but you suceded/failed the last time you did a Check, the chance of suced/faik decreasse to one fourth. In D&D terms, is like you receive one cumulative Advantage for each failed Check in a row, and one cumulative Disadvantage for each Sucess in a row.
This is both less efficient and less effective than using mixed bag. Example: player flips a coin, it can land on either side or balance perfectly on the edge... So you create 256 coin results... Two out of 256 will land perfectly on the edge, 127 will land on heads, 127 will land on tails. Use your rng to merely shuffle the bag. Now you only need to prepare bags in bulk, making the rng code more efficient because it doesn't interleave with your other code, in fact the bigger the bag, the better the performance... Now player flips a coin, simply use the next value from the bag, very fast, simple, and fair. Player will always get 127 heads, 127 tails, and 2 edge cases per 256 flips. Sure, it becomes predictable, but player cannot game it, and in singleplayer game, let player have fun, so why worry? And you can reuse the bags, don't remove any values, simply advance across the bag till the end, then reshuffle for reuse. And no, memory isn't really a problem, even in a MMO if you do this, one shard that holds 2000 players will need only measly 500KiB (256 bools, one byte sized index to know where you are in the bag) to hold a bag for a player where ANY 50/50 chance with a small twist is required. Sure, some more complex systems will need megabytes of memory, but all modern games eat gigabytes of RAM so megabytes is laughable. I can easily build a MMO backend meant to have plenty of monsters, bosses and random other crap using this technique and use under a gigabyte of RAM for one 2000 player shard, go figure. If really pressed for hardware resources, I can do it in C and allocate those 2000 player bags of all varieties statically and have memory usage constant at all times aswell. So it's simply amazing, and I'm not sure why most games don't do this.
@templeofdelusion Cool and well tought. If you get too much of one result, it become less probable because you're increassing the number of options of this result and maintaining the number of ways of the other happen, doesn't it? Thank you for the advice.
Ah so basically I should make the ending of my game based on a coin flip. It doesn't matter what you do, just get through the final dungeon and then a coin flip decides if there's even a final boss, and then more coin flips to determine what happens after you beat the boss (if it shows up) Actually that's really funny, I shouldn't give form to intrusive thoughts...
Fun fact: Computers can't actually create true randomness so instead they just use algorithms to make a number that for all practical purposes is random but isn't actually random
What I remember from computers making "random" things is that they take some... "inspiration" from other things. Basically, you could program your game to look at the date and time, then use *that* to make your "random" seed. You can also use stuff like player inputs, such as how long they were on the title screen for, or where their cursor is located. Stardew Valley for example uses the player's step count to determine events such as weather or saloon stock. This is just from what I remember about computer RNG, I may be *very* incorrect.
Loved the video, but side note (a semantic one, I am guessing you researched this): the terms True Randomness and Pseudo randomness are terms already taken up by math. Usually they describe number sources that are provable to be random and not biased in any way (most computer generated random numbers are pseudo random just based off of computers being deterministic). True random sources are complex and most are resulted from quantum effects, like permisson of photons through a dichroic mirror. If we were to use a term to describe the manipulation of randomness that was used in this video, the correct labels would be skewed and unskewed. I'd go more into depth about randomness but the medium of youtube comments is tough to use.
You are correct, in the context of math and computers the terms mean completely different things. In the context of game design, I've been seeing the same terms be used to mean what I described in the video. Human level perception of true random, and dev manipulated random. If there are better terms I'm sure people will start using them, this is just what I've seen be used to explain these concepts thus far, perhaps because game developers often have degrees in computer science.
Yume Nikki was my bane of existence, the fact the poop hair effect is locked behind luck, because the location of the bed that teleports you to the only area connecting to that specific room, is... guess what! random! Sometimes it can be literally on your nose and sometimes you gotta do alot of stuff to reach it The zippertile event also sucks since it's totally random In my visual novel, i chosen to make RNG affected things not affect gameplay and instead just be silly jokes that have a 10% chance to appear
I don't mind things like the zippertile being completely random. Feels like something that would be amazing to discover on like your 4th playthrough of the game or something
@@FluffinWasntHere I'm fine with it for optional events personally speaking. It's more of a problem when it comes to effects. Since you actually need those to beat the game
well zippertile is supposed to be a secret, wouldnt be much of a secret if it wasnt totally random and the bed thing isnt random since u can always check all the beds, wuts the big deal
@@yandere8888 it is random since the bed that teleports you to the hands world isn't the same everytime It's too much of a hassle to get THAT specific effect, and that's ingrained into progression since you NEED poop hair to beat the game
This reminded me of a thing I wanted to make in pretty much all games I'll make one day. The first one I'm planning to make is (yet another) Phigros-inspired rhythm game, and I wanted to do a little thing with the loading screen tips. Instead of just randomly being chosen, the game would keep track of certain things and pick more relevant tips on loading screens. For example, if you get bad scores too often, it'll tell you to take a break, or if you consistently hit notes too late or early, it would suggest you to change your offset. I don't think I've really seen that in games but it seems like a neat thing to do
Artindi, you forgot about the best type of RNG ever! Randomized characters! That's right, the best games all require you to gamble to get a chance of *maybe* getting the character you want! And do remember to either design your game around a very specific ultra-rare character, you wouldn't want the uninvested player to enjoy the game. Alternatively, design the game around no character at all, I'm sure nobody would be turned down with a story so bland it makes white bread feel spicy!
Xcom2 uses a ball in bag model for random chance. When you fire, you pick a hit/miss ball that can't be picked again. So if you miss 5 times in a row, there just aren't gonna be any miss balls left in your bag
It doesn't matter that much, but usually the term true randomness has a different meaning. Technically, true random numbers are the random numbers that are obtained from actually flipping dice, atmospheric noise, etc. Pseudo random numbers are obtained using a formula and are called that because they aren't actually random and can be predicted and reproduced, if you know the algorithm and the seed, and these are generated using System.Random or the like.
