There's no delicate way to word this

Several people have mentioned that 5five appears to be playing LOA perfectly. I also notice some people have started resigning at the earliest possible opportunity against him, so there's obviously some feeling that he cheats. However, is there any hard evidence? I wouldn't know perfect LOA play from a hole in the ground, so can't evaluate claims that he's just using a computer program myself.

He's clearly taking the piss, playing duplicate games

here and here, but that's not exactly damning.

There isn't perfect play absent the game being solved, but the computer is far above all humans apart from 5five apparently.

That being said if you examine his play he is either using the top computer or he is some kind of savant. Quite simply if he is genuine he would have to be a god of LOA and should call up James Randi. Not only is he playing the same move as the computer, that has slaughtered multiple human World champion he is doing it moving quickly. I checked 3 games and found his play was identical to the computer after the opening (the opening move will vary because of symmetry and opening choice). The opening few move are also excluded as matching the computer with these often happens without the play being suspicious.

Occams Razor says the guy is using the computer in all the games I checked my suspicions were backed up. With practice you can emulate much of what the computer does, but human player miss stuff particularly when moving fast as he seem to do.

If you know LOA then you know it that his play would need a very good explanation. He has a perfect record at chess could be legit (enough players blunder that with a good opening and good concentration you can do well) but I'd suspect that if he is doing it at one game he's doing it at all. So far his LOA play is damning in my eyes, that I didn't need to go any farther and examine the rest of his play in other games in detail.

I was "playing" against 5five (I put playing between quotation marks because I have no idea how to play LOA) and I told him about this thread. His reply was "dont [sic] care about it."

So 5five isn't denying it?

His LOA rating is 1635, has finished only 11 games and the best of his opponents had 1633 rating. I see no game resigned against him!

I see no evidence of cheating whatsoever.

Ypercube - you may not have seen any LOA game resigned against him, but I'll think you'll find he plays other games too. When people feel strongly enough to not play the opponent at one game, you'll generally find that they won't play them at any game at all. However, on probing the resigner, it appears that was more based on gut feel than hard evidence.

KingofhtebeI - I downloaded MIA, and found that, apart from the the first few moves, many of his games were practically identical to what MIA would do. Each game had maybe one or two moves where he played differently, but in those cases MIA had considered his moves as the 2nd best move whilst doing its searches, so it could be just a slightly different parametrisation or engine version. (I checked the same 3 games as you - we might have different versions of MIA.)

So I still wouldn't call it "damning" quite yet. However, I think it's not "no evidence whatsoever" either.

That might be down to the search or different hardware if I play g1-g4 and then g8-g5 it probably is just the almost same as if I play g8-g5 and then g1-g4 yes there are tiny difference but the lines may just be transpostions unless move orders makes a difference.

When I check if the moves matched they did, I think different hardware could account for discrepancies. If you are using better hardware than him/her you would reach higher search depths sometime and where there positions that are a close call the machine may change its mind.

Might be worth examining Mating(connecting) moves and leading up to mate, MIA using a board evaluation and will seem to choose the same mate in 1,2 or 3 even when there are many different ones available. Occassionally you might expect the human to play the same move as the computer but in these instance the human has no programming to go for a higher board evaluation and you would expect variation in what moves is played leading up to mate.

To me this is beyond any reasonable doubt to me that these games were played by MIA.

Not only is he playing exceeding well he seems to move fast and without error. Had this thread not been posted I would have suggested playing predict his move of course it would still be worthwhile to see if his games continue to demonstrate a continued likeness to that of MIA

If 5five has an explanation I would love to hear it.

Human simply can not consistantly play like the computer the game is too complex that eventually the human will overlook something.

If you look over his other completed games that were not resigned I would expect to see further corroboration.

Everything about the play of 5five is just too good to be true and I've seen some phenomenally good games players!

If this was a comparison of chessplay to fritz/rybka on playchess server rather than MIA to it would lead to a deletion of rating and an announcement that you had been caught cheating.

MIA is further ahead of human ability than Rybka is so it is less credible that someone could match MIA so consistently.

The apparently world's best LOA player cares enough to be in this site's championship and monthly cups but doesn't care that are casting aspersions on his play. If I believed 5five was genuine I would seek an appearance fee and prize money sufficient to guarantee his attendance at an offline LOA event.

What would you see as damning fatphil? If machine identical or even 100% correlation to the top 2 choice of the machine over several game isn't good enough then I guess we will never have enough evidence to say it is damning.

