Auto-match variables

  • Question / suggestion - does auto-match, match maker take into account experience parameter?

    Thing is that quitters as mentioned before usually has a very high ratio between ELO and experience - as they quite games which they can't win => gain a lot/max ELO with least games finished.

    Also today I got matched with player who has done ~40 games and has as high ELO as I do with 100+ games.

    So I would suggest following:

    When matching players take into account (maybe even in equal amount) ELO, Karma AND Experience = player level.

    Get on your end some red flag / marker for players who has high ELO/Experience = level ratio and if combined with quitting report just reset his/her ELO instantly!!! Shouldn't be that hard + this system will not be player dependent - whether or not players actually report people (which is a big question of mine given feedback in forums and reaction when I say that my list contains close to 100 people by now).
    Automated player screening should be fast and does not need to be done by a human, just for review of results human involvement would be needed.

    This in my view again would solve / improve multiple MAJOR aspects of this game:

    1. quitter problem
    2. match maker issue
    3. gaming experience
    4. ELO distribution (given players will be closer matched in ELO and experience, in case of a luck/bad luck there will be no significant damage done).

  • I do not agree with any of these suggestions. I do agree that quitters should be punished for quitting games. Not losing ELO for quitting is the real problem not the fact that they have a high ELO. Trying to fix the consequences does not remove the problem.

    You clearly do not understand how ELO works. The amount of games you have played does not mean you should have a bigger ELO. ELO is intended to show a players' skill level. It can not increase infinitely. The fastest way to gain ELO is to win games against players who have a LOT higher ELO than you. There is a maximum cap which is defined by your skill level.
    If you manually start changing the ELO levels of players as "punishments" you will take the balance out of ELO calculations. The player whose ELO was reset can play with differently (less?) skilled players so if he is much stronger than the players in his new ELO, it will be unfair for the players who actually are as skilled as their ELO level suggests.
    Also since you seem so obsessed about ELO... I think it is a bad addition to this game because it makes people angry when they lose. This whole post is probably a result of losing a game to a less skilled player and losing more ELO than you gain from 3 winning games.

    The actual focus in fixing the matchmaker should be to fix the custom game and make that better to use. No matter how you try to modify the metrics for matching players, games will be unfair in one way or another. Player with low ELO but high Karma and Experience can be placed in games with players who have a very high ELO and low karma. So if you want to boost your ELO, you can just create a new user, quit a few games and then play with players who are always less skilled than you (assuming that ELO works in Catan - probably not). Also players who have low experience (meaning have not played much) are forced to play with players who quit the games often, having low karma. This does not really offer a good first experience for online Catan.

    Automated player punishment based on some idiotic metrics are not good. Some people can just be unlucky and have to leave 2-3 games in a row because of things out of their control and the game just punishes them automatically. Guilty until proven innocent?

    (Mandatory start using a lobby mention too)

  • I've come to the conclusion that the quitting issue is complex. After my first two games this weekend ended with no trades my way, at all, during the entirety of play, and though I play for enjoyment and not competition or ELO, my patience began to wear thin. The consequence? I ended up quitting my first two games at Catan Universe in subsequent games when I let other players get to me with their totally inept play or when they placed one of their initial settlement in a spot that killed my game, but did absolutely nothing to help theirs--just because they "saw" through my strategy. I furrowed my brow at that point. It was quite clear I was going for a port strategy, nothing to "see" through. In the past, I would have played through these types of situation just to keep my karma level. However, I quit both games near-ish the beginning. My point, people quit for different reasons. Mine had to do with some sort of code of ethics I hold where playing a good game respectfully (i.e. not being a/an "insert your obscenity here") outweighs winning. My point? I don't believe people like me should be "punished" in the same way as quitters of the last minute type. So the entire punishment thing is unlikely to be viewed fairly by all or most. If quitting late in games because of ELO is the problem, then make ELO an internal CU metric for matchmaking purposes only and don't make it visible to members. Apologies for the longish rant.

  • ELO being hidden would be a good idea indeed.

