Ladder Site Online...

[eyerouge]

Thank you for the feedback guys/girls. I'll take the time to try to answer each and one of you in order, and hopefully it will shed some new light on the issues at hand or lead to an even more constructive discussion. In either case, let's start already. Just jump tp the section with your nickname above it.

Wintermute

. However, winning many games *should* give the player at least a small reward.

Yes, and it does: The more games you play, the more real chances you have to earn (or potentially lose) points. That's still true, despite the new addition. You are rewarded for each additional game you play, just like you were before, granted you win of course.

Let's also keep in mind that the newly added cap is just that - a cap. It still allows people that want to play i.e 14 games a week to do so. 14 games is, from my personal perspective, many within one week if you are a serious player and pick opposition that's about equal to your own skills (unless of course you always happen to agree on blitz settings with all your opponents). Peoples lives vary, we all have different amount of time and dedication, jobs, school, kids and so on. I don't question that, nor do I see my suggested numbers as anything near the final ones or "correct" ones.

I do however strongly believe that the ladder should be something that is for the median players, or some version of them, when it comes to how many games they can/will/want to play per, say, week. I don't want a ladder that primarily rates or rewards peoples amount of time spent on the game. There is of course a correlation between time spent ingame and the skills you might develop, but as we all know, that is just a correlation in the best cases as there are plenty of players who have played the game countless hours without ever improving due to, for example, them only playing weak opposition or not learning the game further for other reasons.

If we suppose that most active players on the ladder would usually have the time/chance to play x amount of games in a week, then it makes sense to allow about that many games/week (for the various reasons suggested in other posts, above) if we want to create a ladder that is directed towards most of it's members, and not only the top time-spenders on it.

Let's now imagine the opposite: A ladder where anyone can play any amount of games, unrestricted. It sounds swell, but really isn't that much, unless of course you happen to be a non-median player, meaning, a person that has, for whatever reasons, plenty of time to put down in playing Wesnoth on the ladder, and which can do it in such a pace and to such an extent that any median or average player will always have a very hard time to a) compete or b) have correct rating in relation to that person that is match spamming / point harvesting. A & b have already been explained in my previous post, and this far nobody has refuted them in any way, thus I still have no reason to do so myself yet.

Personally, I think that limiting the number of games is not a desirable thing for the ladder.

I totally agree, however, I'd like to rephrase what you wrote and change the word ladder with "ideal ladder".

I agree with you when we're discussing the perfect ladder as a theoretical construction. Such a ladder would however also have some things associated with it which the current real ladder has not right now. One of the key elements would be a) heavy participation, something which you suggest yourself, and which I agree with. Another would be b) random distribution/selection of opposition - that players don't get to pick their opponents them self.

The role that b) plays, or rather, that the lack of the condition in b) plays, can't be pointed out enough. It's the core and cause of many problems associated with the Elo rating since the Elo rating isn't really built to let people pick opposition: It is more accurate if people are forced to play randomly, or by other means by the system selected, opponents. In the end one might argue, just like you do, that Elo will eventually self-solve this problem. I agree that it will, but it would take really many many games before an i.e. overrated player that intentionally picks games where he gets just 1p or something like that (in other words a person that always plays games wiuth minimal risk and minimal, but some, gain,.. a so called match-spammer) gets his correct rating. Meanwhile the rest of the community watches such players "flourish" on the ladder, which, in it self, is a real turn of if you're a player that value good competition, sportsmanship and a nice community. It would, arguably, even scare away serious players that don't back down from most challenges or players that have a more normal distribution in regards to their played games history and their opponents different range of skills. In short: Due to the level of (in)activity Elo takes very long to handle the problem at hand. During that time it is likely that a) median players will leave the ladder since they notice that match-spammers get better rating than them by picking their victims and always risk and profit minimizing and/or b) people will adapt and start match-spamming them self in order to compete with the match spammers, creating a new atmosphere and re-defining half the meaning with a ladder in the first place.

I'm not saying that that is what would happen with necessity. Just that it seems somewhat plausible, even more so depending on what you think about the crowd at the ladder and how people react to others actions taken on it.

Summarized - one might ignore the whole discussion and just ask the simple question: For which audience, which players, should the ladder exist? As suggested, I think the answer should be for "as many as possible". To make it so in practice we apparently need to have some restrictions. Why? Because it probably solves problems, as I argue in a previous post. It also makes it possible for more players to join up without feeling they have to play 100 games each week in order to stay competitive. After all, Wesnoth is a strategy game - insta-action and games that last 5 min isn't really the standard norm, nor should it ever be for a turn based strategy game, even if it can be fun to play with such settings from time to time. A ladder for a typical FPS 1vs1 Deathmatch using conventional settings (x amount of frags, y amount of time etc) would have a very different participation rate (amount of players registered and also amount of games played each day) than a ladder for a strategy game like Wesnoth. Most strategy gamers actually think when they play, or at least give the illusion that they do Super-many Wesnoth games using standard settings are not played since it requires thought against equal or harder opposition, and that requires time unless you happen to be godlike, which clearly the average players are not.

I guess my point (if I am even making one ) is that these rankings are still pretty rough, and personally, I don't think there is anything wrong with that. If this is still a big problem a year from now, then I would be worried. But my guess is that things will continue to smooth out.

