Settlers of Catan Placements Study: Results From 754 Games
![Treeckosaurus [Catan]](http://yt3.ggpht.com/ytc/AIdro_nRVKXeGFBefgIvY06ORFQmO0c1cs3nZ5tcP_kSXaGz8Iw=s48-c-k-c0x00ffffff-no-rj)
ฝัง
- เผยแพร่เมื่อ 24 ก.ค. 2024
- I analyzed 754 Settlers of Catan games to better understand what strategies win the most games.
Specifically, I looked at advanced metrics like:
- Number Diversity
- Scarcity of Resources
- The Value of Ore
- Placement Order
And in this video I’m going to share what I discovered.
The Settlers of Catan tips and tricks you'll learn in this video can be applied in every single game you play. These are actionable rules of thumb and strategies I use to have a Grandmaster-status on Catan Universe and to beat my friends and family consistently in a game of Catan.
If you're looking for a channel where Catan secrets and strategies are shared and you can learn about the game, you've come to the right one! I upload online championship videos with insightful commentaries and in-depth analysis of 'every game'-situations, like initial settlement placements.
Also, if you want to apply the things you've learned. I host an ongoing Catan Universe / Colonist.io / VR Catan Tournament called King of Catan, in which you can play against excellent catan players from all over the world. Sign-up is still open and you can do that in less than a minute.
► Check out my Patreon: / treeckosaurus
► Check out Bucks Board Games and use the promo code LongRoad: teespring.com/stores/bucks-bo...
► Sign-up for King of Catan: bit.ly/koc-sign-up
► Short video that explains how King of Catan works: • How Does King of Catan...
► Full KoC rules expained: bit.ly/KoC_Season12
► Play online on colonist.io/
► Upcoming catan tournaments: kingofcatan.net/catan-tournam...
Thanks for watching the video and/or reading the description!
I reply to every comment, so come at me :) - เกม
Truly amazing video! Thank you for all the research and interesting facts you were able to find! We appreciate it! Thanks Treeck.
I'm glad you enjoyed it, DylLighted!
If Dylighted gives the thumbs up you know it is quality content!
I love that your shirt at the beginning just says 'NO WHEAT.' I've always thought it was an overrated resource myself
The full text is: "no wheat equals defeat" actually ;) I think the text itself is just a meme, but I think wheat is the second most valuable resource.
@@Treeckosaurus Second-most to which one? Ore?
Game design is so interesting to me and I appreciate it when people do analysis to put data behind the theories... Fascinating
Solid video Treeck! Some great information in here. Thanks for all your work on putting this together.
Thanks Robert! I'm glad you enjoyed it :)
Very well done. It was very reassuring how closely alligned my intuition is with your results. Thank you!
Thank you and you're welcome!
This is a fantastic video. Well done man.
What a great video! So many interesting points. Really cool stuff!
I'm glad you like it Retired Hero!
Nice effort, but technically speaking counting the number of wins does not constitute a statiscal analysis...I’d love to see confidence intervals, p-values and the like and unbiased results that control for the effect of resource prevalence on the number of wins (for instance, it’s natural for OWS to win more on high ore boards as more people will tend to go for that strategy, and so on). The way most of the numbers were analysed allows for potentially significant biases. Anyways, cool analysis for getting a rough idea of dominant strategies, but hardly conclusive.
Thanks for your feedback Jun. I agree with your comments on the technical aspect of how the research was done. I think for a better research, I need input from all four players.
@@Treeckosaurus What you really need are t-tests and ANOVA analyses! I'd love to see those tests for significance, or a categorical output modelling test of some kind!
Damn man give him a break. He put a lot of effort into this. Constructive criticism is good but I suggest you put your money where your mouth is instead of just criticizing others when you haven't put in any of the work. Treeck didn't have to release any of this info that is the result of his hard work. @Treeck, thanks and appreciate you man!
@@Kickenwing360 Given the raw data, I would happily put some of those analyses together! Presenting raw percent values as comparisons without actually testing for differences can lead to spurious conclusions. I absolutely appreciate the effort that took place here, but we need to be careful about what we define as informative statistical outcomes!
@@Kickenwing360 so you can only give constructive feedback when you redo the research yourself? I think you need to reread the comments including your own. The guys above pointed out a relevant aspect lacking in the research, but they never rejected treeck or his research as a whole. I think they did it in a respectful way as well and, as you said yourself, constructive feedback is always good. Besides, Treeck always actively asks for (constructive) feedback, as he knows it can improve his videos. Kudos for that mindset.