    What comes to the initial settlement placement - blocking someones game is a very valid strategy. Sometimes you take the place with worse numbers if it means other players have to take something even worse. If you get blocked off like that then it is your own fault for putting all your eggs in one basket.

  • Valid, yes, if it helps, or at least doesn't hurt your game too badly. This was a 5:2 hex with sheep port, or a 5 pip port, or something like that, and he placed his settlement directly on the port. He ended up with one sheep hex, a 10, with his second settlement, totally wasting the value of the port for himself too. Destroyed my game and his simultaneously. Sorry, it was a dick move, one that no one with respect for the game or any ounce of maturity would ever make. And certainly not one I would ever make, which is the barometer I use. The other player thought likewise, so I felt a bit vindicated.

    Anyway, on the more relevant issue, glad you like the hidden ELO suggestion.:-)

  • @Stroom read my post please once again. Really will not get into commenting your wrong assumptions.

  • @mrolimpia112 said:

    Thing is that quitters as mentioned before usually has a very high ratio between ELO and experience - as they quite games which they can't win => gain a lot/max ELO with least games finished.

    You assume that every player who has a high ELO and low experience are quitters. They probably are but they might also be highly skilled players who are new to Catan Universe. So they just won a lot of games against players who are less skilled than them, climbing the ELO ladder to where they are supposed to be. The same happens in chess leaderboards - create a new user and play the less skilled players until you have the ELO you deserve. You initially win about 99% of the games until you are matched with a player who is as good as you. Experience is just some number that can increase all the time. ELO is not. Experience number basically means nothing, only how much time you have wasted playing on Catan Universe. Players who come from PlayCatan start at 0.

    Also today I got matched with player who has done ~40 games and has as high ELO as I do with 100+ games.

    So what? You have played more which means you also lost more. You are at the skill level you are supposed to be. That player started playing with less skilled player and now has the same ELO as you. This is not unfair, it is natural that more skilled players climb the ELO ladder faster. Or he was a quitter, does not really matter.

    You are not addressing the source of the problem, you only try to compensate the consequences. Source: the players who quit do not lose ELO as if they lost the game. If this is addressed first, ELO quitters would stop quitting. Your solution only messes with the current system in a way which does not fix the quitters problem but can damage innocent people.

    Also it is much better if ELO is hidden from public. Clearly people do not understand it.

  • My point about ELO is about ELO/experience ratio - as you said majority of the cases will be quitters so if that ratio is higher than average + person is reported as quitter multiple times I see no reason why to spare such person.
    a) when we will stop spearing people and let them get away unpunished for the dammit biggest issue of this game?! even one quit is enough for at least some sort of penalty, few well...multiple - no discussion should be even present!

    b ) noone is forcing skilled players to play lower games by deducting their ELO, it is more like a cleanup so those skilled people are not matched with nub quitters. quite sure they dont like it as much as I dont when I am paired with a person who has questions how to move a robber while having as high ELO as I do.

    c) "wasted" - well dont play it then. I see experience as a valuable factor of not only games played but also finished. Either that has to be taken into account or @Administrator has to implement started/finished game statistics and penalties accordingly for quitting.

    Regarding your final point - I am talking here about cases where ratio is way too high man.
    I mean lets assume you are new user but skilled. You get matched with +/- 50 ELO players around 1000. You will probably beat them BUT you might not gain much ELO due to the fact that you will be matched with sub 1000 ELO at first for sure, you might even lose some of the ELO due to luck / playing vs other new skilled players etc.
    My point is your average ELO gain probably will be 4-6 / won game. AND based on my experience when you will lose you will lose hard - close to 1:2 ratio (meaning 8-12 ELO on some occasions).
    Which brings me to the point that even if you are skilled player you will still need X amount of games (even you pointed that Catan is ~30% luck).
    While quitters usually go for:

    1. NO ELO LOSS!!!! Which gives them boost by ~1-3 games per each possible lost game they dont have to "balance out for" and play more games;
    2. THEY keep gaining at least 4-6 ELO, when certain level is reached and they get into 1100+ ELO people while being 1050 even more per game.