Once again I agree, fully. However, as I mention elsewhere, it takes time for them to self-heal, time which would be drastically reduced if we played an FPS instead of a 4x game since the average player would have time/will to play way more games than the average wesnoth player has.

One thing I think that needs to be spilled out in the clear is that there might even exist a really plausible risk that the problem will indeed be around in 1 year and even in 10 years even if participation is still the same or rises. How so? Easy - because it's not just a question of ladder activity. It's a question of who's activity: Players come and go constantly on the ladder. For things to smoothen out and for players to have correct ratings they'd have to be active for 1-2 years in a row, at least, with the pace it's having now. Unfortunately most players are impatient and/or stop playing ladder games for whatever other reasons they might have. In the end, if you look at the ladder every third or sixth month, you'll notice that the active players, and especially the top 20% of them, are brand new/different players (rating wise on the ladder, not necessarily irl). The conclusion is that, due to player mobility, the problem repeat them self instead of mending.

If there is a feeling that some action *must* be taken, then as one suggestion, perhaps a challange system could be put in place either as part of the sportsmanship rating, or as a seperate rating.

I hope it's somewhat clearer that there are some good reasons to take different actions. I also share your notion that having an additional way to find/meet opponents and count ratings or keep track of some new numbers would be of use. It will come in the future.

Monster

If people see that someone is match spamming they should settle the matter by themselves by challenging such a player to a match and show him just what it means to be a high ranking player, which you don't become just by scoring a high amount of ladder points but by also gaining recognition by winning over truly skilled players

1. A match spammer will in most cases not accept the challenge since it comes from a player where he'll risk losing plenty of points if he loses. Since the match-spammer knows he's a spammer he also knows he's overrated in relations to the really good players, regardless of their Elo rating.

2. Even if the match spammer accepts a game or two against such players he'd get beaten by them and then fall back to th ebehaviour in 1 and then stop accepting games with them. There would still not exist any mechanism that would hinder him from spamming on.

3. I agree with the thought that you don't become good by just having a high Elo. You understand that, I do too. However most ladder players seem not to recognize that as fact, nor do they seem to actually understand what a ladder measures and how it all works. My guess is that most, sadly, focus too much on the Elo and believe that it measures some kind of objective and correct skills. Clearly they have never read even a fifth of what I've stated both in the FAQ/Rules and/or in this thread. On a side note: I'm considering a way to keep track of how many % of your games were played (and won/lost) against the top x players at the time. I think it would be in line with your suggestion...

I don't know if this is the result of just one man's email complaining about his opponents

Here is what it's a result of:

The pure fact that any idiot can accumulate however many points by always playing plenty of games where he only earns 1 - 2 points per game, and that he can continue to do it until he gets ridiculous ratings in relation to all the others who play normally on the ladder. On top of that it would take some serious time before that idiot got so many points that he'd have to start playing harder players to continue earning any points at all. It's a thing that's derived from the way Elo rating works since Elo is designed around the thought that many games are played and that they're played against "random" opposition. If you re-read my original post about the problem and/or the reply to Wintermute (within this post) you'll see that there are plenty of reasons for the cap and few, if any, against it.

1) Such changes of the rules affect all ladder competitors and in its core just slows everything down, 2) especially the
eager newcomers to the ladder that want to gain recognition and and are yearning to prove themselves by climbing the ladder.

1) No, if the amount of games x played within intervall of days y is set correctly it won't really slow down the average player. What you say is mathematically false. The only persons slowed down are those currently "speeding up" the ladder to levels that are absurd, for various reasons, for most strategy gamers. With other words: The cap makes it easier for most players to be competitive, while it makes it harder for those that "competed" by match-spamming. Very simple logic, really.

2) That's partially false as well: Newcomers are in a special mode when they come to the ladder - they are ranked as provisional players. As such they get/lose plenty of Elo points after each game. They hardly need to play and win many games at all to get really good ratings to start with. Once they are out of provisional mode (after 10 games?) they are rated as any other in their own skillclass.

(Also, if you refer to the idea that a newcomer should be able to reach the heights of i.e. 2500 Elo within a week or a month by playing a zillion games then it's a weird notion: Why should a person be able to do that in the first place? What's the rush? Is it even realistic? How come it doesn't work like that in i.e. Chess? Again, even if there were some really good reasons for why we would want people to get an Elo rating of 3000 within a month I'm still confident they will in most cases succumb to the fact that, while being good for the small minority of players that can accomplish it, would still be very disencouraging for the large majority of players since they have no fair chance to compete on the ladder. That in return turns the ladder into a forum for those with the most time to spend on playing on it, premiering them, as XP is awarded in a crappy RPG where you earn it by grinding monsters. The ladder isn't an RPG, and the Elo is not XP. I can however add an XP meter for those interested in it, and that would keep track of how much you have grinded.)

In some way it even takes the fun out of the game and awards calculating players that will want to ensure a definite victory everytime they play ladder

What you write is exactly what would presumably be fixed by the cap I enforced. As it was before the cap people could grind. That means they could pick the easiest possible targets that gave them any amount of points, even 1p. If you keep on repeating that enough you'll accumulate an Elo that's distorted and also wrong compared to other normal players. Why that is has been explained several times in this discussion - check my reply to Wintermute or the replies previous to that. It has to do with what Elo is and how it works, and mostly, how it comes that it fails in some respects due to the huge freedom ladder players still have.