(Not being sarcastic) Try to strive for Input Pseudo randomness as much as possible (at least when the player interacts with it directly, such as with weapons). A good example of randomness would be Slay the Spire (the game RNGs enemy attacks and lets you defend against them), and a bad one would be Xcom (you shoot an enemy, and only then see whether you missed or hit the enemy). Also, if you're making a pokemon style enemy encounters, please make it so that encounters are more rare when you walk in straight paths and more ofthen when you walk in circles
Different categorization, can apply to all 4 quadrants described here. And it's more like independent vs dependent - if you draw a card from a deck, you've now changed the probabilities of drawing each of the remaining cards. While every roll of a die is independent from those that have come before.
I’d say there’s another grouping of RNG: Dice or Deck? 2 roles of the same die have no impact on the next By drawing a card you change how likely you draw a different card For an example of Dice RNG take Minecraft, killing a mob with specific drops; the chances of a mob dropping what you want don’t change by the number of that kind of mob you kill For an example of Deck RNG take Tetris, the blocks you get; the game basically shuffles a deck of the 7 tetrominos, giving you the top one until it decides to shuffle another 7, this makes it so you won’t go too long without a piece you need
Seems like True Random could also work well for games where frustration is part of the appeal and memorizing the progression steps is *not* part of it. So, not "I Wanna Be the Guy" (where you can master it), but maybe games like I Am Bread, Getting Over It, QWOP or the like? Not sure which ones would actually have randomness (or random tweaks) as a mechanic, but I could see them working in that type of game specifically.
Funny enought, I'm currently attempting a cheese wich had a FIVE TO SEVEN layers of RNG.... and ...this...is...pain I'm trying to defeat Jetragon at lvl 1 in palworld (a reminder than the cheese to make jetragon fall to its death has been long time patched). It require : - Rng roll 1 : Forcing spawn and despawn until ONE specific group spawn happen - Rng roll 2 : Wait for Jetragon to go where you want him to be, the mildtest rng of alls but sometime it can make you wait some extra time. - Rng roll 3 : After baiting the two to make them fight, hoping jetragon will catch fire (and no, the flame crossbow don't work with a large lvl difference). - Rng roll 4 : For rng roll 3 , hope jetragon among all its attack don't choose the two attacks wich are almost guaranteed to nuke the whole group in one swoop before jetragon could start burning. - Rng roll 5 : Can be mittigated but it can be very annoying, hope other wilds pals will not bother you by going too close and start aggroing you. - extra rng roll 6 : To mitigate rng roll 3 , aggro jetragon at step 3, and hope jetragon will keep aggro on you rather than doing the worst scenario anyway and then comming after you. - extra rng roll 7 : following step 6 You can put yourself at a place where Jetragon AI can't attack you or are is dodgeable, until his aggro reset because he can't reach you... except it has one heatseaking attacks it might decide to use anyway and you are likely to get hit by it. Repeat step 1 and 3 again and again hoping rng roll 3 don't troll you to death. (and some minor rarer rng rolls too ). This is hell, I though I've found a clever cheese strat but I accidentally created a RNG gauntlet.
Mario Maker is actually genius. As a feature, the RNG is based on player input. Meaning, if you hold right and only right, you can get the exact same results each time.
Brawl Stars had a thing that boosted your odds of a rare pull if you kept failing enough. Still entirely RNG, but you're statistically gonna get something
One thing to note is that unlucky people will still be unlucky no matter how high you rig the chances, and lucky people will always be lucky no matter how high the enemy's odds. However, if you rig the system against the unlucky, they'll do even worse, and if you rig the system towards lucky people they'll do even better. This is always true, you can't make these groups meet in the middle, and even if you tried life would still find a way.
Good thing this has nothing to do with people. Also it's trivial to equalize everyone, just use mixed bag randomness, that way noone can get too lucky or too unlucky, they will have things happen right on the droprate...
@@templeofdelusion Already thought of and failed, you will get what you need exactly when it's least useful and possibly detrimental. Impossible to equalize luck.
This is why I never use a move in Pokemon that has less than 100% accuracy. I don't care if it's the strongest move of it's type, it WILL miss and it WILL get me killed at the worst possible moment.
Great video. It’s worth mentioning that definitions for true and pseudo randomness are different in other contexts including programming in general. However, in this closed context your definitions make sense
A classic failure of RNG implementation, specially in Tabletop RPGs, is to put the whole result behind RNG. An exagerated example: You go to the king and try to persuade them for a higher reward. You fail. Now you didn't just fail on the extra reward, you now have no mission and are banished from the kingdom. Good job. A much more typical example: The sword with the 50-50 chance of doing nothing or ALL damage. Another example: You try to sneak past a guard but fumble stealth, now the whole base is alerted. The better solution is to make the guard alerted to *something happening where you are* and acting accordingly - investigating, calling for backup, etc. But no full-blown alert from the get-go, the guard doesn't even know what's in there (unless it's a very high security area). The key is in respecting the player's decisions. The player isn't rolling dice to see if their plan happens, they're rolling dice to see if their actual action succeeds, to see if they can avoid alerting this guard, not the whole base. Otherwise it's the failure equivalent of a player going "I roll to infiltrate the big bad's lair."
In Warframe each character requires grinding for the different parts from different missions. One of the more egregious ones was Octavia. Her kit completely trivializes the game, but her neruoptics only have a 1 in 5 chance to drop once every 20 minutes in an annoying survival mission. I played for almost 2000 hours and never saw her neuoptics.
Not to be confused with pseudo-random number generators, which spew out a deterministic list of seemingly random numbers (and is most likely what you are using when you go the "true" random route). Computers are actually really bad at flipping coins. I think it's the lack of thumbs.
