ELO points



  • How is currently ELO points calculated in the end of the game?

    I noticed that I get different points for the same places? Does this depend on point different or something?



  • In chess it depends on your points and your opponents' points.



  • :D

    Would be nice to know how it is calculated in catan as well :P



  • Today I had biggest and fastest win (kudos to both players who didn't quit and did battle for 2nd) but not about that, as you can see from screenshot (https://postimg.org/image/a4lvslf6p/) pretty much I won a game by 14 points and almost flawless victory.

    Got +7 ELO points for that.

    My question is how fair current system is if you get +7 by allowing 1 player build 1 settlement in the WHOLE game, while you get at least -5 ELO for being last?!

    If I remember right I once was 3rd just because winner of the game decided to play against me (taking road and 2 points) instead of building settlement and getting 1 neeed point for the win. In that game I got -7 or even 8 ELO. Difference between winner 2nd and me was -3 and -4 points.

    Does not sound fair really.



  • I think your ELO is so much higher than the other players ELOs. ELO never takes into account what happens IN the game, just what your ELO was before the game. This is how the calculations go.

    Take Chess for instance. If Magnus Carlsen plays with a 5 year old and wins in 3 moves, he probably gets 1 ELO. But if that kid manages to win against Carlsen, he might get something like 200 ELO because his ELO was 1200 in the beginning.



  • Good to know if that's indeed the case.



  • Here is one example: lost because of "unusual" roll - 3 and 4 just rolled the *hit and blue player just took off with no return.

    In the end -9 ELO points vs. flawless victory and just +7...

    Like really what here would be fair and encouraging @Administrator ???

    https://postimg.org/image/d8g8lr995/



  • If you can't handle the randomness of the dice, maybe Catan is not the game for you. There are a lot of board games which you could also play online - Carcassonne, Ticket to Ride. Check out BoardgameArena.com - even more good games. Tzolk'in is one of my favorites there.

    Why would you even care about ELO anyway. It is a horrible representation of skill, especially in a game like Catan where 3 players can easily team up against the 4th and always make him lose.



  • @Stroom I got used to the randomness of the dice as you can see and did play this one till the end. My point here is more focused on the fact what should be rewarded - skill or random guess (aided by unusual roll patter)?

    As for ELO and all other reported issues by me - wouldn't call it "caring" rather than my investment in this program development and effort to make CU a better place.

    Finally ELO points seems to be a benchmark in this game and quite often reason why people quit so maybe fixing such an easy thing as mentioned in this topic we all might gain a positive change in community (even if minor).



  • ELO only calculates wins or losses. It should not matter how you won. If you want some other kind of ranking, stop calling it ELO and offer something else. I think there is no need for ELO anyway. Catan is not a game of skill. It is a statistics&probability game with social interaction. Very much like poker. You play with the hand you are dealt with. There is no ELO in poker as far as I know. And as I said before, ranking makes no sense in a game where you can not win all games with pure skill. If luck is such a huge factor, having ELO is pointless.
    Sometimes you win with really good rolls and other players have 4 points while you have 10. Were you a really good player or just got the good resources early? Sometimes the game ends with all players at 9 points and eventually one of the players gets the 10th point. One player got the lucky resources at the right time. But any of the players could have had that chance. There is no way an algorithm could say if a win was "lucky" or "pure skill" (at least the one these developers would make considering they botched the longest road calculation for so long).

    You are constantly reporting that the dice are not rolling by the perfect distribution. This is nonsense in this case because out of 70 rolls per game, you almost never will have each number rolling equally. Keep calm and carry on. This is Catan. It is not a serious game. 30% of it is luck. Got unlucky in one game? Play 3 more games. Eventually the statistics will be in your favor as well. Over hundreds of games and dice rolls you can see how good you are compared to random opponents.

    I do not think people quit because they do not want to lose ELO .They just do not care to stay and finish the game when they know they are not going to win. This is completely different issue the developers have to fix. I have posted in other topics about this, not going to repeat that.



  • I must totally disagree with you Stroom.

    First, chess is the ultimate game for a rating system because there is no random element in the game, we certainly both agree with this.

    Second, Catan IS A GAME OF SKILLS, being influenced by randomness because of dice and cards.

    An ELO rating will provide an acurate rating if all people finish their games. The developers will implement penalties for quitters such as a blacklist, being bared or else, to minimize this situation that annoys everybody (not only in Catan, but any online gaming).

    ELO is a self correcting rating based on each player's rating. Eventually. the good players will have high ratings and the bad players low when everybody will start at 1000 at the same time (which will probably be the case when this site will charge, a reset is expected).