Short repetition: If a grinder get to play 10 games and i.e. Gallifax get to play 10 games, Gallifax will amass more Elo than the grinder since he is way more skilled than the grinder. The grinders supposed skill is still only the sum of grinding. So, even if he would grind those 10 games he'd earn 10 x 1 = 10 additional Elo, while any normal player with real skills would get more. Hence, the gap between all grinders and players like Gallifax would widen with time. Even more important, due to that fact, grinding won't get you far at all on the ladder. On the contrary, you are encouraged to play people around your own skill class and/or harder opponents, something which in my world seems very "sportsmanshippy" and as it all should be.

1) Also such rules in a way feel like our rights to play whenever and with anybody we want, are being cut down. 2) There is a thin line between control and opression.

1) No, that statement is false: You can play whenever you want and with whoever you want. The cap doesn't affect who you are allowed to play or when you play. It caps the amount of ladder games you're allowed to play within a period. By doing that we make life sour for a few and the ladder more feasible for the masses. Add to this the arguments I present, the fact that none of them(?) have been refuted, and also that no veteran which has a real understanding of the game or Elo has complained yet, and I'd say we're on the right track until proven wrong.

2) I bring to you some of the paragraphs which you agreed on when you signed up on the site, if you've done so:

Code: Select all

# Using this ladder and our services isn't a right - it's a privilege that we choose to give to you as long as we want and as long as it's possible.

# If there are breaches in the agreement and/or other for us apparent reasons to do so the site administration has the right to ban users and/or delete/modify results and other info, together with any other necessary measures that a situation might require for us to secure the ladders integrity and the intended functions of it as the admin envisions them.

# The rules are always a subject for change. If many players require a modification, addition or deletion or if the admin sees fit the rules will be revised. Revisions are marked with a version number and they're announced in the news section as well as in here.

The ladder began as my project to learn PHP. I hosted it and I wrote the rules, maintained it and also coded on it an pretty much everything else related to it. It's over a year old now, started out with 0 registered users and 0 played games and also 0 official support and a lot of grief. Along the way I have obtained some amazing help and coding efforts from people like chains and mr russ for which I'm very gratefull. I've also gained unofficial but continuous and valuable input from some of the Wesnoth developers and other prominent persons in the community.

Until now I haven't gotten any feedback that the ladder limits or oppresses players, nor that additions to it have done so. Maybe needless to say, I have a very hard time seeing how something where participation is of free will, can oppress you. If it doesn't fit your shoe then simply don't use it. Since I'm an anarchist I'm also happy to announce that the ladder code, including all my own efforts and mr russ and chains, is free (GPL3) for you to download and use as you wish - you can run a ladder site for Wesnoth or any other game if you'd want to. I see all this as a sign of something radically different than oppression and would probably be insulted if it wasn't for your explaining that you had a bad day and so.

Lastly, I'm only admin for a short while. After that the fate of the ladder will be in the players hands. Personally I look forward to see what you guys will make of it.

Zarak-Kraken

I'm ranked 11th and I think I deserve my rang.I largely plays a lot because I am a very big player RPG and strategy games ... I write in 3 Ladder,wesnoth and 2 another game and I don't never seen such a restriction I'm just above.

As I wrote to Foisprey: Nobody questions you. You as a person/user aren't relevant for the problem at hand. Nor is any other single person. The problem is/was the mechanics of the ladder which allowed a certain type of behavior. Even if the ladder had zero users the problem would still exist in theory unless it was solved.

How other ladders work I can't really comment, but if they are open, free and use Elo I'd say they have similar problems as we have/had. The fact that nobody coded or enforced a restriction doesn't prove anything. However, there might be other explanations for why they haven't felt the need or seen the problem. One apparent candidate is that they don't use Elo.

Gallifax

@eyerouge: Might be obvious by now, that I dont speak up often. I only really do it if I feel its neccessary.
So here it comes... Limiting games.... not good at all:) Contra productiv

I agree that it is contrproductive if the ladders goal was just one: To get as many game reports as possible. Now we will get less reports and probably less played games since the match-spammers can't play as many (In general and in the long run one might on the other hand that this can allow more players to enter the ladder and allow them to stay competetive, resulting in an overall increase of games even if we get a decrease the coming nearest months.)

The thing is that the ladder has more objectives than just getting as many reports as possible. One of the goals are to get "quality reports", or rather, "quality game play" and statistics that are worth having and have some kind of connection to the reality. We get that not by encouraging or allowing match-spamming. We get it by relying and/or forcing most of the people to play players equal to their own skills or above them, or at least get rid of all ways to abuse a system which was originally designed to function in a way more controlled competitive environment than the online ladder. In the long run that also leads to an overall better community, skill wise if nothing else, since it encourages skill development.

I understand it may look contraproductive, and, it even might be so in reality. But, what's the reasoning behind that idea? How do we come to the conclusion? And in what way is it contraproductive?