Yes, in actuality, (unless perhaps you have a quantum computer,) all the "random numbers" are pseudorandom. So we are using the terms "true" and "pseudo" a bit abstractly here for the sake of understanding game design. :)
The best way to implement pseudo-random RNG is to punish your players when they start doing well. Everyone loves the blue shells in Mario Kart, especially when they hit you on the last lap when your only a car's length from crossing the finish line.
Remember, if 100 people toss a coin 10 times, 10 of them on average will get 10 heads. So, if your player base is large enough, you will have a significant number of players getting a 10% Bad Event happen back to back to back, and never land the 90% Good Event.
The people who designed Aegis Cave watched this video and decided that they should make a level where output true randomness should be essential for progression. They should have put more levels like Aegis Cave into the game.
Regarding loot drops, its nice to have a way to trade in lots of bad items for a good item. It can be a thousand to one, just give me a means to actively obtain the good items with all my bad drops. Not a secondary currency which you never use; something which can n be used to make progress in the getting the best item drop.
Reminds me of Bad Piggies in some way where you can get scraps for getting skins you already obtained, and those scraps can be exchanged for better skins
so if you look past the sarcasm its trying to say output rng is bad and given input rng wasnt even mentioned after its introduction at the start that its good kinda makes sense as world generation a possible use of input rng gives every player a more unique experience, while say lootboxes or even crits, an output rng make the player rely on luck which overall just doesnt feel great even when you get lucky because at any moment your luck could run out and its over either way it seems more fun to both make and play a game with stuff like skill based crits instead of chance based ones, aiming for weakspots, or maybe a timed minigame sorta thing depending on your game
The interesting thing about online roulettes that they are pseudo-random. Like I constantly get sequences of 10 red in a row, which only has odds of 0.24% happening.
I approve this video
At last, my offerings have been heard.
In rnjesus we trust(probably shouldn't to be honest)
@@barrygamer7449Exactly: probably.
I would like to get a quote on sponsoring a video, please.
@@Alphanerd2 i mean, you can send me an email, but I have to really believe it's a good product before I would try to sell it on my channel.
Output? Input? True random? Pseudo random? I can't believe it. I almost learned something.
"Almost" So didn't right? RIGHT?
Why are you trying to steal my identity
@@jipcoumou4935huh, why are you stealing my identity
@@jipcoumou4935 Because its funny Purple J
Why you talk like marshu
Remember kids, if you have a 90% accuracy rating, you're going to miss half of your shots.
Also if your foe has a 5% critical chance, you are *going* to die.
This gives me an idea of a game where the actual chances are the inverse of what's actually shown to you.
@@valonyaver600damn you. Well, I guess you did follow the video.
@@valonyaver600 well thats just your average XCOM game
This is why you can’t trust VATS.
I am a fire emblem player and i can confirm this is accurate
Hold on, I need to find a coin to flip so I know whether I should laugh at that joke at the end...
So you're saying there's a chance....
@@Artindi
Heads = No
Tails = No
But maybe there is still a way to get that Yes? 🤔
If it lands on the side ??
@@kambuntschki6314 I've actually had that happen once, flipping a coin on the street - the legend is real
@@Latias38 pseudo random
I feel it’s worth noting that a large enough playerbase is *going* to find out that a system is pseudo random, so it has to still feel good knowing it’s pseudo.
Yeah, or more complex and hopefully built in a way it can't be cheesed.
@@Artindi i mean they found out how to exploit stardew valleys step count rng somehow
Especially if you show percentages, spend enough time with true rng and you'll notice when that pseudo 75% is actually closer to 90%
@@kindacruiseThe important thing is: The average, casual player, cannot naturally discover a way to chease the game.
If they need to search outside the game in order to find ways to exploit its mechanics then is all fine.
(except for multiplayer/competitive games)
Looking at minecrafters who figured out how to use Redstone to manipulate the games random tick generators and make the odds in their favor for anything
Pokemon moves with 80-90 accuracy
rare Shyguymask sighting
We love imperfect accuracy moves
Don’t you mean 50?
70%*
scald burns and focus blast missing are 30% accurate, not 70%
I call it "Bleakwind Miss" cause it never lands when you need it to
One good old tip to fail at RNG is to add it to a speedrunning-focused game. Just ensure the following:
1. The RNG must not be something the player can react to within a reasonable amount of time in a way that no time loss would happen.
2. The favorable outcomes should be very small (but still existent enough for a speedrunner to get lucky once in a blue moon, so the strategy involving RNG wouldn't be banned)
3. The RNG cannot be manipulated or bypassed. Remember to remove any exploits that could do that.
4. If you think your RNG is in a short supply, save it for the end of the game.
Minecraft
@@givlupi2686 Why do people even speedrun the most luck-.based games anyway?
@@artman40 Usually because it has something else appealing about it, or they like the base game.
@@givlupi2686 Or perhaps it's a Nintendo game which has speedrunners no matter how RNG-dependent or downtime heavy the game is.
@@artman40 Nintendo really hates glitches to the point that they'll patch glitches only speed runners would even know about, just to screw them over.
The bane of a speedrunners existance
ha ha, it's true. :)
And also their greatest ally
3:40 "Did you get that joke?" **dies**
I love how League players will tell that if you get 7 losses in a row with 50% WR the system is undoubtedly rigged when it's actually proof of the opposite
So what your saying is my 23 game lose streak is proof I'm actually good at the game? ;)
@@Artindi You are certainly good at dodging bans ngl
Coin flips.
0.5 chance. 23 times. 0.5^23.
0.000011920929% probability.
@@ultimaxkom8728 pretty sure his "true" wr is not 50%, we can't be all pro league gamerz, someone has to touch the grass
@@ultimaxkom8728 the funny thing is that even assuming the lowest possible winrate (20%) you only have a 0.59% chance to get a 23 games losestreak ((1-0.2)^23), bro is even lucky at inting
it's also very crucial to note that rng is always objectively bad game design and should never be implemented for any reason, no matter what. if a million people play your game, not one of them should have a single unique experience
True that.