Fantastic video- thanks for taking the time to record all this data
I'm glad you like it! It was def one of the videos that took the most work to produce, but also one of the highlights of this channel
Great work Treek! Love this video. Great work!!
I'm glad you enjoyed it Justin!
Wow!!! Nice research. That must have taken a lot of work. Thank you
It did, glad you enjoy it!
Thank you for the great insights!
You have also mentioned the disclaimer in your video but it's really important for interpretation, that it is the deta of the winning player, not from all four players. Maybe in a Catan game with no rare ore or wheat there are in average 3 players playing an OWS hybrid or full ORW strategy. It's very likely that one of these wins although your data is not showing that also 2 out of 3 loose with that same strategy.
Exactly, Florian.
Amazing insight from a super strong player
Thanks Alley! I'm glad you like it :)
Thanks for this!
I wish you showed the percentage of people using a certain strategy who win, instead of showing the percent of winners who use certain strategies. This way, we don’t know how successful they are, because it could just be that more people used those strategies (see the OWS with lots of ore, for example)
Wasn't possible at the time. Working on it for future projects though. Glad you enjoyed!
Great video, keep up the content 👏
Thanks Darragh!
What's OWS vs Hybrid OWS?
OWS: A strategy focusing on getting cities and development cards efficiently, by placing on high ore and wheat and decent sheep.
There are variations: Hybrid and Full.
Hybrid OWS also gains a little bit of wood or brick.
Full OWS only gains ore, wheat and sheep at the start of the game.
Thank you!!!! That seems obvious now. I'm a newbie at the game and love it so was trying to learn some strategy.
Great! I will go more in-depth on the strategy soon
@@Treeckosaurus Awesome. I just subscribed so i won't miss it.
Great video! Would love to see this done for C&K, as the placement is a LOT different!
Thanks! Me too. Unfortunately, I don't have enough data for C&K at the moment
Great results. I suggest to do a modelling with all these results. I am sure we will get some interesting insights. Also it will be fun!
Thanks! Yeah, that would be interesting
Would be nice if you would link to a video giving full explanation of your "hybrid" strategy
Will do!
I loved the video. I saw this video for the first time two years ago when I was just starting my Ph.D. in data science. Now I'm almost done with it and preparing for the national Catan tournament. I came back and found that many conclusions are misleading (I'd love to discuss in detail). Can we have access to the dataset?
Hey Fred! Awesome to see you've done well with your Ph.D. and you enjoying the video. I'm aware of the fact that the technical aspect of the research are poor, I'm keen to know your thoughts on why you label the conclusions as "misleading"? Feels like a very strong word to me.
Great video treeck, loved the analysis. From the 754 games you analyzed, how many of those did u have a picture of the board?
Thanks Austin! I have them for basically all, but I didn't check with all these boards as it isn't time-efficient.
Man, top level Catan players (at least the youtubers) are also good at academics. Dyllighted is into software development, Treeckasaurous' statistical insights are also brilliant and more importantly he understands where the results are inconclusive or data cannot be taken at a face value. Maybe if someone collects enough data, my hypothesis about Catan players would not be rejected :)
I appreciate your words Tushar! I didn't know DylLighted is into software development. What is your hypothesis about Catan players?
Hahaha I'm not actually into software development, but I appreciate the comment! I'm studying Biology at College though!
Firstly, happy to have both of your responding. It's a pleasure. Somehow I got the gist about Dyllighted being in software when he complained about Catan Universe in one of the videos. :) @Treeckosaurus Hypothesis is that good Catan players are good at academics too. Bdw your username is difficult to spell man
Haha okay, even youtube has trouble with it: "trakosaurus catan"... I think I'd assume your hypothesis is a very strong one.
Great video! I've read in the comment in one of your replies that you don't share your dataset. Can you at least show in the video or other videos how did you collect the data from different results, the different sample sizes (HOWS, OWS, Road...), etc, so that it will add more context to the results?
You could also account for the placements and the dice distribution to account for luck. (might be helpful for the results of luck or skill)
From the results you have shown, I think you can still do more. a LOT. A lot in the descriptive part and also modeling.
Really great video. It can also lead to other researches, or encourage other researchers. It would be nice if we have a Github repository for these data and results, so that others can contribute.
Really great content. Thank you!
Hey Dave!
Thanks for your comment and feedback!
Right now, I'll keep the data to myself. This might change, because I'll do some more data research videos in the upcoming months with new data.