All
Granted you have read most of what has been written on this subject, please come with the following feedback: What kind of settings would you use for the restriction and why? Please don't forget to explain why and argue for your opinion. The variables we can easily change is a) the amount of games we allow and b) the time span of most recent x days. It's all coded properly so we could set the value to whatever you guys want. I do however require something rational incentive here.

[Wintermute]

Granted you have read most of what has been written on this subject, please come with the following feedback: What kind of settings would you use for the restriction and why? Please don't forget to explain why and argue for your opinion. The variables we can easily change is a) the amount of games we allow and b) the time span of most recent x days. It's all coded properly so we could set the value to whatever you guys want. I do however require something rational incentive here.

That was quite a lot to take in. I will focus my reply in terms of the requested feedback.

I think that you pose a difficult question to answer, because as I understand it, everyone posting in response to the idea of fix number of games is suggesting that there should not be a fixed limit. So asking "what should the limit be" is hard to determine. I, for one, have no idea what a "reasonable" limit would be - because I don't think that any limit is reasonable. So if you want a number: I suggest 1000 games per month.

More seriously, I think if this is a problem worth addressing, I think the way to do it would be closer to how it is handled in chess: setup ranked events where you do not have control over whom you play. Now, clearly this would not work for Wesnoth in the same way that it does for chess, but I would suggest the following which might have the desired result.

Currently, to keep your rank you need to play 1 game every month. Why not make that one game vs. a player within ten ranks of you in one month? Or two games? Maybe also within 15-20 ranks (since the players at the top may have a hard time matching up with only 10-12ish people. It is a bit of a restriction, since it may legitimately be hard to match up with the right players, but if you really care about being ranked, I think you owe it to the system to play regular games against people your level. Regular in this case being 1 or 2 a month (not too bad I think, but maybe enough to take care of this problem).

An alternate idea would be to apply a rule to the top 20 (or top howevermany) spots where they have to play each other regularly in some fashion. Something like: to move up from #7 to #6, you need to have the points to do so, AND have beaten someone ranked above you. Though this might lead to players having higher points but lower ranks than other players - it still might work well to make sure you can beat top players to get top spots. Granted there would be some flip-flop in the spots over time as players beat each other back and forth, but that happens now to some extent anyway (and maybe it's a good thing overall?).



Actually, the more I think about this second idea, the more I like it. It would mean that to be ranked number 1, a player would have to have beaten at least 20 top-20 players, including the top few at the time they moved up. That seems like it would give some addititional credibilty to the ranking system that it does not currently have (namely, that top players to not need to have beaten other top players - which is one clear advantage that the ToC system has over the ladder rankings). Potentially, it might also lead to problems with scheduling. But perhaps not very many - I imagine it would only be an issue near the top (with few players in your available pool to move up from a win over), and keep in mind you would still need to amass the points to do so.

Would anyone else go for this idea? I still feel that it is not really enough of an issue (given that the ladder is still 'young'), but I would rather see something like this than a fixed game limit (which has the potential to drive people away or at the very least be annoying to some players IMO)

[Pelopidas]

I mostly agree with Wintermute.

As an alternative or add-on one may suggest that players do not get any more points if they do not match in a certain percentage of matches around equally well rated players.
I heavily oppose the limitation of matches - reasons:
1. Beginners and bad players have a right to get an opponent. I never refuse a match against a bad player although I do not get significant points from it, because it could give them a bad feeling.
2. Some players (including me) unfortunately have changing amounts of leisure time - I can enter in weeks where I can play and want to play more than 14 matches and there are many weeks when I cannot play at all. The present rule thus disencourages players with changing timetables AND active players.
3. A limitation of match numbers is far less efficient than a limitation of percentage of matches played against far-off players. A naturally selecting player will always have a mix of opponents he is playing against, so that one can easily determine those who are just keen on points. If you limit the overall number of matches the rating-maniacs will suffer no significant disadvantage since they still can match weak players.
4. Additionally I would never want to punish active participation in the ladder like this. You force players into inactivity although they could play matches and increase overall fun in the ladder. The ladder lives from an active community of players being online. Limiting match numbers you decrease this.

If that is not fully convincing, there will be many more reasons. I think it is a good idea to use some measures against players just keen on getting a falsely high rating. One should however always consider the price one has to pay for it and personally I cannot really see why another player's high or overly high rating should affect my personal happiness. If he consequently denies a match it does some harm, but I do not get unhappy with a small number of players trying out (legal) strategies to increase their rating a bit.
Anyway for the above reasons I strongly oppose the set rule.

Best,
Pelopidas

[tsr]

I seem to recall from my chess-playing days that highly ranked players would reach a cap where they wouldn't gain a single point from beating a player with a lot lower ranking. Why isn't it so in the ladder? Or has it to do with that cap being set very high?

/tsr

[Gallifax]

@eyerouge: You want quality games and good attitude. I consider it also good attitude to play with anyone, no matter oppoents rank our outcome of match. Low cap ruins that.

Also If players play with no observers or dont upload replays the ladder will hardly benefit from quality games. So what you want cant really be force upon players if they dont want that.

Just for your consideration.

