Ladder Site Online...

[nani]

Just because something beneficial (attention drawn to multi-aliases) comes from something wrong (cheating)

Get it please:

As I've already said, using multiple accounts isn't cheating, but a common rule breach...

PLUS it gives the genuinely new person a wrong impression that this is the type of skill level they should expect from a beginner Ladder player.

Starting Ladder doesn't mean starting with wesnoth 1v1 in general. The intention of provisionally ranked players is: You don't know about their skill yet. Everyone, even the other new ladder-players have to deal with that anyways.

And don't even try to argue that they don't play against weaker players! I checked the top ten, and all of them played a majority of their games against underrated players, in some cases nearly 70%!

Do me a favor and count Demogorgon out here explicitly, would you? I'd love to read that, just because...

pretty ridiculous that nani's main account isnt unblocked yet. im dissapointed in you rigor

Dude, you do know that Nani stated he's ok with it, don't you?

Kind of, yes, according to Rigor's plan, I got treated better due to confession, but let's wait...

[Insinuator]

And don't even try to argue that they don't play against weaker players! I checked the top ten, and all of them played a majority of their games against underrated players, in some cases nearly 70%!

Do me a favor and count Demogorgon out here explicitly, would you? I'd love to read that, just because...

Ugh, get over yourself already. The_Black_Sword actually has the highest percentage of games versus underrated players.

As I've already said, using multiple accounts isn't cheating, but a common rule breach...

I don't care what Faello thinks cheating is. He's wrong. This is the dictionary's definition of cheating: "to violate rules or regulations".

Starting Ladder doesn't mean starting with wesnoth 1v1 in general. The intention of provisionally ranked players is: You don't know about their skill yet. Everyone, even the other new ladder-players have to deal with that anyways.

But when you create a second account, you DO know your skill level! You've demonstrated it with your last account and you can use that knowledge against other players who can't tell!

[Scatha]

And don't even try to argue that they don't play against weaker players! I checked the top ten, and all of them played a majority of their games against underrated players, in some cases nearly 70%!

Err, "underrated" is not at all the same as "weaker". A player is underrated if their ELO rating is lower than it should be. For the sake of your own ELO rating, playing against underrated players is bad and playing against overrated players is good.

But when you create a second account, you DO know your skill level! You've demonstrated it with your last account and you can use that knowledge against other players who can't tell!

I don't really see how you can "use that knowledge", but I'd appreciate it if you could explain. I'll just note that while you're provisional in your rating (first few games) other people stand to lose very little by playing against you.

[Velensk]

What is more, if you are the top ranked player, then 100% of your games will be against underrated players.

The only reason that the person at the very top of the list would have any games against people ranked below him would be that he played games before he was at the top of the list.

You would have to be very, selective to be in the top 10 and not play most of your games against underrated players. Personally, I don't see the point in being that selective in any sense.

[Dauntless]

just wanted to comment on that too
As velensk says, its quite hard to play vs "overrated" players when you are on the top

[Scatha]

Sorry, I think I didn't explain this very well in my post just now. Being underrated has nothing to do with being weaker. If a strong player joins the ladder then they might have a real strength of 2000, say; until they get up to that they will be underrated. Vice-versa if a weak player joins the ladder and has a real strength of 1200 then from their starting point at 1500 until they get down to 1200 they will be overrated.

(Edit: apart from that the point Velensk and Dauntless is making is a good one, of course -- it just isn't actually relevant here because the statistic Insinuator found wasn't what he or she insinuated it was.)

[Dauntless]

You are right Scatha, however in the Opposition table on the ladder site the terms underrated and overrated are used in the way Insinuator used them, that is people with lower/higher elo than you, so I stuck to the terms and just put them into " " to make it more clear...

[Scatha]

Dauntless: I'll have egg on my face over this if I'm wrong, but I'm pretty sure that that page of the ladder uses them in the way that I describe, not the way Insinuator uses them. Off to bed now, so I can't provide evidence, but have a look at which players were underrated/overrated and I think you'll see it can't be the meaning you think.

Edit: OK, if you want to look at the strength of their opposition, look at the average ELO points the players won/lost for a victory/defeat. Unsurprisingly (for the reasons you and Velensk point out) the top players played against on average weaker players, by a sizeable but generally not a huge margin.

[nani]

I don't care what Faello thinks cheating is. He's wrong. This is the dictionary's definition of cheating: "to violate rules or regulations".

"Cheating refers to the breaking of rules to gain advantage in a competitive situation." - Wikipedia
I don't see where I can get advantage while breaking this so called "rule" the way I did. (No need to start a vicious circle here...)
Playing Wesnoth without fog would be such an advantage while breaking a rule of the game itself e.g.