    The way you pick your first settlements (that determines a lot depending if you are first or last), where you develop, which are the priorities at the different stages of the game(in particular at C&K), how you trade and so on. Multiple parameters, the best player will pick the right ones to at the right time (again in particular at C&K where you have multiple ways of improving and gainig points). Of course we have all lost games where we are ahead by 5 points and the dice will just roll for the opponents and we lose at the edge of victory by one point. But in the long run it will happen to everybody.

    So the ELO rating is appropriate for Catan. There is a player I cannot name because it is forbiden to do so that I affectionate playing with. He wins almost every game, whatever he starts first, second or last. So your say that this is not a game of skills is a pretty naive statement. He also sits at one of the top ELO ratings from the list, and he ALWAYS finishes his games. I have never seen him quit.

    Such as poker, where the world champion is not the lucky guy who only got the best cards. If you think that poker is a game where ''You play with the hand you are dealt with'' you certainly don't understand the essence of poker. Poker is a game where position is more important than the cards you are dealt. Poker is a game of intimidation where your cards have very little to do.



  • Adding to above said - given this game has been developed as is ELO role here is to:

    a ) provide ranking (for whomever care about it);
    b ) to provide better matchups as explained by @Administrator it is one of the parameters in auto-match alongside region and karma (if checked).

    That 2nd point is what I am more interested in. And for me to get "better" matchups ELO must work otherwise we get situations like now where many complain about matchups and quitters.

    Yes Catan is partly game of luck and games like in my post above happens rarely (ends fast while roll is very distorted from "normal") but still there should be some logic / protection available there or at least a fairness principle - if I am beaten hard I get -7 ELO but if I win by a margin I should get at least the same +7 ELO.

    After ELO rest I have tried different approaches and slowly gained current ELO (close to 1070) but I still wonder how people get 1200+ given current system.



  • Yes, of course there is some skill in Catan. I wanted to make a point that better skill does not guarantee victory. The skill difference per one game only matters when playing with total noobs who don't know how to play. Yes, over a ton of games, some people might have more ELO than others. Still, Catan is too random to rank players based on games won. At some point when you have the basic strategies for initial placement and trading down, the game becomes quite equal due to the dice factor. IMO that ceiling comes really fast.
    This means that the max ELO won't be around 2800 like in chess, rather around 1800, maybe even 1400. If you have 4 players in a game and 1 wins to get ELO... You must win at least more than half of the games to maintain a positive income of ELO. Or maybe 2nd place gets some ELO too, but then if all players have 9, who is the 2nd player? Playing online, the average player will win 1/3 games (or 1/4 if playing with 4 players)
    If you start matching players based on ELO, the top players will have long waiting times. And then the matchmaker would place them in a game with lower ELO guys. If you know how ELO works, you know you will get more points (or lose less) if the top ELO guy does not win or even gets the last place. So no-one has any incentive to trade with him and probably block him first. A game where players can pile up on one player is not suitable for ELO. If you become the target, you can not win unless you are really lucky with the dice. This artificially created bias makes good players want to play less if they become targets only because they won more games than the other players. It just is not fun if these rankings are visible.

    Also, since Catan has different expansions, player numbers, home rules etc, then having a unified ELO for all these modes does not make sense. 3 player base game works entirely differently from 4 player games. Also the ELO will fluctuate much more in 4 player games because there are more losers. If you are good at C&K does not mean you would be good at base game.

    If you play with someone who always wins... Maybe it is your fault that you trade him the good cards or don't block him at the right time.

    Poker - I meant about playing one hand - sometimes it makes sense to make a bet, sometimes not. To know whether you have a chance or not, you need to know the statistics and read people. Similar to Catan in that sense. If you talk about the bigger picture, yes, it works differently. Winning with skill is much easier in poker than Catan. Poker does not have ELO rankings. Catan shouldn't have one either. It just distracts you from playing the game and obsessing about imaginary points which will fluctuate horribly through the games.



  • Getting back to my point: I want to be matched with players with higher ELO / higher skill rather than visa verse which means that assuming auto-match takes ELO into account I have to build up mine, which is hard to do given I get as much if not less points for flawless victory vs lose even by a small margin.

    I am not calling myself a great player but eventually people like me get stuck in the pool of noobs who are easily pissed off given I usually win them (or take the lead before they quit) by quite a margin while I don't get to play better players as I do lose some of the games due to skill as well as randomness of the dice etc.



  • I finished game with 16-5-2 yesturday and won +5 ELO points. I can finish with 16-2-2, but they want to leave, and i didnt use robber for them. What was the best way to collect more points?



  • @Stroom Yeah not quite. Poker involves chance and social interaction but it IS a game of skill. That's why you see the same relatively small group of people at the end of the World Series of Poker year in and year out.


Log in to reply