    So their ELO/EXP ratio in reality will be at least 2 time higher than average up to even 4 times. You can just take my banlist and see some of the examples.

    As I said in comparison to me I have done like ~120 games and gained ~70 ELO = average ELO / game for me is 0.6!!!! while we are talking about at least 4 in case of constant quitting = almost 8!!!! times higher AVERAGE!!!! I mean you have to blind or stupid to defend such player (not even talking about cases where average gain is higher than 4 and difference already 10 times or higher).

    As for hiding ELO - Amen to that!

    1. Your solution does not fix any problem. The problem is that quitters do not lose ELO at the moment. If that is fixed, quitters will not have a higher ELO than their actual skill level.

    2. You still do not understand what ELO or experience mean. You think ELO has to be giving free points to players all the time. In fact it is false.

    a ) My whole point is that your solution does not fix anything. Quitters should be dealt with but not like this.

    b ) The matchmaker actually IS forcing skilled players to play with less skilled players. If you have to wait long enough you will be placed in the queue with less skilled players. (Also, lobby is better)

    c ) Experience in Catan Universe is just a number showing how much you played on this platform. It does not take into account the thousand games played in PlayCatan. Also Experience only increases. You should instead make a new statistic about started games and finished games instead.

    ELO - some measure of skill level. It can not increase infinitely. If a super skilled player wins a game vs new players, he would not gain any ELO because it was predicted that he would win anyway. If a player who has lost all games so far wins a game vs a champion would gain a massive amount of points because he was the underdog. It is basically like betting in a horse race - in the beginning you might have 1:2 bet for the high ELO player and 1:4 to the low ELO player. If it seems unfair to you, you are just wrong.

    Experience - just a pointless number that indicates how many games you have played. You gain experience in every game - win or lose. See what the difference is? With experience you could have 10000 experience after playing a lot. But ELO would probably be around 1300. But you start at 1000 ELO and 0 experience. How the hell can you calculate the ratio difference based on these numbers?

    You clearly do not understand what ELO means. If someone gets more ELO per game than you then they CAN be better than you. The current problem is that the quitters do not lose ELO. THAT is the problem. (I have to repeat it because you clearly still do not understand)

    You played a lot of games and gained 70 ELO. Good for you. This is your skill level. You are blaming quitters for totally unrelated reasons, though.

  • Dude really...this is getting ridiculous.

    Not sure if English is your first foreign language or do you use some translator or there is some other problem.

    My suggestion clearly shows how quitters could lose ELO (which they have gained using system loophole).

    ELO principle and ELO in CU is totally different story. We are talking about CU here and obviously in CU ELO is a sign of rank.

    **If you keep telling people that ELO is (bla bla bla what wikipedia or whatever say) but 90% of community still thinks that that's a unit of how big your dick is, you will not prove them otherwise even if you post 1000000 times. **

    So we have to find a solution for current situation as again otherwise I could argue that in real life if player quits I get to beat his face up, unfortunately that is not viable in virtual environment...

    Who cares about PlayCatan or any other possibly relevant BUT STILL NOT CU game?
    We can debate that I can ride a bike but that is not mentioned here or more relevant - maybe I am playing Catan board-game 24/7 again not represented here.

    Here we have clear** 3 parameters which are being evaluated**, compared and used in the game: Experience / ELO / Karma.

    And my suggestion ties two of them (which are in my view more important than Karma as that filter is never used / karma is not checked by players before the game).

    ELO actually can increase indefinitely (theoretically), depends on community, skill and activity, plus assuming that posted formula is rounded up to full integer (0,0[0]1 => 1 not 0).

    Plus even with current scenario you can't gain HUGE ELO gain or lose because match-maker still ha a limit - haven't seen it matching player with sub 1000 with 1100+ player. So yes for some maybe +/- 16 ELO is HUGE but i mean it is still far from like 100!

    Problem at the moment is more of a situation that skilled players take one step forward and one step back rather than two forward and possibly just one back.