Look at it this way I challenge anyone to find an example of him making a move the computer considers bad in a game where there is computer dominant prior to this thread being started.

No games resigned, really? A win in 1 move or nought moves is a good indicator that either the opponent is unwilling to play, careless about time manage, a sock puppet or deliberately being annoying. Since 4 different players have lost in the minimum possible time I would say it likely that at least 1 or 2 of them did so to avoid playing. These were not players I have seen resigning against everyone or timing out a lot!

Those who resigned had sound gut instinct in this case, but if you don't have the software or know of the software then you can't say he is using this software. His/her misfortune was to come up against someone actually knew enough to see what happening and back up that suspicion. Just because don't know how the magic trick was done doesn't mean you should believe the conjurer if he sincerely tries to claim he is really a psychic. Eventually a James Randi will come along and expose the fraud.

If all the MIA moves were clearly obvious to someone skilled at the game, then doing the same as MIA would not be damning, it would be obvious. I am not in a position to judge how obvious the moves were, as I'm not sufficiently knowledgable in LOA.

Were he to be seen beating players skillful enough to see clearly obvious responses, then that would imply those responses aren't clearly obvious. Were an expert in the game to say that his moves are not obvious but demonstrate deep insight into the game, that would alter things too.

A very skillful reversi player was unfortunately accused of cheating, the accuser was 100% he was correct; it took reports from others who had met the accused in real life at international tournaments to persuade the accuser of his error. I would rather such embarassments not be repeated.

MIA moves are not obvious or it would not crush the human world champions. Put any human in the world up against MIA in a match they can start every game I would predict the human would get beaten in every game and don't think anyone who would reasonably be considered an expert at LOA would disagree with that assessment.

If he could demonstrate this ability was entirely human it would be something for James Randi as it would demonstrate a level of play that hasn't come close to being generated. Pick random games of the top ranked players on this site there will be moves that MIA considers wrong not 2nd choice but so bad as to actually lose the game if played out by machine.

His play doesn't demonstrate deep insight or any insight a monkey could be trained to copy the moves from one interface to the other.

When I said I could play predict a move I meant you start a game aginst him play any random move after the open and compare his reponse to that of a computer and then compare other top players to a computer and see how ofter they match.

How knowledgeable was the accuser, I haven't come across anyone who has even claimed to have beaten MIA, let alone shown they had the potential to do it.

I would say my knowledge of LOA is probably greater than this guy who complained has of reversi. The fact I have mixed with some of the best board games players in the world and have seen just how good a legit player can be. Some of those world class players play on this site and I would expect the level of play unassisted by computers would be higher that in live play through greater thinking time and the option to set up a board and analyse.

There is no embarrassment to be repeated, I know personally just how good the top humans in the real world have played and this guy is putting in a performance that would be as surprising to me as a seeing 5 second 100m run by a human.

Program a 9x9 interface for this site and see if this guy will play, I doubt it while there isn't 9x9 MIA!!

Oh this innocent reversi player actually claimed innocence and made a reasoned rebuttal of any evidence presented I presume?

Several of the MIA moves in the games I stepped through were completely obvious to me, even as a novice. Therefore one cannot universally say "MIA moves are not obvious".

You might like to say that your knowledge of LOA is probably greater than that of the accuser in the reversi equivalent, of that I have no doubt. However, you say that whilst simultaneously not knowing who the reversi accuser was, and also after having not yet demonstrated your expertise at LOA, with only a few dozen games under your belt, most against novices. Even I have beaten better LOA players than you and I claim little more than superficial familiarity with the game. Which makes it currently a fairly hollow claim. It may be true, but whilst you are unable to grasp the difference between that which is true and that which is provable, this thread isn't going to go anywhere.

The innocent reversi player didn't even see the accusations when they were first made, he was here to play games, not waste time on fora. People had already adequately defended him before he was brought into the thread.

Well the championship hasn't started I played those who entered monthly cups so can't control who I have faced. Online anyway play is not sufficient proof of ability other than a game where the ai is inferior (even that would be assuming no MITM tricks).

Several of the move were obvious, several isn't all, have you played against MIA? Just becausee the best most/very good move is sometimes obvious does not mean you can spot all best moves.

Since I never competed against this guy you've beaten I couldn't say whether not he is better than me but I wouldn't welcome the chance to face him or anyone else in person I rarely lose in the real world and that's what matters to me and have no fear of any opposition. You have no basis to say this person is better than me.