But when you create a second account, you DO know your skill level! You've demonstrated it with your last account and you can use that knowledge against other players who can't tell!

Who cares if I know, we were talking about the others who shouldn't expect to know their enemy's ranking.
I can use this knowledge? For what? Stealing points from others maybe? Must've done that quite uneffective...
Being provisional, the enemy's didn't lose many points, and afterwards, I think I had a record that didn't allow me to mislead anyone.

We should stop arguing here since it obviously doesn't lead to anything.
I think I can safely claim you wouldn't even admit if I was correct after failing my little test:
You couldn't just say: Yes, you're right, Demo wasn't like this, no, instead you preferred browsing the database for probably the only one with a higher percentage of games vs underrated players.

Useless discussion, it'll never end, I should be out as well.

[Rigor]

ill tell u a little story now. today i woke up early, went to work, worked. then i found it was such a beautiful weather outside that it would be great to go to the beach i have now that i am abroad. after coming back from the beach i had an incredibly tasty dinner with my new friends from the dorm and my belly was about to burst. when i came back to my room it was very,very,very late already and i turned on the computer to check out the progress of this discussion. my plans for tomorrow? im going to wake up early,...

i would like to come to the most interesting part now because i am waiting and waiting...and it is not going to happen it seems. you know, i learned from experience that its a better idea to let the other guy say something first, but if you decide to just play dead, or avoid the topic, this is also an answer for me. and please dont tell me some real life events keep YOU busy, when i am writing at such a late hour during working days only to get it finally over with.

HINTS:

However, I really think that players like leocrotta or Dauntless (just to name famous ones, everything said here also applies to all other ladder players) would certainly reveal their aliases (as they already did in this thread), because I also expect them to want to create a nice ladder atmosphere (which isn't possible if you don't say who you are and someone discovers the use of the alias...).

I'm favor of Rigors 2 ladders and public aliases.

I'm super curious.

Dauntless, what is your alias?

[...] Another question is crossing my mind, too. Who of you did discuss with his alias here?

Just cut the crap and send Rigor your second accounts to delete & don't do it ever again.

It's not really easy to 100% confirm that 2 accounts belong to the same player. So the admins can't just run around and paint everyone red who could be an alias...

hope the other players with aliases confess and we discover the other ones.

What we could do is just ask the players to be honest, tell us all of their accounts and ban all accounts but their best.

Spot on, on all accounts!

Your idea seems quite good to me, it's essentially disclosing which accounts are aliases, which I said I am for.

you just say "i have one alias but im not telling" and thats it, you kidding me? everybody else doing this kind of thing before you had the same choice, tell or dont, some with exciting public viewing (soul_steven), some none at all (neki) - what i would prefer to do is to just hear from you very soon which one (or 10, very funny Q) you use and then act accordingly. until then i dont see a reason why you should be able to continue playing as if nothing happened. could you alias-players please stop playing ladder games until this matter is resolved?

[Insinuator]

A player is underrated if their ELO rating is lower than it should be.

You're right in a general sense but wrong in a technical sense. The statistic for playing against "Underrated" players merely rates one player's elo versus another's. Thus, a player with 2000 elo playing versus one with 1900 is playing an "Underrated" player. If the 2000 elo player then plays a 2100 player, he'd be playing against an "Overrated" player. Simple as that.

But when you create a second account, you DO know your skill level! You've demonstrated it with your last account and you can use that knowledge against other players who can't tell!

I don't really see how you can "use that knowledge", but I'd appreciate it if you could explain. I'll just note that while you're provisional in your rating (first few games) other people stand to lose very little by playing against you.

Ok, another illustration using a fictitious Ladder player named Roy: Roy plays Ladder for 3 years and gets up to 2150 elo. Now, he grows tired of playing that account for whatever reason and decides to create a new one. This account, named Cheeta, is a provisional account starting with 1500 elo. Now, another player, little Johnny, decides that it's high time he joins the Ladder. He's played Isar's Cross for a few months and shows quite the aptitude. So he joins and starts with 1500 elo.

Now Cheeta is a Ladder veteran with a very low elo account. He spies little Johnny's game and notices he's a new player. Looking for an easy boost to his ego, he beats Johnny in several games, improving his elo by leaps and bounds without much effort. Now normally, little Johnny might object to playing a 2150 elo player. He's a modest guy and prefers to play with those of a similar skill set. So he doesn't object to playing Cheeta. Is it fair to Johnny that Cheeta is deliberately misrepresenting his skill? Does it matter to Johnny that the Ladder will "balance" itself out eventually?