I will have some time to think on what I want to do with this data-set.
I'm glad you enjoyed the content :)
I am surprised about the last fact in cities and knights. I usually play it and thought, same as you, 4th is the best position. What should I do if I have no comodities open for my city placement at the start? What is my win condition? recommendations?
Great videos!
Thanks!
As fourth? You could go with high brick + brick port or something that makes it so you can get a city on your first spot fast. Win-condition can be road. But yea, Cities are a must-have (99% I'd guess) to win games.
Interesting I've tried port and 5 resource strat. I noticed that the 5 resource does better since no one focuses you. Played a few ows and seems the most promising.
I assume you're referring to online gameplay (like auto-matches on Catan Universe). My experience is: port-strategies get targeted heavily by casual players and are therefore a no-go unfortunately.
Awesome research, Henk!
Thanks!
@@Treeckosaurus I am Bonaventurianus on the KoC discord server😉🤙
One of the things I would like to know if the strategy patterns and percentages change for C&K.
Hey Arturo! I don't have enough data for C&K to give you good answers. Maybe Babilarian can anwer your questions better.
love the videos on the channel. I like that you are starting to analyze games with a lot of data. I am just skeptical of how useful this data is.... I don't think the strategies listed are mutually exclusive. you could have a road building strategy that relies on production and ports, for example. Also having people self report why they won seems fishy.
what i would like to see is some sort of model that can guess at everyone's win probability given a game state (every good player has a mental model like this). The model seems very hard to make though. Any ideas on how such a model would work?
I'm glad you enjoy the videos! I share your skeptics towards the technical aspect of how this research is done.
I think the model would be based on the amount of resource needed until a player has 10 points.
- odds of drawing a VP
- resources needed to get a certain win-condition (road, army)
- total resource gain
Combined with the same stats for all other players.
Would be really difficult to make, I think
@@Treeckosaurus Your model would be (Final Rank: 1/2/3/4) ~ Intial_Production$ore + Initial_Production$wheat + Initial_Production$sheep + Initial_Production$wood + Initial_Production$brick +Diversity(1/2/3/4/5/6 different numbers)
You could also run a logistic model asking whether a player wins/not but I think that model would be less informative. Of course, you could include a ore*wheat*sheep interaction term and a wood*lumber term as well (to account for the roadbuilding/OWS strategies), but it would be up to you! Use deltaAIC to compare your model parameters and bam: we could know more definitively how initial placement influences success.
If you wanted to get crazier, your parameters could also account for the proportion of the board's production you settled on for each resource. Perhaps instead of how many points of production you settled on, you instead measure whether a player settled on a high or low proportion of the boards available production for each resource. That strategy would tell us whether OWS/Roadbuilding etc. are more effective when the required resources are in abundance or not. You could also add a port term that reflects how many points of production they had relative to the ports they settled on in initial placement.
Thanks Evan! If I ever do another research for Catan, do you mind if I contact you? You clearly know a lot more about doing research than I do.
@@Treeckosaurus this guy definitely gets it, would recommend. however, I think there is indeed a way to circumvent most of the critique: more data. while this won't get rid of any of the 'conceptual' issues (lack of further parameterization, inherent luck-based aspects) I think it will greatly lessen their influence. I don't know enough statistics to calculate when that would be accurate (it would be rather hard since the distribution of dice rolls in a single game rarely will rarely conform to the normal distribution since the number of throws is too low), but, my gut just says that 700 some-odd games is too little. (though, it's impressive already!). I would be interested mainly in just plain more games/data
What do "All Games". 4-7, 8-11, 12-15 columns mean on your low/medium/ore? I don't believe that was explained clearly? Is that total # of pips for ore across all tiles? It looks like it may be worth targeting a five resource strategy rather than hybrid if total ore is < 11? Would also be curious to see this for low wheat.
Yes. I've asked the players how many pips of the ore there on the board in total.
Don't have data for wheat unfortunately.
Man Treeck this is pretty cool. It helps confirm what people probably already suspected, going first and hybrid OWS is the best; but only if ore production and the desert are in their proper proportions. Interesting road building won more games even if it was barely. I suppose that is because it’s easier to swipe and win in a big final turn, while the winner can’t snag large army because third or fourth place started on it earlier or the winner just wasn’t able to buy knights for some odd reason. I was really shocked to see how having three unique numbers was big for hybrid ows and five resources; suppose feast/famine and their numbers hit.