Replayability is horrible, why give the player an opportunity to enjoy the game as much as possible for as long as they want when you can make another game for them to buy and enjoy again?
Your unique experience depends on what you do in the game. Not what some generator devices for you.
Although true output RNG is the worst type of all, it has 1 advantage over others. The emotions that generate when you get the crit you very well needed are absolutely priceless
X-com: This character has a 95% chance to hit! _Miss. Miss. Miss. Miss. Miss._
So called “+45 Sanity Sinners” After their 4 coin skill only flips tails
Statistically, that's a 1 in 3,200,000 chance!
@@Poke238s To be fair to Limbus, a single turn probably calls up the chance roll like 50 times more than a turn in XCOM, so that fail chance has way more opportunities to show up. That said, I have no excuse for whatever the hell XCOM is doing with point-blank misses...
That's funny since XCOM always rigs the chances in player's favor.
@@benmaks but it decides the outcome before you move and other thing to prevent save scumming
You see how many people make randomizer mods for games, where random items or events happen at random and (typically) incorrect times? That means that's what players want in a game, so your game should be like that, too. Don't worry that players don't have context for when these events are meant to happen. It's all part of the fun, and it's not going to confuse players at all.
1:17 fear and hunger:
Funger
The more random your game is, the more personalised their play experience should be, so therefore you should make everything random to deliver an experience that can cater to everyone
This one is really good because it keeps you guessing on whether this is actually good or bad game design.
( ͡° ͜ʖ ͡°)
If XCOM taught me anything, it's that a 90% chance to hit at point-blank range means jack shit.
I guess they were employing true random.
Didn't know about the distinction between pseudo and true random before this video, so I did learn something from this after all.
Thanks!
NOOOO YOU WHERN'T SUPPOSED TO LEARN ANYTHING! Oh, wait, that means I failed right? Maybe this is a good thing. :)
jokes on you in xcom (at least in enemy unknown) they only use true random on the highest difficulty
@@Artindi "How to fail at being a bad teacher"
XCOM uses true randomness, but on lower difficulties they have a constant multiplier to the displayed values (1.1 or 1.2), stacking aim bonuses for every miss in a row (but only if the base chance was at least 50%) and even MORE stacking aim bonuses if you have less than 4 characters alive. Plus equivalent debuffs for enemies.
@@tipoima The hidden bonuses for repeated misses are also capped at 95%.
But also none of this is quite enough to deal with people not properly understanding just how often that 10% miss chance will come up.
Unironicslly all jokes and satire aside, I ended up discovering this on my own. I was making a superboss, and thr rng of their attacks compared to their cooldown phase sometimes made it really frustrating to work with, so eventually I implemented a counter that increments based on how many attacks were used before forcing a cooldown, where the counter resets.
It made the boss go from annoyingly difficult to actually entertaining to replay while keeping its difficulty.
I honestly wished this video came out sooner because it took me a month to realize on my own.
Ah yes… the Valve method
Well, they make money. ;)
@@ArtindiDon’t worry the money also has random crits
I don't get the joke can you explain?
@@steluste Valve is notorious for their unbalanced and overused rng to the point that people started calling their ceo and founder, Gabe Newell, rnjesus, that’s why it’s the Valve method.
@@Raspingpython-10x can you give an example where they used rng? Cuz I don't remember anything random even though I played all hl portal left4dead cs games
I first learned of this kind of stuff from the Slay The Spire wiki. That game has a luck system where every non-rare card you get bumps up your chance of getting a rare card until you do get one!
so pity system
This sounds similar to a chain system that's used in a pokemon-like game i play. You KO multiple creatures of the same type, and each one that is normal contributes a point to boost your chain. The higher the chain, the higher your likelihood of catching a 'shiny'. And it resets one you do get one.
True RNG sounds like an absolute nightmare, and that's because it can be - but that frustration can do something interesting. In Fear & Hunger, the main gimmick of every action is a coin flip. Looting a container? Flip a coin! Dodging an instant-kill attack? Flip a coin! Sleeping? You better have a quarter in every pocket, because you are FLIPPING A COIN! The game is ansolutely brutal and sounds like a nightmare to play, but the coin flips further the game's central themes - that life is cruel but ultimately fair. There is no force altering the world against you... the world is just built like that. It's a really interesting case study on RNG in games, and I would highly suggest anyone looking into RNG at least give the game a peak. Just, please don't copy it one-for-one, I don't want to feel more pain!
I felt less pain in fear and hunger than i felt trying to farm for all the dev sets in terraria and killing 300 fucking destroyers AND STILL not getting that last set because rng
"Pipe be upon ya"
-demoman tf2
Gotta be honest, I will never design an indie (not because I don’t think I can, it just doesn’t interest me), but starting your videos with “the best way to fail is to never try” is genuinely some inspiring shit. Really rings true. Keep up the great work!
0:10 also known as Really Not Good 🗣️🗣️🗣️🔥🔥🔥🔥🔥
There was someone that made a challenge run of Pokemon where the RNG will always be against them. The opponent always crits and triggers side effects, while the player never does and will always miss if there's a chance to. Funnily enough, it's actually less frustrating because all of these become the expected results.
And when players complain about the RNG just point out that it's not technically random because computers aren't capable of actual random number generation
One must also adjust their glasses by pushing up in the middle while giving this explanation.
Wait what
@@jirachido3997 Computers are purely logic based because what a computer boils down to is a insane amount of true and false statements (1 and 0). So we cannot do something like lets say a dice throw because in the computers world, unless we change the way the dice is thrown, the dice will always fall the same way on the same side. So how do we give random numbers?