    0.6 ELO gain on average in my view would be a normal case for player at Elite ranking - given posted equation, as they would play only vs elite players and if won gain closer to 1 ELO and if lose still lose just around 1 ELO.

    Experience alone is not very useful that's what I suggest using ratio. Which in case will show not only how many games you have played but how well you have played how many games.
    Again this relates back to pros vs losers. As you say maybe player has gained 1300 ELO (which is quite high) using 10000000 games => player is nub.
    But given current system actually this nub eventually will be placed vs players who had gained comparable ELO in much less games => nub vs pro game scenario.

    Ratio EXP/ELO would place that 10000000 game player vs current equivalent of ELO sub 1000 or more precisely sub 100.

    To sum up:

    EXP/ELO ratio solves matchmaking (nub vs pros cases);
    Gives new "dick unit measurement" which I cant see to have loopholes;
    Is self-managing given odd cases are reviewed and ELO adjusted to quitters.

  • I understand English perfectly. And from reading your posts I can clearly see that you have no idea about how ELO works or how to do math.

    You complain about the loophole but you offer a totally irrelevant fix for it which will not work.

    CU ELO is pointless and we should not even have this conversation. ELO should be removed entirely. @Administrator pls.

    The solution to this problem has already been offered and it is to make quitters lose ELO as if they lost the game to all other players. Adding your "solution" on top of that would not make sense if ELO works correctly in the first place.

    Who cares about CU playing experience? Oh look, that player has played so much online, must be a good player... Umm... no. All metrics except finished games/started games are useless. At the moment only karma is the one metric I would need (and it needs some fixing too as I have seen from other posts). Experience is just a random number to make you feel happy about playing the game. Another "big dick" thing to give players so they would keep playing.

    Ok, let's start combining our metrics. How would it work...

    Yes, theoretically ELO can climb indefinitely. But if you have learned anything about logarithmic functions, you will understand that the higher your ELO is the less points you will gain when playing with people of lower skill level. Very few people will have a high ELO so when they enter matchmaker, they either have to wait a LONG time or the matchmaker will at some point place them in games with less skilled players (See League of Legends matchmaker - same problem). So they have a much higher chance of losing ELO than gaining it. And one lose would bring you down hard. There will never be players with more than 2000 ELO in Catan. Maybe even 1500 is the practical limit.

    First of all I say that experience is a pointless metric. It does not show skill, it shows how much a player has played the game. If you combine them you can get highly skilled player with low experience playing with a low skilled player with high experience. This will allow players with higher skill gain ELO faster when they make new accounts. Therefore if you want to win constantly, create "Smurf" accounts and have fun playing with less skilled players. This will drive people away from the game. Example 1 why this is a bad idea

    You offer EXP/ELO. Let's see. Everyone starts at 1000 ELO or something. ELO will usually be around 800 and 1400 in Catan. Experience would be from 0 to... let's say 10000. A player might have EXP/ELO from 0 to 10000/800 = 12. Let's say we have a user with 10000 EXP and 1000 ELO -> score 10. Another player has 1400 ELO but he has played less so has less experience - 8000 -> score 5.7. Why would a more skilled player get a smaller score? Because he has played less? But he has won more! The amount of games you have played does not reflect your skill. Example 2 why this is a bad idea.

    Let's use a notation - Experience is roughly a number or games. 1000 EXP = 1000 games. ELO is a number of skill. EXP/ELO = games/skill - what does that even mean? How many games you have played to earn your skill level? Why is it relevant to the matchmaker? If a player is good without playing thousands of games, he should be able to play with people who have played a lot more but also have his skill level. Your suggestion is just stupid. In terms of Chess it would mean that a young player (Magnus Carlssen) can not play with an old player who has played much more than him (Garry Kasparov) unless he plays a thousand games and "catches up". Although they are similarly skilled. Example 3 why this is a bad idea

    ELO is more or less a constant value compared to experience (experience practical limit is infinity, ELO practical limit is 1400) so the more games you play the higher your EXP/ELO rating will be. Even if you are a good player but who plays less games, you will start being placed in games with less skilled players who just play 5 times as much as you. Example 4 blah blah

    The way "nubs" and pros are matched together with a well implemented ELO system is only if the pros have not reached their appropriate ELO yet. This is inevitable. No matter how you combine other metrics about the games.