So does this mean we should play the odds and basically do:
Five numbers with
Hybrid OWS or five resources and
army/road respectively
And switch between the strategies,
depending on table position/ore production/and desert placement?
Hey Robert, it also depends what your personal skills are. If you're really strong at trading, playing with an imbalance in your resources is probably better than playing with all five resources. I would always make your decisions board dependent.
Army / Road: yes, I recommend to choose one of the two at the start (before doing placements).
Alternating strategies is a good idea. I think inviting yourself the challenge to play with one strategy for a while is also a good idea to conduct a research yourself.
- What does work?
- What doesn't?
- What perceive opponents to be too strong?
- What is just the right balance?
Yes, always board dependent.
With weak sheep/grain road building strategy is a decent strategy. I've noticed most games I still achieve longest road with a hybrid OWS (settling at a port as soon as possible is big though)
I believe "lebensraum" is key to win. The player with the best 3rd village wins. But I'm not sure if it needs to be a 3 hex production place or a 2 hex and 1 port (preferable 3:1 port).
3rd settlement spots are indeed important, but I think games can be won in a lot of ways. If your opponents believe they should always target the player with the best third settlement, it becomes difficult.
there is something like reverse- causality. So win % jumping from 1 to 2 VPs can also be because to get 2VPs statistically that player will have bought a lot of VPs. Which means those players statistically had good production, prob. cities and perhaps largest army. Might have pulled Monos etc. So it would be interesting to see how many Dev Cards pulled wins.
Yeah, good point. Didn't know it is called 'reverse-causality'.
I think it is called “sample bias”
@@kaushikdr it has many names, "confusing the common cause" is another one
Impact of number diversity. The Problem with your winrate could be that the possibility for the diversity is not the same. Having 6 numbers implicates many things. Sitting one 6 different hex spots with 2 3hex settlements. I guess the probability for that is very low. I mean it is much more likely to start with 4 or 5 numbers. But anyway your results show some stats even without the reasons behind this.
Except of that nice stats collection. Good work
Thanks Sime! I agree with your calls on the technical aspect of this research.
Great Video! one thing I wanted to point out is even though you have 754 games worth of data I still think there is a lot of variance in these results. If you look at the scarcity table there are a few massive differences in the results of wood and brick. In theory these results should be about the same because wood and brick are used for the exact same thing (you need one of each to build a road and settlement). But there is as much as a 14 percentage point difference in five resource and a 8 point difference in road building. Despite that though I still feel like these results hold merit but need much more data to eliminate some of the variance.
Hey Chris! Glad you enjoyed it. I agree with your comments, it's why I had to -this- to some interesting topics within the research.
It's worth to remember thou that there are four wood fields, while only three brick fields. It kind of gives some sense to the results for road building and rare resources, where getting the bricks would make you less likely to be blocked and more likely to make good trades, respectively. Still, doesn't explain the difference for five resources.
Cool video I was wondering if the data you had was available to the public or if this was just a personal dataset that you made?
The data-set is currently not available to the public, as I'm still doing research with it
Any chance to get the data set used for this research?
No, sorry.
Do you have any data on with which colour they played?
No. I wish I had that data, would be super interesting to know if it has any impact
Hey, where did you get the dataset from? Where could I get it?
I host an online catan league called King of Catan. The participants always fill out who won the game and who they played against. I've also asked some extra questions players could fill out voluntarily, that's where I got the data set from.
what is production points? and how do you calculate them?
A 6 and 8 = 5 production points (pp)
5/9 = 4 pp
4/10 = 3 pp
3/11 = 2 pp
2/12 = 1 pp
This system is based on the odds of a number rolling. Per 36 dice rolls, a 6 rolls 5 times on average. And a 9 rolls 4 times. 4/36 = 1/9 = 11.1% every dice roll to roll a 9.
Does it answer your questions?
What exactly is the 5 resource strategy? Just, everything? Thanks (:
It's focused on having balance in the resource gain, so porting and trading isn't necessarily needed.
Is there an original data set that you derived these results from? If so, can you share it?
There is, but I won't share it. I'm sorry.
Nice data you got there, be a shame if they, rolled an abnormal distribution😌
Yea that happens from time to time
Hybrid OWS dominates in high ore boards (12-15 points of Ore). But does about the same as a basic 5 resources strategy
Hybrid OWS does better with fewer dice numbers at start of game (3-4 rather than 5-6)
I can vouch for Mono being a way to win cuz I got shafted at 9 vps by a player with about 5 when he played two monos two turns in a row. If not for that I would've made a city and won. I'm salty still.