[eyerouge]

wintermute

Thanks for extended feedback. I'll try to keep my replies short, not that I wish to sound impolite, on the contrary - so I won't eat too much of everyones time I know several people have complained I have a tendency to write too long forum posts... *hides*

More seriously, I think if this is a problem worth addressing, I think the way to do it would be closer to how it is handled in chess:

It is a problem. Not only is it a match-spammer problem, it is an overall problem with most players ratings on the ladder. The problem is, as previously described, derived from the combo of Elo + freedom to choose opponent. The implications of it can't ever be overstated. In short it makes the whole rating less useless, something which you correctly put in contrast to arranged tourneyratings. [In the above posts I also openly write that nothing beats tourneys when it comes to skill measuremnt and that a ladders primary function is something else and that ladder is a middle step between casual play and tourney play. That said the current ladders rating mechanics can still be improved, hence these creative discussions.] The problem is just more visible with match-spammers, they "personify" it pretty well.

I fully share your belief that there needs to be some kind of mechanics in place that somehow direct the player towards certain opponents, following certain criterion. I also aim to get such things implemented in the code, either by doing it myself or by begging on my bare knees that somebody else gets involved and does it for us It is on the roadmap, or rather, mindmap at the current stage it is in. Btw, all suggestions and developments of the ones that have already surfaced, are welcome. There's a lot of stuff to think of in regards to this, as I'll soon show.

Currently, to keep your rank you need to play 1 game every month. Why not make that one game vs. a player within ten ranks of you in one month? /../

An alternate idea would be to apply a rule to the top 20 (or top howevermany) spots where they have to play each other regularly in some fashion. Something like: to move up from #7 to #6, you need to have the points to do so, AND have beaten someone ranked above you. Though this might lead to players having higher points but lower ranks than other players - it still might work well to make sure you can beat top players to get top spots. Granted there would be some flip-flop in the spots over time as players beat each other back and forth, but that happens now to some extent anyway (and maybe it's a good thing overall?).

The above ideas and most possible variations of them have the problem which you mention yourself: They are restricting a) as to which people must be around in order for you to be able to play a game. Depending on different timezones and what not this can easily become a serious problem and even reduce the number of games way lower than a hard-cap. b) On top of that, and possibly more seriously, players could use their power (being near above you in the ladder, or below for that matter) to "freeze you out" of the competition if they set their mind to it or form some kind of cartel.

Another problem is how activity would be effected, even if we'd remove the current cap: Because of only about 100 active players there either won't be any meaningful distinction and it could just result in a slight real life effect due to the few person around and which "Elo class" they happen to be in, or, the contrary could become true - if we instead narrow it too much (say you are only allowed to play the 10 people above or below you) then it will perhaps be impossible to find games. What would the balance be? I'm sure some decent one can be accomplished, question is what that would be and what the effects would be with i.e. 100 active players and the elo ranges we currently have distributed over the whole ladder, which is everything from 1300 to 2090. (And nevermind the people that are first or last, I already have some solutions ready for that)

Your second idea did cross my mind, and it was also strongly advocated by chains several months ago. At the time we didn't find it a pressing matter and let it go, mostly due to my skeptical attitude towards it, partly because of the problems mentioned here and now.

I have however changed opinion about it and I would love to have a system that somehow took more notice about which players you actually played. I'd like to see statistics of how many % of your games were played against the top x players for example, also see how many different opponents you have played, how many times you have played which guy etc. All this can be fixed and will be. The question is how we should approach the main question for now - how to implement some version of what Wintermute suggests. If we had such a system, I'd also believe that we could remove the cap that's currently in place, as I fully agree that the cap is crude and the alternatives are way better.

Pelopidas

1. Beginners and bad players have a right to get an opponent. I never refuse a match against a bad player although I do not get significant points from it, because it could give them a bad feeling.

That has been answered in detail already in the above posts. In short: Beginners are not affected the way you and somebody else describe due to the provisional system which is also in place.

2. Some players (including me) unfortunately have changing amounts of leisure time - I can enter in weeks where I can play and want to play more than 14 matches and there are many weeks when I cannot play at all. The present rule thus disencourages players with changing timetables AND active players.
3. A limitation of match numbers is far less efficient than a limitation of percentage of matches played against far-off players. A naturally selecting player will always have a mix of opponents he is playing against, so that one can easily determine those who are just keen on points. If you limit the overall number of matches the rating-maniacs will suffer no significant disadvantage since they still can match weak players.

Point taken. I agree.

4. Additionally I would never want to punish active participation in the ladder like this. You force players into inactivity although they could play matches and increase overall fun in the ladder. The ladder lives from an active community of players being online. Limiting match numbers you decrease this.

Currently, when there is no alternative system that solves the problem in a satisfactory way, it may very well be a good call to use the cap since we could lose more players in the long run without it. Reasons for this have also been explain in above posts. In short: Average "normal" players won't ever be affected by the cap. If they are, then the cap is set too low and it should be loosened up.

I do however agree with what your aiming at and yeah, ideally the cap should be removed and we should have another system in its place which doesn't cap activity. Until that system is coded by "someone" the cap can be useful to hinder distorting ratings by the influence of match-spammers.

tsr

I seem to recall from my chess-playing days that highly ranked players would reach a cap where they wouldn't gain a single point from beating a player with a lot lower ranking. Why isn't it so in the ladder? Or has it to do with that cap being set very high?