Most of the time its by using our clock. Either the program itself has a build in clock that counts the time since startup or they use your computers clock. Either way, time is a value that has a lot numbers (hours, minutes, seconds, miliseconds) and will almost always be different for any player engaging with a specific RNG event. So we might both start the game and immediatly throw a coin but I thrown the coin 5 miliseconds later than you did so our outcomes might be different.
@@diefontysstagiere5395 Or if you wanna have a blast, you hook the computer up to the random website which monitors atmospheric pressure changes for RNG
@@diefontysstagiere5395 adjust glasses
Actually quantum computers use qubits so they aren't just 1 and 0. so they should in theory be able to give us a more true random.
Brings back Terraria memories, where the game has a 0.1% chance of dropping something, but whether you actually reach a moment where you see it dropping throughout your entire life is 50%.
Fun fact, pseudo randomness was used in completely wrong context in this video. All simulated randomness on the computer is using pseudorandom algorithms. Computer can not give you "true" randomness. So what video calls true randomness is actually pseudo randomness. And this computer pseudo randomness can have interesting effects on the games. Like reloading same save state giving same "random" results, due to keeping same seed for random numbers, etc...
He's going deep! Let him cook! Someone hold his beer! 0.0
Many modern computers can do true randomness these days. From using sensor data of fans and temperature, to using quantum effects.
It's just much faster (and easier to debug) to use pseudorandom algorithms seeded with true random values.
@@tipoimaFalse. Computers cannot generate true randomness. In fact with the possible exception of quantum states, nothing in our universe is truly random
True, from a development POV, but everything here is from a game-design standpoint. In that sense, pseudorandom is more akin to curated, or rules-based random, vs just grabbing whatever Mathf.Random() gives you.
@@tipoima No, @david's right, that's just randomizing the seed, doesn't mean you're getting true-random output. But yeah, you're right that for game design & testing, working with fixed-seed (or at least saved) is the way to go, whether you're adjusting it afterwards or using it raw.
"If it's not 100% accurate it's 50% accurate"
Your explanation of the 4 kinds of randomness was so good i almost forgot it was a how to fail episode.
Everyone gangsta till someone makes a game accordingly to all the ‘how to fail’ videos
I actually expected you’d criticize pseudo randomness rather than true randomness. I thought you’d mention how a game might use a fixed list of "random" numbers, and that, over time, players could recognize patterns and even memorize the sequence if they played long enough. I vividly remember learning the RNG of some games such as the minigames of new super mario bros ds and the CPU's decisions in mario party games.
This is a good point actually, I considered talking about Stardew Valley's random system which can be controlled by players who understand it, so yes, pseudo-random can be damaging if not implemented well.
Same thing with the example I gave, players will purposefully hit something weak until they've reached the max number of hits before a guaranteed crit, then hit the stronger enemy with that starting advantage. Kind silly how far players will go for min-max a game. :)
if you can get a player to remember a list of numbers, you did great
In older games like the original DOOM, this was mostly a hardware limitation, since generating random numbers on the go was apparently too expensive for early CPUs, at least while also trying to run a game at the same time. id Software sort of got around this by having the game make random number calls for nearly everything, including enemy idle animations, so that the number you were going to get from the lookup table was always changing. The final effect was something reasonably close to "true" randomness.
On the other end of the spectrum is the Happy Lucky Lottery from Paper Mario: The Thousand Year Door, which is straight-up rigged. The winning numbers aren't randomly generated at all, the game simply checks how many IRL days have passed since you bought a ticket and awards prizes according. To add insult to injury, if you manually change the console's time and date, the NPC running this fake, rigged lottery will accuse YOU of cheating and have a big meltdown about it. The nerve of some people.
@@BarioIDL Unironically, yeah. In a single-player game, if people are dedicated enough to put together forums, compare notes, and write tracking spreadsheets, then congrats, you made a helluva game.
@@generalrubbish9513 I honestly had no idea DOOM used that method for its RNG. The date and time RNG feels like such a stupid move from the developers. And looking back, it’s funny how younger me could predict a game’s random number generator where it really showed just how flawed that method of pattern based RNG is. After a while, instead of feeling random, the game just becomes a predictable grind, where the developers expect you to be surprised by the randomness, but you just end up memorizing the pattern and going through the motions like a robot.
This first bit is effectively a critique of how Pokemon handles randomness with accuracy.
Focus Blast is an essential metagame move. It is 70% accurate but it does not feel like it is. It feels closer to 50%. It seems that alot of testing is necessary to find good ratios to make randomness feel fair.
Most extreme case of bad RNG you've seen?
The random critical hit mechanic in Team Fortress 2. You just… randomly deal triple the damage.
@@Soupnator its not bad its funny
hello neighbor's fear school level
Oh yeah, I remember hearing about that... lol
@@milkmilk6702 it is kinda BS that you die because the enemy happened to be luckier than you, especially in a multiplayer game
bro I love your pixel art style
Just to be clear that I hate when the game change your "luck" on hard instead of enemy's Intelligence
You should have done a section on input random just to say it’s only for rougelikes. Great video!
ha ha. yeah. I think I considered that while writing my script, but it never made it in. :)
The most randomly helpful random video I've every randomly stumbled upon.
I came here for a fun time and actually learnt something useful about gamedev while I was at it. Great vid artindi
NOOOO! You weren't supposed to learn anything! ARGH!
DND idea, the amulet of pseudosis, it gives them the ability to roll twice if the first roll failed horribly.
I'm not sure if the coin flip bit was because it actually fit into the rng bit, or bc RNG related stuff bothers Artindi so much that they found a way to give actual advice to try and stem the RNGesus bullshit in indie games.
Missed opportunity to make thus video 4 min and 11 second😔😔😔
I guess so, I'm not sure what that's a reference too though. :/
@@Artindi good for you😮💨😮💨
I guess they randomly picked it
I think a good way of makijg a balanced pseudo-random mechanic would be implementing the Gambler's Falacy in your game's RNG.
