20 Oct 05
1 edit
Originally posted by DoctorScribblesGreat, now suppose the cheater alternates search times and ply-depths. You yourself have claimed the following in this thread:
My proposed algorithm would in fact be very likely to classify such player as a cheater! In particular, when examining the games of the user in question, it would observe several features:
"Matches engine X on every other move."
"Matches engine Y on every other move."
"Matches engine X with a rate of 50%."
"Matches engine Y with a rat ...[text shortened]... observed in the game of a cheater or non-cheater, and the player can be classified accordingly.
...surely you realize that there are numerous trivial algorithms for cheating with engines and avoiding detection. Anybody who wants to can do it and get away with it. (If they can't figure one out themselves, I'll be happy to tell them how.)
So, suppose a cheater employs one of these simply algorithms and thus avoids detection by your cheat detection algorithm, yet still manages to beat even the very best opponents, even when moving rapidly, etc. The appropriate "Turing Test" in this case is not merely avoiding detection by your algorithm or some variant algorithm. You should read Gatecrasher's nice post in the general forum concerning detection of engine users.
I'm signing off of this debate. This is taking up valuable time that could be spent making fun of religious people in the Spirituality forum. Again, I understand and share your concerns and those of No.1. I would also like increased transparency, but I'm having trouble seeing how to do that, given that Tebb and other mods have claimed that revealing evidence will make their jobs significantly harder (and it can already take months to amass evidence sufficient to ban a user).
Cheers.
20 Oct 05
2 edits
Originally posted by bbarrThe algorithm is a probabilistic one. It will erroneously classify players as non-cheaters with a non-zero probability. Anybody determined person with half a brain could outsmart it. He just needs to be more clever than the 90% or whatever chosen threshold of the rest of the cheaters.
So, suppose a cheater employs one of these simply algorithms and thus avoids detection by your cheat detection algorithm, yet still manages to beat even the very best opponents, even when moving rapidly, etc.
But those same half-brained people can also outsmart whatever methods the Game Mods are currenly using. Subtract out the equal parts and what you have left is dummies getting caught and some non-cheaters being falsely accused, in either method. Under mine, the falsely accused could appeal to some flaw in the algorithm, that everybody could then objectively analyze and rectify if need be. Under the current system, there is no such appeal.
20 Oct 05
Originally posted by DoctorScribblesSince fw is gone now, may I ask the esteemed Doctor to verify a memory of mine. Did your bayesian algorithim say that ironman did not cheat?
Many months ago at the FW Round Table, I described a statistically sound and objective algorithm, based on Naive Bayesian Classification (http://en.wikipedia.org/wiki/Naive_Bayesian_classification), for classifying players into cheaters and non-cheaters. This algorithm could be tuned to achieve whatever false positive/false negative balance was des ...[text shortened]... cs of the algorithm I described and thus the consensus about it was, "I doubt that will work."
20 Oct 05
Originally posted by GatecrasherA "partial", testable conclusion would be far better than the present situation where the community is asked to take on faith that "X cheated" on the bare assertion by the Game Mods that they have "overwhelming" evidence.
Unfortunately, a statement like "We believe Exy used engine Y in games 1,2, 3 .......... X" may represent only partial evidence.
It is certainly true that several tools have been developed for game moderation at RHP. But I cannot see how any of the evidence gathered using these tools could be publicized without jeopardising their effectiveness.
...[text shortened]... indignation. I'm sure any suggestion that might lead to a workable compromise would be welcome.
20 Oct 05
3 edits
Originally posted by NyxieI'm quite certain I never implemented it. It would require a significant amount of effort that I wouldn't undertake for my own amusement. I don't recall hypothesizing about whether any particular users, other than Feivel, were cheaters. If you recall, the particulars of who cheats and who doesn't bores me, as I stated numerous times at FW. It is the method that interests me. I couldn't care less if every player on this site cheated. It wouldn't affect my enjoyment of the game one bit, and aside from players in the running for being the best of the best with possible ineligible opponents, I don't understand why people get so worked up about it.
Are you certain of that, I do believe you stated that applying this algorithim showed im to not be a cheat. Were you just hypothesizing?
20 Oct 05
Originally posted by DoctorScribblesYes I remember that. I just remember you mentioning this fact in an offhand way, almost in passing. But as you said, not many seemed to care. I still think that most don't care one way or the other. People will continue to not care untill it affects them directly.
I'm quite certain I never implemented it. It would require a significant amount of effort that I wouldn't undertake for my own amusement. I don't recall hypothesizing about whether any particular users, other than Feivel, were cheaters. If you recall, the particulars of who cheats and who doesn't bores me, as I stated numerous times at FW. It is the method that interests me.
As much as I agreed with no1 and others that something needed to be implemented, I am shocked at what has been created. I find myself embittered by it and am slowly reaching a form of appathy towards the whole subject.
I never thought that my vote or oppinion would be used as a rationale for creating a form of secret police. The very concept of this type of system stands against all that I believe in.
There are still cheaters here, people will still play them. Only now we have a system that can be used as tool to threaten and harrass others mercilessly. The very destruction of worlds come not in the hate that we share but in the weopons we choose.
20 Oct 05
2 edits
Originally posted by NyxieThere is still time to sign my petition, a few threads down. As I mention there, and as you reiterate here, the poll was misleading. The Game Mod concept would not have received such widespread support if it were made clear that they would operate secretly and unaccountably, without even publishing standards of what constitutes evidence of cheating.
I never thought that my vote or oppinion would be used as a rationale for creating a form of secret police. The very concept of this type of system stands against all that I believe in.
If our input created this monster, our input should have the chance to destroy it.
20 Oct 05
Originally posted by NyxieHow can anyone compare the game mods to secret police?
I never thought that my vote or oppinion would be used as a rationale for creating a form of secret police.
They are subject to the same rules as everyone else, at least I would assume they are.
20 Oct 05
Originally posted by Dr StrangeloveBecause I still don't know why exy was banned. Do you? Have you seen the evidence, or even the charge? He was there one night, and then the next he was gone.
How can anyone compare the game mods to secret police?
They are subject to the same rules as everyone else, at least I would assume they are.
It sounds like a fair anology to me.
20 Oct 05
Originally posted by zeeblebotThey SHOULD at least post their specific conclusions because:
why SHOULD the mods post the evidence? .... to what purpose? ... the reasons stated don't seem sufficient to risk exposing the collection methods ...
1) The system is unjust as the accused is given no meaningful way to dispute the charges;
2) The system does not have the confidence of many members of the community as they are asked to take the Game Mods' assertions as gospel truth without being shown any colloborating facts:
3) The Game Mods bald assertion that the making of a statement such as "User X used an engine in game 1, 2, 3 ....... T" would be damaging to their ability to detect future cheaters is illogical and almost certainly false;
4) We have no way of knowing that those actually expelled were cheating under the present system except our trust in the inherent fairness of the individuals concerned (contrary to what has been suggested, only one was among the top vote getters in January). This is causing some anxiety among honest players who fear that the possibility of "false positives" is significant under the current, inquistional regime. An airing of testable findings would alleviate this concern.