Great points. Thank you for your insights!
Good video, lets have an update with 5k+ games ( professional matches )
Sounds good! I'm work on it already :)
Cant do statistics like this and diregard standard deviations.
A lot of this is debatable. More clear data and regressions are necessary.
Agreed, but I still think it's valuable
What is Hybrid OWS? Good wheat and ore and port fuel for sheep?
@Woolybulli It's a strong Ore, Wheat, Sheep number with an addition either wood or brick number. Full OWS means you are going no wood or brick and going all OWS. Hybrid is the large portion of your numbers are OWS, but you either have a brick or wood tile to assist in getting your next settlement down.
There are two OWS strategies:
Hybrid: Ore, wheat, sheep and something else (brick or wood)
Full: only ore, wheat and sheep
What is your order of importance on the resources?
Is it
1) Wheat
2) Ore
3) Clay
4) Sheep
5) Wood?
Great question!
I'd change up ore and wheat, because there's ore on each board than there's wheat (3 vs 4 hexes).
And it will depend per board. If there's only one player who has wood in the game, it can make wood the most important resource.
Hey Treeckosaurus. I was led here from the Colonist.io Discord.
I'm a college student that is interested in performing some data analysis of my own on a large set of game data. Are you open to sharing your game data, if you still have it?
Hey Jorge, I don't share the data set.
The most interesting to me was that second position usually won the game
What position did you expect?
What’s a hybrid OWS strategy?
OWS: A strategy focusing on getting cities and development cards efficiently, by placing on high ore and wheat and decent sheep.
There are variations: Hybrid and Full.
Hybrid OWS also gains a little bit of wood or brick.
Full OWS only gains ore, wheat and sheep at the start of the game.
My friends and I use Colonist to play online and we have found these Two things:
1. If you place a robber on a number it some how triggers that number to be rolled more frequently. if the Robber is on a 6 wood it will trigger the 6 wheat to be rolled a bunch of times.
2. Sometimes numbers like 5 or 9 will never get rolled after over 100 rolls of dice. So this idea that their dice are actually RNG seems bogus.
Hey Andrew!
1. Yes, it seems like the same for Catan Universe. I've never done valid research for it, but my hypothesis would be that it's true.
2. Well, if you look at dice odds. A game without a 5 or 9 should happen once every 167 games. I haven't done the calculations myself, but someone with a lot more math skills has: cs.stanford.edu/people/nick/settlers/DiceOddsSettlers.html
Can you provide your data? Would love to use it for a school project.
I'm sorry, but I won't. I might do it in the future, but not right now.
@@Treeckosaurus no problem! Was worth a shot.
Hi there, i dont want to be harsh, but this data is not very useful. The key metric you wanna be looking at is win% for each strategy, and you wont get that unless you ask for data regardless of the position the player ends in.
Nevertheless, it is really cool seing people investing so much in the game, keep up the good work!
P.D: If you need further clarification, feel free to contact me, i´d be glad to help.
Hey Alvaro, that's unfortunate it wasn't helpful for you. If there's ever a round 2 of data research I will ask for the strategies of the losing players as well.
@@Treeckosaurus I am looking forward to it! ;)
I usually employ what I call the high production (HP) strategy: high ore and wheat combined with decent and balanced wood and brick, ignoring sheep altogether.
This gives me fast city building capabilities as well as decently paced FLEXIBLE expansion (if you're getting brick and wood you build a settlement, if ore and wheat you city up), focusing solely on upping my production numbers and ignoring dev cards and LR altogether at least until I reach 6-8 production points. Constructing triple ore and double wheat position is also important as it rapidly quickens your city building. Sure, blocking and robbing is annoying, but at the end of the day it's just one square and with a strong enough production you can ultimately power through. More or less the only thing I trade for are sheep, which are usually in common supply as people tend to overload on sheep.
Also, it has been my experience from early on that the obsession with having 6 different starting numbers is highly misguided. In some cases you actually WANT coordinated numbers, like wood and brick, giving you a free road on each roll of the number or double ore/wheat drastically lessening the time needed to city up.
Yea, that's a great strategy. Especially on boards with strong sheep (easier to get the resources you're missing with trading/robbing). Your strategy didn't get much love during this video, but I agree with you it's very good.
To not combine the percentages for Hybrid OWS vs Full OWS makes it seem like perhaps an OWS strategy isn’t the best option to win the game.
It’s OWS or the odds aren’t in your favor.
I see them as two different strategies. I agree, and you can make the same case for 5 Resource Strategy and Road-Builder.