Basically, you make the chance of a event happening depends on the results from before, increassing in 1 the Power of the Denominator of the fraction representing the chance of it happenong in 1 for each time it didn't happened in a row.
Like, in this system, if the chance of sucess/failure is one half, but you suceded/failed the last time you did a Check, the chance of suced/faik decreasse to one fourth.
In D&D terms, is like you receive one cumulative Advantage for each failed Check in a row, and one cumulative Disadvantage for each Sucess in a row.
This is both less efficient and less effective than using mixed bag.
Example: player flips a coin, it can land on either side or balance perfectly on the edge... So you create 256 coin results... Two out of 256 will land perfectly on the edge, 127 will land on heads, 127 will land on tails. Use your rng to merely shuffle the bag.
Now you only need to prepare bags in bulk, making the rng code more efficient because it doesn't interleave with your other code, in fact the bigger the bag, the better the performance... Now player flips a coin, simply use the next value from the bag, very fast, simple, and fair. Player will always get 127 heads, 127 tails, and 2 edge cases per 256 flips. Sure, it becomes predictable, but player cannot game it, and in singleplayer game, let player have fun, so why worry?
And you can reuse the bags, don't remove any values, simply advance across the bag till the end, then reshuffle for reuse.
And no, memory isn't really a problem, even in a MMO if you do this, one shard that holds 2000 players will need only measly 500KiB (256 bools, one byte sized index to know where you are in the bag) to hold a bag for a player where ANY 50/50 chance with a small twist is required.
Sure, some more complex systems will need megabytes of memory, but all modern games eat gigabytes of RAM so megabytes is laughable. I can easily build a MMO backend meant to have plenty of monsters, bosses and random other crap using this technique and use under a gigabyte of RAM for one 2000 player shard, go figure. If really pressed for hardware resources, I can do it in C and allocate those 2000 player bags of all varieties statically and have memory usage constant at all times aswell. So it's simply amazing, and I'm not sure why most games don't do this.
@templeofdelusion Cool and well tought.
If you get too much of one result, it become less probable because you're increassing the number of options of this result and maintaining the number of ways of the other happen, doesn't it?
Thank you for the advice.
Roblox devs should take notes
I bet they use pseudo random, but not to benefit unlucky players, but to mess up lucky players
@Smiley_404 somewhat yeah, they have pity systems which is similar to pseudo random. But sometimes the rates are just unreal 💀
Ah so basically I should make the ending of my game based on a coin flip. It doesn't matter what you do, just get through the final dungeon and then a coin flip decides if there's even a final boss, and then more coin flips to determine what happens after you beat the boss (if it shows up)
Actually that's really funny, I shouldn't give form to intrusive thoughts...
Fun fact: Computers can't actually create true randomness so instead they just use algorithms to make a number that for all practical purposes is random but isn't actually random
What I remember from computers making "random" things is that they take some... "inspiration" from other things.
Basically, you could program your game to look at the date and time, then use *that* to make your "random" seed. You can also use stuff like player inputs, such as how long they were on the title screen for, or where their cursor is located.
Stardew Valley for example uses the player's step count to determine events such as weather or saloon stock.
This is just from what I remember about computer RNG, I may be *very* incorrect.
This is actually really interesting
No, it's supposed be unhelpful! >:(
Loved the video, but side note (a semantic one, I am guessing you researched this): the terms True Randomness and Pseudo randomness are terms already taken up by math. Usually they describe number sources that are provable to be random and not biased in any way (most computer generated random numbers are pseudo random just based off of computers being deterministic). True random sources are complex and most are resulted from quantum effects, like permisson of photons through a dichroic mirror. If we were to use a term to describe the manipulation of randomness that was used in this video, the correct labels would be skewed and unskewed.
I'd go more into depth about randomness but the medium of youtube comments is tough to use.
You are correct, in the context of math and computers the terms mean completely different things. In the context of game design, I've been seeing the same terms be used to mean what I described in the video. Human level perception of true random, and dev manipulated random. If there are better terms I'm sure people will start using them, this is just what I've seen be used to explain these concepts thus far, perhaps because game developers often have degrees in computer science.
The existence of the quantum human brain means you actually flipped a lot of coins making this video.
2:02 i actually cant tell if this is supposed to be bad advice because it actually sounds pretty good to me
The coin made me do it.
@@Artindi you said you didn't flip a coin >:(
was it a dice?
@@topminator2991 a dice with two faces
Great episode! I've never actually thought about using pseudo random this way, and thanks to your waring, I never will!
Yume Nikki was my bane of existence, the fact the poop hair effect is locked behind luck, because the location of the bed that teleports you to the only area connecting to that specific room, is... guess what! random! Sometimes it can be literally on your nose and sometimes you gotta do alot of stuff to reach it
The zippertile event also sucks since it's totally random
In my visual novel, i chosen to make RNG affected things not affect gameplay and instead just be silly jokes that have a 10% chance to appear
I don't mind things like the zippertile being completely random. Feels like something that would be amazing to discover on like your 4th playthrough of the game or something
@@foursidekm but it's true random tho and there's noway to manipulate the RNG to open it
@@FluffinWasntHere I'm fine with it for optional events personally speaking. It's more of a problem when it comes to effects. Since you actually need those to beat the game
well zippertile is supposed to be a secret, wouldnt be much of a secret if it wasnt totally random
and the bed thing isnt random since u can always check all the beds, wuts the big deal
@@yandere8888 it is random since the bed that teleports you to the hands world isn't the same everytime
It's too much of a hassle to get THAT specific effect, and that's ingrained into progression since you NEED poop hair to beat the game
I really liked this video in particular. Great job!
I really liked this comment in particular. Great job!
The goat has finally uploaded!
I'm not a farm animal! oh... wait. nevermind.