We use the exact same system - Elo. And yes, it would eventually happen in here, as it eventually happens in chess. The process is slow because the average active players don't play enough games for it to go faster, nor will they ever play in a pace where such a process will be clearly visible to the normal participant due to the fact that Wesnoth is a strategy games and takes time and effort to play. Wintermute has mentioned this when he spoke about the ladder being young. It is indeed in the world of Elo. =)

What makes this process even slower is this: Imagine I, on purpose, always play the easiest players I can play in order to get 1 p per win/game. Eventually, after a massive amount of games, my rating will change and become higher.

Now, imagine I repeat the process until my rating changes again, and then repeat the process again, and again. As you see, it takes very very long time before I get a rating which forces me to play against a player that has a real chance of beating me. This mentality that I "have" is exactly what has been addressed when speaking of the match-spammer problem. A temporary solution to itm, and crude, is the cap that's enforced. It makes match spamming less attractive and in time it would also make it useless since good players that don't match spam get more points per match in average than a match spammer ever will, meaning, the overall ladder ratings will be more correct.

>Gallifax

@eyerouge: You want quality games and good attitude. I consider it also good attitude to play with anyone, no matter oppoents rank our outcome of match. Low cap ruins that.

Interesting perspective, indeed. I think it's a really tricky question. While I agree that it is good sportsmanship to play everyone, reality tells us that in a competitive environment most players tend to try to profit-maximize while they try to risk-minimize, especially in a system like Elo's. In other words: While you are noble and nice, the average player isn't, he/she is more or less your opposite. Average player will usually try to play people he/she thinks he/she can beat.

I know for a fact that you and a bunch of other veterans have such a nice view of what gaming is and what it entails, and that you're honest in what you write. The sad part is that I also don't believe, for a second, that your attitude would embody the behaviour of the average players on any ladder, in any game that is open to the public and has many users on it. While I still admire what you write, and love your attitude as well as hope that more people will embrace it, I also must question how realistic it is to apply it within a competitive body. Most people aren't as wise as you Gallifax - they wouldn't play anyone on the ladder, because they i.e. have mental issues with losing and/or worship the statistics. With other words, they don't use the ladder as you or I perhaps do - as a personal tool for funs sake, keeping some stats. They see it as something way more serious and try to protect their Elo. This protectionism is widely known, even among chess players.


Also, by implementing say, Wintermutes / chains idea, we would actually force people to play "everyone" - people would be forced to play people they wouldnt play in normal cases becauase they were afraid of losing.

Lastly, I agree that a low cap ruings plenty, even the purpose with the cap to begin with. What would you set the cap to, if I might ask? I'm getting the impression that I'm the only one that thinks the current is sane or feasible...

[Cernunnos]

Hi all,

Loyalty :

I was coming here with the idea to support you somehow eyerouge, understanding the "problems" that can be created by match spammers...

Voice :

I changed my mind.

Just to explain my point, I was playing Warcraft before Wesnoth, there on Battle.net, each match you do ranks you. And, as I don't think too much, I used the same logic playing Wesnoth. So here I am with my 100 games played on the ladder which are nearly the only (but 2vs2, and some let's say 20) games I ever played.

And today I can't play my some ladder games cause I already played 22 in the week. "Please try again tomorrow!"
(I'm noticing a bug here btw, cause you can play how many games you want if you loose them...)

1) + 1 with Pelopidas :

I can enter in weeks where I can play and want to play more than 14 matches and there are many weeks when I cannot play at all.

This is I think, the best argument against a limitation system.
So that's my point, I don't like this "cap" system.

Since I take too much time to write, and two players already posted while I was, I'll just say one last thing, you wrote :

Eyerouge :

Granted you have read most of what has been written on this subject, please come with the following feedback: What kind of settings would you use for the restriction and why?

Another system than pure limitation of the amount of games because of what I read all along this subject.
An alternative way should be discussed :

Limit the amount of games reported by ranked players among provisional players and not the opposite.
Force players to face a better ranked player each X amount of time as said Wintermute.
Do a tournament once in a while so that we'll see who's there because he's good and who's there because he spam.

Many other ways could be disscussed I think,

Best regards.

(you post while i was writing, i go read it, maybe edit)

[Gallifax]

@eyerouge:

uh...oh.. this wasnt about me. Used bad english. I was speaking in general and not about me.
There are quite some players who do that.


And about all other things you image on me. Thx but I feel embarassed by it. I am not jeasus nor perfect. Too much credit for me if you know my meaning.

Anyway I wanted to bring this thoughts of mine to your attention. And again this isnt about me....

[Monster]

Ok perhaps limiting the amount of matches u play in 1 week isn't such a bad idea after all. But I must agree with Cernunnos that it isn't fair if you can and want to paly as much as you can and wish in one week and then in the next you don'te have any time at all. Perhaps some sort of "reservoir" system should be used so that the matches you didn't play in 1 week transfer to the next one; for example you play 7 matches one week and the other 7 you diddn't play transer to the next one in which you can play 21 matches and so on. Spammers will still only be able to play 14 matches a week or they will just have to wait and won't play any matches at all if they were to rack up the number of matches they could play which is not so bad as well

Tournaments are also quite a good idea as it would give a more exact view of who is good and who is not so good.