That is one way to "use your knowledge" to gain elo quickly and easily. And it is damaging, particularly to new players.

[Kolbur]

I just want drop in that Insinuator is wrong and Scatha is right about "underrated" in the opposition table of the player profile.
A player was underrated at the time of the report of a game when he had a lower elo than he has currently. It has nothing to do with having a lower elo than your opponent. It becomes quite obvious if you look more closely at it. There's no need to argue about this.
And yes, playing vs underrated players is generally bad for your elo, though most people are underrated anyway since they are still climbing and there is some inflation going on over time.

Also I support Rigor's request.

[Oook]

Insinuator:

Spoiler:

Firstly, please don't shout, we're perfectly capable of reading your full message. If you want to highlight any particular point, I'd suggest italics, bold, or simply making it clear from context. The posting guidelines suggest you don't type in all-caps, and I know how keen you are about rules being followed. Hope this helps

Oh, my bad. I guess I thought ROBBING A BANK isn't "widespread rulebreaking" that needs to have "attention drawn" to it. Please, Oook, link me some more pages on economic policy, racketeering, and murder cases, because those really demonstrate how confession makes it OK. I know what he was trying to say. But the people in those cases should not be labeled as heroes, nor should they escape punishment. They did not commit those crimes to "draw attention to widespread rulebreaking". They happened to be caught and then reduced their sentence in return for cooperation. Those were negotiated terms, not tender-hearted confessions.

Nani suggested that people who assist the authorities by confessing early tend to receive more lenient treatment, as this encourages people to come forward with evidence. You said this was unlikely to happen in real life. I linked to a variety of examples of this happening in real life. I don't see nani or myself here arguing that Rigor should see nani as a blameless hero of the situation. If you're arguing against that, please provide the relevant quote where one of us said that?

This use of incentives for speaking out is especially relevant in cases like this where there was no way of identifying use of aliases without people speaking out. If you really have a problem with this kind of policy, you'll probably be too worried by what your government gets up to to care about the wesnoth ladder To improve the bank analogy, consider it not as armed theft, but, say, widespread misuse of bank software.

Just because something beneficial (attention drawn to multi-aliases) comes from something wrong (cheating)

Yeah, it's really not helpful to call it cheating. Just because something breaks a rule, doesn't mean people automatically start referring to it as cheating. For example, in football / cricket / ice hockey etc. players frequently commit fouls, that are against the rules, but are not regarded by anyone as 'cheating'. Or another example: you've already gone against the posting guidelines, but I'm not claiming you're cheating at this discussion In short: cheating implies rules have been broken, but rules being broken does not necessarily imply cheating.

Cheating tends to be reserved for more serious offenses. In this case, if you showed that a player had set up a second account, and claimed victories against it with their main account, that would be cheating. But there is no evidence at all that that has happened.

So now there are "conflicts of interest"? For someone who identifys himself as very trusting in the next breath, you seem oddly suspicious of Pelopidas now. This sets a double standard. Why should an admin have to report his alias if another player doesn't? Particularly if said alias is inactive after the second is created, obviously indicating nothing more innocous than a name change.

In case you weren't aware, a conflict of interest is simply a situation where a person with power has to make decisions which will affect them in a personal or other capacity. It does not mean that they have misused their power, or made poor decisions as a result, merely that there is an overlap of professional and personal interests. If they were to take advantage of their position, that would be termed corruption, or abuse of power. I thought I was very clear that I was not accusing Pelopidas of this - I'm confident he has not misused his admin capacity. My comment about trust referred to the way people such as WhiteMage and yourself seem very reluctant to trust people not to do wrong

However, when a conflict of interests exists, it is useful for the person concerned to make it obvious that they have acted fairly, so as to avoid future malicious claims of corruption. See the section on Disclosure on the wiki page.

As for the use aliases part, you've got it the wrong way round. There's no problem with admins having old aliases that are not used any more. But they should publicly report / acknowledge any current aliases they use, for the reasons outlined above. The reason this is more necessary for admin is the greater power that admins have - they get to settle disputes between players, so must be seen to be open and impartial in this duty.

1. What evidence are you expecting to see? Only a moron would make a second alias and play it all the time then beat it. That's something that would be noticeable. It would be far smarter to create half a dozen accounts, keep them at a moderate level and beat one every once in while. Spreading the games out like this would be virtually impossible to track and just because someone confesses to one prominent alias doesn't mean they don't have a bunch of little ones out there.

Well, if they're abusing one or two aliases, then it should be reasonably obvious, as you say, or else have very little effect. If you're claiming that someone has a large number of aliases that they are using to feed points directly to their main account, that's a very serious accusation. The onus is, with good reason, on the person making such an accusation to provide evidence for it. If you can't provide any evidence at all that it took place, you should probably stop making those claims.