This reminded me of a thing I wanted to make in pretty much all games I'll make one day. The first one I'm planning to make is (yet another) Phigros-inspired rhythm game, and I wanted to do a little thing with the loading screen tips. Instead of just randomly being chosen, the game would keep track of certain things and pick more relevant tips on loading screens. For example, if you get bad scores too often, it'll tell you to take a break, or if you consistently hit notes too late or early, it would suggest you to change your offset. I don't think I've really seen that in games but it seems like a neat thing to do
Oh so like how doom eternal shows a tip of the enemy that killed you
@@THTB_lol A bit like that yes but a bit deeper
just one game...
just one game
Fear & Hunger
Artindi, you forgot about the best type of RNG ever!
Randomized characters! That's right, the best games all require you to gamble to get a chance of *maybe* getting the character you want!
And do remember to either design your game around a very specific ultra-rare character, you wouldn't want the uninvested player to enjoy the game.
Alternatively, design the game around no character at all, I'm sure nobody would be turned down with a story so bland it makes white bread feel spicy!
Ah, yes. True randomness without using quantum descomposition as source of randomness.
In the context of game design, it's true random. In the context of the universe, it sure ain't. ;)
Xcom2 uses a ball in bag model for random chance. When you fire, you pick a hit/miss ball that can't be picked again. So if you miss 5 times in a row, there just aren't gonna be any miss balls left in your bag
It doesn't matter that much, but usually the term true randomness has a different meaning.
Technically, true random numbers are the random numbers that are obtained from actually flipping dice, atmospheric noise, etc. Pseudo random numbers are obtained using a formula and are called that because they aren't actually random and can be predicted and reproduced, if you know the algorithm and the seed, and these are generated using System.Random or the like.
I flipped a coin on whether or not to like this video. I guess you earned it :D
Nice! :D
(Not being sarcastic)
Try to strive for Input Pseudo randomness as much as possible (at least when the player interacts with it directly, such as with weapons). A good example of randomness would be Slay the Spire (the game RNGs enemy attacks and lets you defend against them), and a bad one would be Xcom (you shoot an enemy, and only then see whether you missed or hit the enemy).
Also, if you're making a pokemon style enemy encounters, please make it so that encounters are more rare when you walk in straight paths and more ofthen when you walk in circles
That's a good idea with the encounters bit, I wonder how they could have fixed that issue in Skyrim, and interesting thought.
I thought randomness was pooled, like deck of cards, and non-pooled, like a dice.
Different categorization, can apply to all 4 quadrants described here. And it's more like independent vs dependent - if you draw a card from a deck, you've now changed the probabilities of drawing each of the remaining cards. While every roll of a die is independent from those that have come before.
That sounds like an interesting way to categorize randomness, very interesting. :)
@@Artindi That's what I thought. With pooled randomness you are guaranteed to get the percentages promised.
Well said .
So that’s why Fire Emblem is the way it is… 🧐
indeed. 🧐
Should charge money every time a player flips a coin
I’d say there’s another grouping of RNG:
Dice or Deck?
2 roles of the same die have no impact on the next
By drawing a card you change how likely you draw a different card
For an example of Dice RNG take Minecraft, killing a mob with specific drops; the chances of a mob dropping what you want don’t change by the number of that kind of mob you kill
For an example of Deck RNG take Tetris, the blocks you get; the game basically shuffles a deck of the 7 tetrominos, giving you the top one until it decides to shuffle another 7, this makes it so you won’t go too long without a piece you need
I MUST FLIP A COIN TO DECIDE IF I LEARNED ANYTHING HERE OR NOT!
2:38 this is true (citation: OSRS)
Please teach us how to fail at code readability and optimisation
Seems like True Random could also work well for games where frustration is part of the appeal and memorizing the progression steps is *not* part of it. So, not "I Wanna Be the Guy" (where you can master it), but maybe games like I Am Bread, Getting Over It, QWOP or the like? Not sure which ones would actually have randomness (or random tweaks) as a mechanic, but I could see them working in that type of game specifically.
*Heartstone anxiously smoking in the corner*
That joke at the end caused me physical pain :(
anyways, please make more videos (specifically ones on how to fail) :D
Funny enought, I'm currently attempting a cheese wich had a FIVE TO SEVEN layers of RNG.... and ...this...is...pain
I'm trying to defeat Jetragon at lvl 1 in palworld (a reminder than the cheese to make jetragon fall to its death has been long time patched).
It require :
- Rng roll 1 : Forcing spawn and despawn until ONE specific group spawn happen
- Rng roll 2 : Wait for Jetragon to go where you want him to be, the mildtest rng of alls but sometime it can make you wait some extra time.
- Rng roll 3 : After baiting the two to make them fight, hoping jetragon will catch fire (and no, the flame crossbow don't work with a large lvl difference).
- Rng roll 4 : For rng roll 3 , hope jetragon among all its attack don't choose the two attacks wich are almost guaranteed to nuke the whole group in one swoop before jetragon could start burning.
- Rng roll 5 : Can be mittigated but it can be very annoying, hope other wilds pals will not bother you by going too close and start aggroing you.
- extra rng roll 6 : To mitigate rng roll 3 , aggro jetragon at step 3, and hope jetragon will keep aggro on you rather than doing the worst scenario anyway and then comming after you.
- extra rng roll 7 : following step 6 You can put yourself at a place where Jetragon AI can't attack you or are is dodgeable, until his aggro reset because he can't reach you... except it has one heatseaking attacks it might decide to use anyway and you are likely to get hit by it.
Repeat step 1 and 3 again and again hoping rng roll 3 don't troll you to death. (and some minor rarer rng rolls too ).
This is hell, I though I've found a clever cheese strat but I accidentally created a RNG gauntlet.
Whoever came up with those terms did this specifically to annoy computer scientists.