In any case this cap system still needs a lot of tweaking done in my opinion, as well as my thoughts about it

Reply to Fosprey's Problem: Don't just wait for people to join your match, go out there and challenge them. I doubt anybody will take it personally I myself rarely start a match on my own because I almost always challenge the player I want to play with, problem solved.

[Fosprey]

i would play 2 or 3 games all this days, but as a matter of fact i won't because i can't.
That means that the ladder and proably the wesnoth server will be without me until monday. I'm a match spammer or whatever, so if you think that the ladder is doing better without me, as is doing the wesnoth server, then go for it. but i think several people will be in the same situation, wesnoth is not the most played game in the world and is hard to find opponents, so you are making it so it's harder.
Aditionally i think this won't make change my attitude toward finding an opponent. Before this rule i was aleady trying to cap it so i don't play anything with less than 1450. but i can't really afford to make it much higher. So even if the cap is introduced i will keep playing low ranked opponents, and the reason is, that i already have to wait between 10-20 minutes to get a match, having such paramaters. If i make it any higher say 1650+ i would start to wait something like 30-45 minutes.
and i'm not even talking about putinga 1800+ restriction, i think sometimes i will have to wait an hour.
I alrady have troubles with games taking so long, in fact thats the reason i want a faster time, and that's of gameplay, 45-60 minutes of waiting? no way in hell, even 30 is too much., in fact, i don't like to wait 10-20 minutes.
So i will keep playing low rankered people, because they are who join my game.
So what you are doing is reducing the amount of time players spend in the ladder and/or wesnoth, if you think that's good for the sake of the search of the "best player" go for it, but if this goal is so important, then the problem will remain. I don't think you are actually going to acomplish that goal by this means any better.
If you really want to do that, more than a ladder you need to do some kind of tournament that may be related to the ladder (maybe say, tournament X that only the first 20 players cna play) or some crap like that.

Edit:
sorry but i have to make it personal because i think an issue is disscussed that may well not exist.
Do anyone know someone in the ladder that purposly match spam and , very important REFUSES to play to skilled player. Did some of those so called experts enter one of those so called match spammers and those spammers said "no, i won't play against you"?
If it happens this person should be named, because the case being, this is not happening.
In fact i see several times, myself included, a game hosted, and top ladder players, sitting there in the lobby.
Most of those called match spammers, myself included, host the game, and don't say no to anyone, and if he does, is against too low ranked opponents, not too high.

So if before it was hard for high ranked players to find each other, what will happen now? even less encounters, because high ranked players like me, will be less time in the lobby, wich means less chances to be matched against a top ranked player.
So this does in fact reduce no only the amount of weak vs strong matches, but also reduces the amount of strong vs strong matches, wich there is a lack of .
So i wonder how this actually makes things better

[Pelopidas]

@ Fosprey:
I would not think that anybody really would call you a match spammer, at least not people like me who matched you so far.

To the rest:
- With the cap being present I would suppose that distorted ratings are just a consequence of the rapid increase of the elo constant. If one considers this mathematically the ladder is at the moment stratching to about the double span in points, since the overall equilibrium has been shifted. By this stretching players with more matches are definitely moving upward quicker at the moment than players with fewer matches - a state that will end after new equilibrium is reached. A simple solution to sudden effects might be a slower and step by step stretching of the elo constant, but the effect will always be present. Personally I do not have any problem with this anyway. One has to ensure a good "mixing" of players' matches in the system, nothing else and the whole ratings will be ok.

- A quicker adaption of newcomers' ratings might also reduce the effects. Concerning chess ratings freechess.org for example has a random number assigned to each player. One starts with a very high one and this reduces over time by matches played, as well as it slightly increases during times of inactivity. By this a new player quickly reaches a proper rating and one cannot gain significant points by matching newcomers. This is of course already basically implemented - a refinement of this rule might however also be used to slow down the point motions of frequent players. At this point one might also think about putting players' points to float values instead of integers (one can show only integers) - makes the system a bit smoother and enables fair points even for outlier matches.

- For ensuring the good mixing of players in matching each other I find the idea of obligation to play matches from time to time against a certain class of players ok. I would not favour to simply force a player to do that. One might think about biasing the elo of a player, if he e.g. has not matched any equally strong opponent for a certain amount of matches or even more elegant - increasing his random variable (the amount of point change) in matches against equally good players step by step until he plays and if he still fails to do that after a larger number of matches force him to play against some equally strong opponents. If he really pumped his points artificially he might lose a large amount of points.

Overall it is not really clear if one CAN even pump one's points in the long run. I would suspect this is only the case by a) the current adaption of the ladder to a new state or b) creating artificial players/accounts and inventing matches.

Best,
Pelopidas

[Yogibear]

Hi eyerouge,

here are my 2 cents as well .

Why i think limiting the number of games is not a good idea:
First, as has been said before, it hits all players alike, match spammers and non-match spammers. Then again, i think the limits you proposed are rather at the low end of what is feasible. Now let's assume i want to maximize my rating and my "true ELO" is around 1700. So i start normal playing until i get there. From that point i start to match spam. Let's say i get 2p average per game (which is rather low IMO). That makes 28 points each weak. So i need like 9 weeks to get into the top 10 (current conditions).
IMO, limiting the number of games at best slows down the process of climbing, but is not preventing it.