The reason I was confident was the fact that my knowledge comparably was better to the complainer he/she made a complaint against a guy who was a well known in offline competition and this person played in a credible way or people wouldn't have believed falsely that this play has to be cheat.

May I point out that I was able to identify my presumed source of his play and that MIA is not perfect but merely incredibly strong. Also I had to actually be suspicious enough I don't have the time to check every 1600 rated player (at the time 1550ish) for sign of computer use.

No human plays error free at very complex games or matches the machine 100%! Compare chess grandmasters to rybka.

What it comes down to is how likely it is that a human could play like he is doing. If I thought there was much doubt I could add his chess games to rybka to compare with, if rybka actually decided that his play had significant errors at least for chess it would suggest there is a decent chance of human play.

Statistical comparison with engine is the only likely evidence that will ever likely be available absent the player actually confessing. Some might argue that the fact he didn't deny anything could be seen as tantamount to confession.

The closest thing to a defender this guy has is someone saying that matching a computer is no evidence and someone who acknowledges that there is evidence, but due to not being very familiar with the game and software cannot say whether the evidence is sufficient and even his best defender ended his first post suggested the accused is taking the piss.

I just realized this discussion is very academic because Lines of Action is one of the games dominated by our resident cheater with the thousand aliases. In time will find out if:

a. 5five is a great player who will be crushed by our resident cheater thus proving he's an honest player

or

b. 5five is just another alias of our resident cheater

or

c. 5five is a separate cheater and will battle it out with our resident cheater to see which cheater has the strongest program

By the way congratulations to our resident cheater with the thousand aliases for reaching six of the top seven spots in the just completed reversi championship.

Victim of Chomskian elision - "I have beaten better LOA players than you" was supposed to be interpreted as "I have beaten better LOA players than you have". I never meant to imply that I could place any upper bound on your skill level, which I can't. (OK, I can, it's less than MIA's!) Apologies for the misunderstanding.

kingofthebesl, my advice is to relax about 5five. He might be cheating, but consider that you might be wrong.

Better to spend some energy of making cheating less attractive. One thought I had was to require every player when they start the Championship to certify they are not using computer assistance. This would discourage players who casually use computer assistance to desist or exit the Championship. Policing and banning nicks or IP addresses is a lot more effort than people think.

I should also point out that the definition of "cheating" varies from person to person. For example, I consider using any program to assist in any way once the game has started to be cheating, but I know this is an unacceptably narrow defintion to many.

I should also think that cheaters come and go. Spending hours of your life mechanically entering computer moves in game after game will quickly become boring and pointless once the strength of the program you use gives you too much of an advantage.

On a side note, and I see this has been discussed elsewhere, I second Ray Garrison's suggestion that computer or bot players should be clearly, and uniformly, identified. Ray is a very strong chess player and has pointed out that on the ICC computer players have " (C)" (space + bracketed C) after their name. The current naming conventions employed here on LG are insufficiently clear in my opinion.

quartastella

The resident cheat if you are referring to the Jamaican crew, the JaMaIcAn appears to be a man in the middle and sock puppeteer. The Jamaican appears to actually be beatable so I would expect 5five to win. Of course if they both use a computer then black will win around 88% of the white 11% and draw around 1% of the time assuming similiar hardware if they blindly follow the machine.

The JaMaiCaN, EnCaL etc all use the mix of capital and lower case so I'd say this is someone separate.

Of course if those that were beating the Jamaican were using a program then the fact he has lost would not remove him/her from suspicion of using the program as the program will naturally lose to itself. But I've only been inclined to look into players who don't lose or only lose in one move. If someone is a sockpuppet master without the necessary knowledge/assistance to back it up and beat the good players then you can deny them victory by outplaying them.

Fatphil

Ah I see and I see why you would claim Adam as a better than anyone I've faced on this site his rating has taken quite a drop since that game.

Anyone who is able to come to London towards the end of august will get a chance to face some of the best LOA players and possibly become champion if you have enough game.

alain

What it would mean for me to be wrong that is that someone is capable of play that matches the computers candidate moves and plays seemingly without obvious error despite moving very quickly. Being proved wrong would be amazing and would be really exciting to see.