It was real scientists, and yes, we did that to annoy people who will never be real scientists.
Mario Maker is actually genius. As a feature, the RNG is based on player input. Meaning, if you hold right and only right, you can get the exact same results each time.
Good job. If you mentioned pity, they might succeed.
Brawl Stars had a thing that boosted your odds of a rare pull if you kept failing enough. Still entirely RNG, but you're statistically gonna get something
1:05 FEAR AND HUNGER REFERENCE?!
Nevermind.
this video was definitely a video with puns that are very pun-y
One thing to note is that unlucky people will still be unlucky no matter how high you rig the chances, and lucky people will always be lucky no matter how high the enemy's odds. However, if you rig the system against the unlucky, they'll do even worse, and if you rig the system towards lucky people they'll do even better. This is always true, you can't make these groups meet in the middle, and even if you tried life would still find a way.
Good thing this has nothing to do with people.
Also it's trivial to equalize everyone, just use mixed bag randomness, that way noone can get too lucky or too unlucky, they will have things happen right on the droprate...
@@templeofdelusion Already thought of and failed, you will get what you need exactly when it's least useful and possibly detrimental. Impossible to equalize luck.
@@flameofthephoenix8395 then you're shit at problem solving because this IS a solved problem and I even gave you the name of the solution, idiot.
This is why I never use a move in Pokemon that has less than 100% accuracy.
I don't care if it's the strongest move of it's type, it WILL miss and it WILL get me killed at the worst possible moment.
I WANTED THIS ONE FOR SO LONG
Great video. It’s worth mentioning that definitions for true and pseudo randomness are different in other contexts including programming in general. However, in this closed context your definitions make sense
A classic failure of RNG implementation, specially in Tabletop RPGs, is to put the whole result behind RNG.
An exagerated example: You go to the king and try to persuade them for a higher reward. You fail. Now you didn't just fail on the extra reward, you now have no mission and are banished from the kingdom. Good job.
A much more typical example: The sword with the 50-50 chance of doing nothing or ALL damage.
Another example: You try to sneak past a guard but fumble stealth, now the whole base is alerted. The better solution is to make the guard alerted to *something happening where you are* and acting accordingly - investigating, calling for backup, etc. But no full-blown alert from the get-go, the guard doesn't even know what's in there (unless it's a very high security area).
The key is in respecting the player's decisions. The player isn't rolling dice to see if their plan happens, they're rolling dice to see if their actual action succeeds, to see if they can avoid alerting this guard, not the whole base.
Otherwise it's the failure equivalent of a player going "I roll to infiltrate the big bad's lair."
Great point! Perhaps the best random is the random you can't really feel at all. :)
In Warframe each character requires grinding for the different parts from different missions. One of the more egregious ones was Octavia. Her kit completely trivializes the game, but her neruoptics only have a 1 in 5 chance to drop once every 20 minutes in an annoying survival mission. I played for almost 2000 hours and never saw her neuoptics.
I had to flip a coin to decide wether to give a like to the video or not.
I'll take those chances!
This is why I removed randomness from my Fire Emblem fangame.
Not to be confused with pseudo-random number generators, which spew out a deterministic list of seemingly random numbers (and is most likely what you are using when you go the "true" random route). Computers are actually really bad at flipping coins. I think it's the lack of thumbs.
Yes, in actuality, (unless perhaps you have a quantum computer,) all the "random numbers" are pseudorandom. So we are using the terms "true" and "pseudo" a bit abstractly here for the sake of understanding game design. :)
The best way to implement pseudo-random RNG is to punish your players when they start doing well. Everyone loves the blue shells in Mario Kart, especially when they hit you on the last lap when your only a car's length from crossing the finish line.
Remember, if 100 people toss a coin 10 times, 10 of them on average will get 10 heads.
So, if your player base is large enough, you will have a significant number of players getting a 10% Bad Event happen back to back to back, and never land the 90% Good Event.
Trivial to fix.
The Yo-Kai Watch this episode took this episode to heart.
Every coin is technically a d3, just with the 3 having slightly less chance to be rolled than 2 or 1
The people who designed Aegis Cave watched this video and decided that they should make a level where output true randomness should be essential for progression. They should have put more levels like Aegis Cave into the game.
There might be a 1 in a 8.1 billion chance that Artindi might make another episode. MIGHT! (pseudo-random)
Those are the numbers I'll tell people to make then feel special when another episode drops. ;)
Regarding loot drops, its nice to have a way to trade in lots of bad items for a good item. It can be a thousand to one, just give me a means to actively obtain the good items with all my bad drops. Not a secondary currency which you never use; something which can n be used to make progress in the getting the best item drop.
That's a good idea, if you have them, provide ways for the player to use it in a way they chose. :)
Reminds me of Bad Piggies in some way where you can get scraps for getting skins you already obtained, and those scraps can be exchanged for better skins
At that point, the bad-drops become a secondary currency. Your time playing is also a fixed-resource.
so if you look past the sarcasm its trying to say output rng is bad and given input rng wasnt even mentioned after its introduction at the start that its good
kinda makes sense as world generation a possible use of input rng gives every player a more unique experience, while say lootboxes or even crits, an output rng make the player rely on luck which overall just doesnt feel great even when you get lucky because at any moment your luck could run out and its over
either way it seems more fun to both make and play a game with stuff like skill based crits instead of chance based ones, aiming for weakspots, or maybe a timed minigame sorta thing depending on your game
Can you make a video about Rhythm Games?
The interesting thing about online roulettes that they are pseudo-random. Like I constantly get sequences of 10 red in a row, which only has odds of 0.24% happening.
Love the new art
Would you ever do a video on specific indie games? Like how to fail *fill in game name* ?
Like, how to fail at playing them? or how to fail at making that specific game and I just do sick burns the whole time? Both could be fun.
how to fail at making that specific game, and just do sick burns the whole time. lol