Some general stuff to get things clear:
I know this is not so popular but i think we have to face it to get things right: The ladder is a tool for competitive gameplay of wesnoth. And the central measure of this tool is the ELO rating you get. Therefore, people expect the ELO rating to represent the skill of a player. Before you bash on me: I know you stated differently in the FAQ, but i am pretty sure that is what peoples expectations are. Ask yourself the negative question: How many people would be playing the ladder if there were no ranking? I guess not many, or we wouldn't have all this discussion about more adequate ELO numbers.
Most every player around here will try to get a better ELO rating and climb up in the ladder, which is a completely fine goal in itself. But because of the expectations of competitive players, we want to guide that process to give us a more or less representative view on the ladder who is a good player and who is not. And we are discussing on how to accomplish that best.
There is also a small amount of players who don't really care too much about their ranking and just play for fun, using the ladder to keep additional statistics.

On a side note: I don't think there is anything wrong with competitive gameplay even if it is to be considered potentially harmful among the developers of wesnoth. And the very best reason i can back this up with is the "community culture" within the ladder, which i consider to be *way* better than outside of it.

Now, here is my suggestion:
To increase your ELO number, you will have to play an opponent that is xxx points maximum below you or higher. To just throw in some numbers: If you ELO is 1650 your opponent has to have 1450 at least to change your own ELO (based on a range of 200). The range could of course be different, this is just an example.
With this, people can play as many games as they want and climb fast if they have the skill to do so.
If you want to play someone below your range you can of course do so. It just doesn't change your ELO if you win (it is debatable if the same applies to losing, which would result in a upper ELO limit as well).

This change is simple and it would remove the problem if the range is chosen accordingly. And yes, there is a downside: It will be harder to find opponents to change your ELO, depending on the width of the range. But i think it addresses the problem better than limiting the number of games.

[grrr]

It seems to me that the term "match cap" is misleading unless you really check for it in both cases: a) player A reports a win, b) player B reports a win vs player A. If not, then what you have is basically a "victory cap". At least that's the impression I got from reading through the description.

Anyways, I like your corrections about the ladder as a "skill measurement" tool, eyerouge.

[Yogibear]

Anyways, I like your corrections about the ladder as a "skill measurement" tool, eyerouge.

Well, i might be completely barking up the wrong tree here, but the intrinsic purpose of the ELO number for me is just that: A skill measurement.

Granted, the ladder might not reflect the reality as the current calculation model and the conditions it is used in can result in inadequate numbers. But talking about the "ideal ladder", i think that is what we are heading for: A ranking that shows the skill of a player.

This is just a wild speculation but i think much of this caution about competitiveness and skill measurement results from the fact, that this is viewed to be neutral or even bad by most of wesnoth's developers and that the potential for conflicts should be kept low. I don't think that is a good thing, though (if that speculation is wrong, never mind ).

[Cernunnos]

To Fosprey : I think it's another "problem" you're talking about. I noticed that top ranked players had difficulties to find opponents. (Which is not my case )

Once Gallifax created a ladder game on the lobby, i asked him if i could join, cause it was like 1pt for him to win and like 40-50 to loose. He told me something like I could join cause nobody would, cause it was in a certain way too risky for them. (You tell me if I'm wrong Gallifax)

So this can't be called match spamming or whatever. You top-ranked-extremely-good-players (tell me if i do to much ) are in a certain way forced to match versus bad players. It is the way it is, nobody else wants to match against you because he'll lose 10pts and go 5 places down, oh no! How can I be able to live with this! I feel so depressed...

On the other hand, bad players think that on a hold-up they can defeat you and gain a lot of points or they don't care too much cause in the end they'll lose like 1 or 2 pts...

[ while posting, Yogi Bear posted, i think the above is related with what he said, i just thought that rather than always play between +200/-200, it could be only once in a while]

As I'm thinking about it, what about classes?
Eyerouge, you implemented this class system which is interesting enough, why not make something like you must play with people of 1 class up/your class/1 class down once in a while? This "while" could depend of your elo, for example 1 about (elo/100) games : 1 about 20 for Gallifax, svek, leocrotta, Pagan, Moonletters, 1 about 19 for Fosprey, Loser...... and so on. (Or another system, that's just for the idea)

The class system is an idea, but anyway playing with a better ranked player sometimes is interesting enough, this way, normal players will do it even not thinking of it, while "spammers" will be forced to have a "real game" once in a while.

Well, this was not to bother you, this is just to try to help finding something relevant enough.

Anyway, a tournament from time to time would reveal the qualities of players and would add some fun too (But this is a lot of work, finding something that solve the problem would already be great)

Thx, (please excuse my english)

Bye

[EDIT]
Important to Eyerouge : while thinking of it, you should make a news as you did while the site was done, so that people mail you their results with date and time, cause maybe some won't notice they can't report before they try to report, and the elo will be changed at the wrong time. After a player who lost made a game with another, with false elo etc... you know what i'm talking about, i hope... bye.