2. The elo is self-correcting over the ENTIRE Ladder, yes. But it is not so for one person. Imagine a strong player such as Pumpkin, who is accused of being a second alias, playing five games against an actually new person. Now, they both start at 1500, but if Pumpkin wins every time, he's boosted himself past 1600 in a breeze and smashed the new guy into the 1300s. That is a hard place to recover from, PLUS it gives the genuinely new person a wrong impression that this is the type of skill level they should expect from a beginner Ladder player. That could be very discouraging and not representative.

No, elo is self correcting in individual cases as well. If you lost too many points playing someone who was greatly under rated, then you will be slightly under rated, so will gain slightly more points than usual over your next few games, until you're back to your 'correct' rating. As I said in the message you quoted:

Assuming you don't spend most of your time playing players who are greatly under-ranked, your rating will still be more or less correct.

If you want to avoid that, whether from aliases, strong people joining the ladder, or simply a rapidly improving new player, it's easy. Don't play too many of your matches against people who's rating is rapidly increasing.

As for your example: if you expect a new player on the ladder to have a strength of 1500, you fail to understand how the ladder ratings work. 1500 is simply the arbitrary value assigned to a new player, and the 10 game period is designed to help them quickly get closer to a more accurate one. There are plenty of new players (not aliases) who join the ladder and find their strength was already well above 1500. As has been pointed out, the first 10 games you play won't have that much effect on other's ratings. I agree that it's not an ideal system, hence my preference for a Glicko one, but it's not a problem you can blain aliases for.

It is far more important to consider the trees than the forest, in this instance. It doesn't really matter what the overall picture of the Ladder is (though, off-hand, I would argue that having two people hold the top four slots shows a skewed picture). It's more important to consider the individual players. If the strongest players use these multiple accounts, especially against new players, they're artificially suppressing the elo of lesser players. And don't even try to argue that they don't play against weaker players! I checked the top ten, and all of them played a majority of their games against underrated players, in some cases nearly 70%!

As I pointed out above, players will only lose points on average to accounts that are still gaining rating quickly. So any points lost there will be regained once they play several games against more settled accounts. As for the comment about underrated players, I think enough people have set you straight on that already

WhiteMage:

Spoiler:

I get your point, and agree that it's not implausible that some vote stuffing took place (though see Huumy's post for other suggestions). But that was not all you claimed:

Apparently some aliases feel very strong about the issue of keeping aliases. . When people are willing to go to this extreme, I am not sure how will you re-establish trust on people using their aliases for honorable reasons only.

There you were claiming it was done by people who use aliases - I was simply pointing out there was currently no more evidence of this than that you were responsible for the vote stuffing

With regards to your stats discussion with Huumy: no offence, but the last stats example you used in this thread was horribly simplified, with so much relevant context removed that your conclusion was pretty meaningless. So in this one, it would help if you made explicit your assumptions, and why you think they are reasonable.

If you need statistical evidence, that is pretty easy to do with some p value calculation and standard alpha value (0.01) giving you probability that this happened by pure chance.

As Huumy pointed out, it's really not as simply as you suggest here, unless you make a lot of convenient (and unrealistic) assumptions.

Huumy,
You have some partly legitimate questions, and scenarios, I agree. In statistics we call them: lurking variables and correlated data. Let me assure you that there are well accepted statistical methods to deal with these common problems in real life and I happen to know them in depth. Since you do not know this field well, here are a few topics I recommend you read some books and articles about before you engage in deep discussion: cluster analysis, statistical pattern analysis, multivariate statistics, Occam’s razor.

WhiteMage: I agree that you can use those techniques to analyze this kind of situation, but you need a lot of additional data to come to any meaningful conclusion. As Huumy suggested, you'd need to estimate probabilities of such a cluster being caused by other factors, and make sure you took all reasonable scenarios into account. I think it would get pretty complex, and you'd be unlikely to reject the null, especially at the 1% level. Since you were the one suggesting that this analysis would be reasonably doable, and useful, I'd be interested in what formulae you'd be using, with relevant parameters / assumptions. Thanks

As for Occam's razor, that's generally more useful as a heuristic, than for providing any actual input when calculating confidence intervals. A recognised problem with it in cases such as this is that the complexity of the assumptions is rather subjective - is there a meaningful difference in complexity between the assumptions "vote stuffing was done by an alias supporter", "vote stuffing was done by a non-alias supporter", "someone told like-minded friends about the poll"?