Thanks for the hard work, but it's really hard to follow all this raw data. You should try explaining the key points ans conclusions. You could watch Spirit of the Law, for inspiration. He analyzes aoe2. He crunches numbers and presents conclusions in a friendly manner.
Sounds like a type of channel I'd watch yeah, will check it out!
What is Hybrid ows?
OWS is the strategy where you have strong Ore, Wheat and Sheep. The aim is to get cities easily, as well as development cards to win the Largest Army.
There are two variants: Full OWS (you only have OWS) and Hybrid OWS (where you also have a bit of wood or brick to make the necessary road+settlement a bit easier to pull off)
what is hybrid OWS?
OWS: A strategy focusing on getting cities and development cards efficiently, by placing on high ore and wheat and decent sheep.
There are variations: Hybrid and Full.
Hybrid OWS also gains a little bit of wood or brick.
Full OWS only gains ore, wheat and sheep at the start of the game.
what is hybrid ows?
OWS: A strategy focusing on getting cities and development cards efficiently, by placing on high ore and wheat and decent sheep.
There are variations: Hybrid and Full.
Hybrid OWS also gains a little bit of wood or brick.
Full OWS only gains ore, wheat and sheep at the start of the game.
thanks
What does OWS mean
Ore Wheat Sheep
What exactly is hybrid OWS?
Ore wheat sheep + either brick or wheat on your starting placement.
OWS: A strategy focusing on getting cities and development cards efficiently, by placing on high ore and wheat and decent sheep.
There are variations: Hybrid and Full.
Hybrid OWS also gains a little bit of wood or brick.
Full OWS only gains ore, wheat and sheep at the start of the game.
Hi Treeck,
Can we work together on your data in Catan? I have a research proposal for Catan about developing new strategies that might help, but I really need more data. Are you open for collaboration?
Hey Dave, yes I'm willing to discuss! My mail = treeckosaurus@gmail.com
@@Treeckosaurus Didn’t notice this. TH-cam notifications off. Just sent an email. Thanks, Treeck!
How come you can't fix Catan Universe?
I don't work for Catan Universe
Just from observation, as your percentages are in many cases very close compared to the number of games (observe; sqrt(625) = 25 ==> 1/25 = 4%, I doubt you in many cases even have a stasticially significant result, as the 'test sentitivity' will be on the order of 4%. Look into how to do binary statistic tests, like the Fischer/Chi tests, before drawing any conclusions. For alpha = 5% and 1 class; example; win rate is more than 25 percent, it's N * (r - 0.25)^2/0.25 > 3.84. 0.27 does not meet that bar, and only gets you a statistic of 1.16, or confidence of about 0.3. 30% of the time, your conclusion is wrong.
Second; you didn't draw any conclusion about using clumped versus spread out numbers, because what you need to compare between the two are win rates, not absolute number of wins, as the ability to settle on multiples of the same number is determined not just by player choice, but also by the board's randomization and other player's decisions, which should be corrected for by including wins and losses in the data.
I appreciate your level of detail in your answer, as it's something I'm trying to learn more about.
Good WORK Trek! KOC WSOC!
2nd performs the best!?!?!.
That opinion of online players isnt accurate.. You dont know VR players. 3 world champions.. 3 national qualifier.. Catan VR isnt in this study with that.
Catan VR is the premiere league.
(Discord) catanVR.CLUB
Thanks Kerr!
Why is it that when calculating number diversity, you chose a completely useless metric, which is the number of numbers someone placed on? The only metric that matters is how many unique pips of production someone is on. Think of it as there actually being thirty different numbers. A guy who placed on 9/6/10 8/4/10 is on 20 different numbers. Saying "5 numbers" doesn't matter at all since a 12 holds only one of the 30 production slots and an 8 holds 5 of them. An 8 counts as 5 times as many numbers as a 12, so your analysis of number diversity using this metric is completely worthless.
Personally, I'm quite interested in this factor, since I'm curious if number diversity actually helps win games, but your figures are completely useless for determining that because they hold almost no meaning.
Je kan je Nederlandse accent niet verbergen haha
Klopt... I mean true
@@Treeckosaurus Achteraf zie ik dat mijn comment gemeen over kan komen... geen verkeerde bedoelingen, het was een goede video!
Dankjewel! Blij je 'm goed vindt
too bad it doesn't have access to losers
Yea, that would have been better. It would have given a better data size, but it's valuable nevertheless - in my biased opinion.