This is as good as the evidence is going to get when it comes to LOA IMO. If we can't be sure enough this is the computer that is playing, then the game possibly dies online (it already seems to have JaMaiCanN pRoBlEm from what I have gathered!!) for the same reason 6x7 connect 4 dies if a human can play perfectly then the game is too easy and the best players will get whatever the theoretically correct result is when they meet each other. This is at odds with the response I got suggesting 9x9, 10x10, 11x11 loa...etc as potential new variants which was a suggestion the game is "complex enough" and from someone who is one of this site top players if I'm not mistaken!

The reason to raise the possibility I'm wrong is because I'm human I make mistakes. Which is exactly what 5five hasn't shown, any errors in his play seeming or move inconsistent with a computer what a computer might chose.

I'm be willing to change my mind if there is a reasonable alternative explanation for the play other than machine assistance. Clearly 5five isn't guilty of a crime or even of violating any rule of this site so really the only issue is do we as individuals think his play is very probably that of a computer or not and how that makes us chose to interact with 5five, if at all.

As mentioned in the cheating threads elsewhere there isn't a rules page other than how the each game is played.

I Suggested that those using assistance/sockpuppetry be labelled (C), stripped of cups and rating. Don't Think anyone who is fair would disagree with honest description of how someone plays a game and that people should play the way they expect others to play. So I think I would also go along with what Mr. Garrison suggested.

The idea of someone using the computer to play is more amusing than stressful in this instance.

No, wait a minute, the resident cheater with a thousand aliases plays man-in-the-middle only in games where there are no computer programs available to cheat or where computer programs are weak. In other games he is a full-time cheater "winning" 100% of the games he finishes. Just check the just finished top reversi league where he owned six different aliases. Magic Jim managed to finish tied for first because - as he explained on the reversi board - he used a computer program to countercheat.

By the way, the resident cheater with a thousand aliases changed one of his nicks into aLiAs_nr_1000. I consider this a personal homage to the people like us who helped unmasked this poor soul.

aLiAs_Nr_1000 sound like part of the so called JaMaIcAn crew, Fatphil could tell you more about them.

BTW is magic Jim seen as justified it his counter-cheating. There is something appealing in denying victory to the dishonest player of course when is it acceptible. The difference between machine and human play has been compared to an adult beating up a young child, no-one wants to be the child, but it would be nice to fine out you were beating up the child either.

Note that this justiification doesn't apply to 5five, if a respectable player to most people suggest that UsEr DaTa, PoWeR UsEr...etcera were to be beaten by any means necessary, then this player should still not be using assistance against those who are not playing unethically. If the computer user is truely has noble intent he should explain to the innocent player what is going on, so they can resign before they have rating jeopardy or if they don't wish to resign then the game should be played unassisted.

There is a problem the MITMer MAY see themself as justified in countering the use of machine and the Machine user in countering the MITM, if true there is a goal beyond being top of the rating list and winning a cup! Each is now justified in stopping an undeserving cheat in there own mind, this sense of morality allows them to not feel guilty about steamrollering innocent players and overcome the monotony of being a postman that alain raised.

we feebly tried to persuade Magic Jim to resign, and make the cheater's victory a hollow one, but he said that instead he'd just see if he could play "advanced Reversi" better than the cheaters, and we all backed him, for a non-negligible value of 'we'.

Well any victory that is gained unfairly is hollow.

Better honest cheating than dishonest "cheating" wins I guess.

Another possible thing make the monthly cup never end by have 2 or more honest players have a never ending game when a cheater enters a heat, once the cheater realises that they won't get given the Monthly cup win or even the heat they will lose interest hopefully, and hopefully that will free up the MC win for for someone who uses a brain rather than a cpu or a sockpuppet.

FYI: pedropajarito just indicated that 5five cheats in the reversi forum.

There is another account created the same month as 5five with perfect stats in games that have publicly available super strong AI I have not done any experiments with computers to see if this someone is computer like or just someone who has faced weak opposition.

simply i can undestand whats funny in do cheating.

I expect that for QUAKE, WOW, or that kind of massive games...

that any imbecil can play.

But Cheat on Abstract Strategy BoardGames ???

Thats weird..

The subject of why people "cheat" (I would use unethical play until we get rules on this site) is an interesting one perhaps deserving of its own thread in the main forum as you will get a wider response.

Some possibilities that spring to my mind:

Someone could be "cheating" as a protest at the lack of rule and by making things so ridiculous the hope is to provoke change.