[Huumy]

you just say "i have one alias but im not telling" and thats it, you kidding me? everybody else doing this kind of thing before you had the same choice, tell or dont, some with exciting public viewing (soul_steven), some none at all (neki) - what i would prefer to do is to just hear from you very soon which one (or 10, very funny Q) you use and then act accordingly. until then i dont see a reason why you should be able to continue playing as if nothing happened. could you alias-players please stop playing ladder games until this matter is resolved?

I'm not sure Rigor if you are suggesting that people who you quoted have multiple aliases?
Anyway just to be clear, I don't have multiple aliases I never had.

[Faello]

I don't care what Faello thinks cheating is. He's wrong. This is the dictionary's definition of cheating: "to violate rules or regulations".

1)This rule was breached multiple times by multiple persons, even by the ladder admins which validated this breach several times as Pelopidas already stated [he's also an example of such a validation]. Besides playing under different nickname, it's hard to accuse players that used aliases of anything else than using aliases - most of them probably played the games as they'd normally played them, gaining positive sporstmanship.

The main flaw of this breach is that players that used more than 1 account falsified the ladder stats [from the "legal" point of view we could say they did it intentionally, because it was easy to predict].

I've stopped discussing "does it matter or not" topic, because it's perfectly logical to me that ladder was created to rank players - if they're playing under multiple accounts they're using ladder only for it's auxiliary purpose of finding skilled opponents but that's it => that was the reason why I wasn't interested in repeating myself over and over again because I felt like I was talking to the wall.

No serious ranking in any game allows the same person to be ranked multiple times.

That fact - actually a nail to the coffin of this "is it ok or not?" discussion - wasn't recognized by some players not only in this discussion, but also since the beginning of their ladder activity - and that's also why we are discussing this problem today.

Majority of this rule breaches cases probably can't be called cheating, because cheating would be a delibarate activity that harms the other players directly (even if it would be veiled by using the alias account, like "hunting" the better ranked player with our alias accounts) - still such a cases probably already happened BUT if we're considering ONLY the action of using another account to play ladder games, it's just a rule breach.

In this case, to call it cheating, you'd have to prove that somebody used his ladder account in malignant way to harm others - in civilized society we've something called "presumption of innocence" and this presumption protects ppl from harming their good reputation. At this moment we've only a proof of rule breaching from nani and some other players, but NOBODY presented a valid proof that somebody used this rule breach to cheat - probability says that somebody probably has already used this rule breach like that but without a proof we cannot accuse anybody of cheating.

2)Considering all of the above, banning ONLY nani was unobjective. Either ladder admins are banning EVERYONE that did the similar thing or they should abstain from banning anyone. Considering ladder mechanics, banning nani looked like he was some scapegoat, only because he was the "most visible" example of this rule breaching. It reminds me a bit of a medieval way of hunting crimes:

"We've no idea who committed this crime, we've also no means to effectively pursue the perpetrators, so let's punish cruelly the most likely perpetrator and hope that the rest will tremble in fear hard enough to never do it again"

Rules about using multiple ladder accounts were not precised well enough and if they were not precised, ladder admins should react at first sight of such actions [it's hard for me to believe that ladder admins heard about multiple accounts for the first time just couple weeks ago, while ladder is online since...2 - 3 years?]

Furthermore, ladder admins should set an example for other ladder players how to act and behave on the ladder - I find Pelopidas/Coriolan behaviour somewhat inadequate considering his admin position - his Pelopidas account is still unblocked in spite of his ladder admin position. It's inactive right now, but it can mean as well that even the ladder admin could use second alias while his first account was still ranked, which AGAIN proves that banning nani was a very harsh treatment. I'm not even talking about ingame behaviour <I guess using alias sometimes let your dark side on the loose >

Considering the ladder management area of this case, I see here the classic "split-responsibility" management problems [if 1 person would be responsible for the ladder rules, they'd be probably more precised than they're now].

Considering the multiple accounts players, I see here the "inflation" of their activity on the ladder, maybe even some form of addiction to the ladder games (it's easier to obtain the points on the ladder on fresh accounts and every loss in some high-level game can be more stressful). It's probably also nice thing to feel that our victory in wesnoth game is NOT just a waste of time and it will give us some points or something, but that's just a proof that some players should think a bit about their activity on the ladder, because most of the arguments that pro-multiple account players presented, like for ex. "I cannot play more than 1 ladder game at the same time" [which is untrue] were totally missing the point of this discussion [ladder objectivity//ladder ranking//ladder players trust to the ranking and other players] and were mostly concentrated on the auxiliary ladder purpose - playing a good games - which was already possible under current rules...

All in all, I support Rigor propositions, but I'd still count on the second/more serious ladder.