    To sum up your sum up:

    EXP/ELO does not solve matchmaking - nubs and pros will still play against each other.
    The "dick unit measurement" is just a hidden experience number which still has loopholes - just play more and it rises.
    As said in a post before, manually lowering ELO of some players only hurts the integrity of the ELO system. The same players, if they are more skilled, will play with players with less skill and they will ruin their fun. Some players would be unjustly detected by the system and the admin will have to deal with it.
    The real fix is to either make quitters actually lose the game and ELO or remove Experience and ELO entirely - useless development feature whereas you should fix the gameplay.

  • Dude my English is good, my math is good and my understanding of posts is good.

    Sorry I can't keep up with this ... I write a sentence you write 4 cross-referencing everything and from what I see stumbling away from the point.

    I guess this is the place I call it and agree to disagree.

    @Administrator feel free to clean up all that bla bla bla and just focus on initial suggestion.

  • Can you say where I am wrong with my post?
    Especially about EXP/ELO score still matching players of different skill level.

    It seems like you just ignore my post not even trying to read it.

  • I agree with whomever said hide the elo.

    I think more people will play with a better experience of each game played.

    Quitters, for whatever reason-good or bad-I will try to think of a neutral term for quitters-do harm the game experience for the other players.

    I would hope the Catan Masters/administrators could figure out a method of determining how much the game would grow if game completion went up with no elo vs. how much the game will grow if they keep the current elo system.

    I suspect hiding the elo would grow the game faster, but would attract older, more mature players. Maybe Mayfair wants to grow their younger audience and they may think that elo has a big appeal for them.

  • @Stroom as I said before I actually read all that.

    You assume too many things that's your biggest problem/shortcoming.

    Another big issue is - you are way to critical. E.g. You just proclaim that experience is useless. And after continue with a paragraph of text which is based on your proclaimed truth...just a reminder you are 1 out of thousands of registered players.
    If i disagree with your initial statement (experience is useless I dont even read all what follows because it is based on that fake fact you create).

    Just reread my initial post and then follow up and try to see what main focus here is on matchups not so much on quitters while you kept focusing on quitter issue on your replies.

    As for the math - you can do it on your own, not really into teaching it here.

  • OK, so you did not read my post. That's all I needed to hear.

    I don't think I make too many assumptions. The ones I make seem reasonable for me so I use them to provide real life examples to base the math on.

    I say that experience is useless and it is. What does it matter where the player has played the majority of his games? Is he supposed to play 100 games on Catan Universe before he can play with the "big boys"?

    You actually made an assumption that ELO and experience are both connected to the player skill level. This is not the case at all. You think that if you play more games then your ELO will always increase. You clearly do NOT know how ELO works.

    Experience roughly shows how many games you have played in Catan Universe. If you have played offline or on another platform, it does not matter. It does not show anything, really. Is there a difference between 100 and 200 games? 1000 and 2000? At some point is becomes useless to count the amount of games a player has played. If you start using this as a metric for assigning players to match each other... Will not work well.

    After saying that Experience is useless, I wanted to show what would happen if you use experience as a part of matchmaker. I showed that when you use experience as a metric for assigning players to play together, you will start matching less skilled players with more skilled players. It does not help against quitters at all.

    1. You specifically list quitter issue as the first thing your solution fixes. I say that it doesn't and the best solution to all this is to make quitters lose ELO. No need for additional calculations after that.
    2. I also showed that using experience in matchmaker will not solve matchmaker issue.
    3. And it will make gaming experience worse because players of different skill level will play with each other.
    4. Your ELO distribution claim is flawed. Experience and ELO are not related metrics. A bad player playing 100 games is not equal to a good player playing 10 games. Your matchmaker would think that they are.

Log in to reply