Perception that everyone is "cheating" at the top and that make it ok.

Work and family have tired you out and you are busy so you take a sneaky look at your program to check if you've blundered that can lead down a slippery slope to full blown assisted play.

The MITM gang must be stopped and someone like "Magic Jim" comes along and sees things as the ends justify the means (I'm not necessarily anti Magic Jim but it can lead from beating up cheaters to beating up everyone since you are on a just mission in your mind).

The person has a personality disorder they are pathological and doesn't care, they see any thing as justified to get what they want and don't care whare people think. This person would not care if there were formal rule forbidding the "cheating" behaviour.

The person is hyper-competitive and wants to win by any means within the rule and since there aren't any they see it as anything goes. This person wouldn't violate the explicit rules but would happily exploit any loophole or use gamesmanship.

Most people on this site are here to play fairly and just have some fun.

5five has used another alias Yaschny (see well known cheaters thread) to sabotage my rating, resigning against everyone in order to have a low rating then have that account play computer-like.

These account also make moves minutes apart.

Has anyone had a chance to analyse any of

this guy's games yet? A perfect record, and an id only one different from one of the well-know sockpuppets in the Jamaica Posse, is perhaps a little suspicious?

Isn't Ye_L_o a Jamaica account http://www.littlegolem.net/jsp/tournament/tournament.jsp?trnid=loa.ch.13.4.1

So far in his game against you not one of his moves is one MIA wouldn't play.

Would love to no how fast this player is moving from when the game first loads up.

Suffice to say championship was one of those on my radar and I didn't even know about the consecutive I.D. number with another sockpuppet.

Was mulling up posting something in "well known cheat" about championship, akin to your post here!

"So far in his game against you not one of his moves is one MIA wouldn't play."

Thanks, I'll tell RoRoRo about that, for his records. I don't have evidence that he's in the Jamaica Posse yet, but for him to have appeared on the radar and been caught cheating after only a handful of games makes me think that the real evidence will appear pretty soon.

(And yes, Yelo = Jamaica)

I'm more careful with championship because I have no experience facing him.

The very nature of championship play championships no reason there would be any link to Jamaica for several tournament cycles. The game against Ye_L_o was looking MIA-like I quit half way through.

BTW offtopic does RoRoRo still play GWG?, I would like to try facing him.

You can invite RoRoRo to a game, but he doesn't sign up for any tournaments.

LOA variant on 9x9 is not enough, because cheaters can run MIA, choose 10x10 board and setup stones on 9x9 board.

.

Therefore I create Black Hole variant on 9x9, where central field is Black Hole field.

Richard, why do you allow a very few users (often with multiple accounts) to so badly interfere with the intended structure of Little Golem? None of the solutions are perfect, but many are far better than the status quo.

The mechanics of 9x9 board are a lot different from 10x10 board with the 9x9 position set up. If you relied on the on the 10x10 board to play 9x9 you would probably come a cropper.

Will attempt some tests on how a game played using the wrong size board might play out, but IMO only people who don't understand LOA would trust AI designed for a different sized boards.

Thanks Richard for this variant, but where can we find the rules on how the black hole impacts initial LOA rules?

It seems we can jump above (or go through) it, is it only a dead square?

Marius Halsor "The anti-cheating board has examined the player "5five", and concluded that this player is a cheaters. He/she will be banned from further participation in the tour, and will also have his/her name removed from the results list."

Note that we ONLY deal with the "tour of LG", and nothing more. We can decide that players are cheaters as much as we want to - only Richard can remove them from Championships and MCs (and has so far never done so). We can only blacklist them from participation in the Tour.

Richard wrote:

> Therefore I create Black Hole variant on 9x9, where

> central field is Black Hole field.

I just saw it - and replayed a few games from a

user tournament. My impression: on 9x9 LOA is a bit

like a race of slow caterpillars.

So, what about 7x7 with a black hole in the center?

Ingo.

That is why outside my discussion of the tour I, although others do not put quotation marks "cheater," since as the rules stand there are no formal rules! There is only a kind of social contract that using a near unbeatable bot is not cool.

Using these methods against unsuspecting innocent players is unethical, it isn't, as things stand cheating.

Until the rule change a don't advocate any more punishment than that stipulated in rules for the Tour. That said adopting similar rules from the tour and applying them to the championships would be a good move and popular move.

Ingo perhaps we could take this to the Blackhole thread this discussion concerns unethical play.