02 Dec 08
1 edit
Originally posted by KeplerA self-professed statistician that doesn't even realize the importance of his assumptions regarding the data-generating process. Poor kids.
Well carry on then. I am going to be teaching in about 30 minutes and won't be back for about three hours. You will have to talk to yourself for a while but, rest assured, I will return.
02 Dec 08
Still not clear to me why anyone would object to Kepler's efforts. If he can strengthen the cheat test, that's good. If he shows it to be weaker than thought ... well, he can't, or so the cheat police auxiliary are telling us, so what's the problem? Korch seems to think that questions of the Kepler type will give new hope to cheats, but how is that? If the current test is sufficiently strong, the cheats will be detected sooner or later, regardless of their morale.
02 Dec 08
2 edits
Originally posted by luctrucThe problem is Kepler's continuous insinuations that his results somehow undermine the use of match-up rates to detect cheats, when they don't. Of course, he says this in one post and then contradicts himself in the next by denying it.
Still not clear to me why anyone would object to Kepler's efforts. If he can strengthen the cheat test, that's good. If he shows it to be weaker than thought ... well, he can't, or so the cheat police auxiliary are telling us, so what's the problem? Korch seems to think that questions of the Kepler type will give new hope to cheats, but how is that ...[text shortened]... ufficiently strong, the cheats will be detected sooner or later, regardless of their morale.
02 Dec 08
Originally posted by PalynkaNot quite the next post. I didn't insinuate either. I said it quite boldly. And it wasn't continuous. There was a reason for my actions earlier in this thread and I got the result I desired. I don't actually think my results undermine the use of match up rates to detect cheats at all and have said so. If you or anyone else wants to think otherwise I have no objection.
The problem is Kepler's continuous insinuations that his results somehow undermine the use of match-up rates to detect cheats, when they don't. Of course, he says this in one post and then contradicts himself in the next by denying it.
02 Dec 08
Originally posted by luctrucThey've already "flocked" here, as they have "flocked" to every internet chess site. The problem is how to at least reduce the problem. Suggesting that the current standards that require match up percentages above what the greatest GMs and correspondence chess GMs achieved isn't stringent enough to virtually eliminate the possibility of false positives isn't how to achieve that goal.
And these insinuations will cause engine users to flock to the site?
02 Dec 08
Originally posted by no1marauderIf the system works then no amount insinuation or praise will change that. Suggesting that anything I say may change how well that system works is preposterous. If anything, suggesting openly (not insinuating) that match up rates don't work may cause cheats to be more careless, not cause those who have confidence in the method to change or stop. I suggest that if you do not have confidence in that method then the problem lies with you not anything I have said.
They've already "flocked" here, as they have "flocked" to every internet chess site. The problem is how to at least reduce the problem. Suggesting that the current standards that require match up percentages above what the greatest GMs and correspondence chess GMs achieved isn't stringent enough to virtually eliminate the possibility of false positives isn't how to achieve that goal.
02 Dec 08
Originally posted by no1marauderRefresh my memory. When somone submits a fair play ticket showing a high match-up with a particular engine -- Cheatsoft, say -- on a particular vector of settings and running on particular hardware, how many games from the Old Masters' DB are analyzed with Cheatsoft, on the same settings and with the same hardware, in order to determine the big P, as in "P is the probability that a strong, unassisted player will match Cheatsoft's first n moves"?
They've already "flocked" here, as they have "flocked" to every internet chess site. The problem is how to at least reduce the problem. Suggesting that the current standards that require match up percentages above what the greatest GMs and correspondence chess GMs achieved isn't stringent enough to virtually eliminate the possibility of false positives isn't how to achieve that goal.
02 Dec 08
1 edit
Originally posted by luctrucNone; the wheel isn't reinvented every time you have to drive to WalMarts.
Refresh my memory. When somone submits a fair play ticket showing a high match-up with a particular engine -- Cheatsoft, say -- on a particular vector of settings and running on particular hardware, how many games from the Old Masters' DB are analyzed with Cheatsoft, on the same settings and with the same hardware, in order to determine the big P, as in "P ...[text shortened]... e probability that a strong, unassisted player will match Cheatsoft's first n moves"?
02 Dec 08
Originally posted by KeplerPressure from users here is the driving force behind the Site Admins taking a firm stand against engine use. There would be no Game Mods if there hadn't been a huge outcry in the forums almost 4 years ago (much of it fueled by accurate cheating accusations in the forums). Reduce the confidence that the average subscriber has in the efficiency of the system (particularly in an area where the system is without reasonable doubt efficient) and there will be less incentive for the Site Admins to act aggressively to address the problem.
If the system works then no amount insinuation or praise will change that. Suggesting that anything I say may change how well that system works is preposterous. If anything, suggesting openly (not insinuating) that match up rates don't work may cause cheats to be more careless, not cause those who have confidence in the method to change or stop. I suggest tha ...[text shortened]... do not have confidence in that method then the problem lies with you not anything I have said.
02 Dec 08
Originally posted by luctrucBy much analysis of pre-computer era human games, both OTB and correspondence. In fact, if SB's numbers are accurate (and they sound about right), then they err quite a bit on the conservative side.
OK, then how, exactly, were the magic P's developed in the first place?
02 Dec 08
Originally posted by no1marauderYes, well, "much analysis," at 30 sec/move (saves time), first 4 moves, but on how many engines? No need to name them, just how many? And why were OTB games included? After all, we play postal chess here, and blunders on the part of strong postal players are rare, whereas even the OTB Olympians occasionally pull a "??"
By much analysis of pre-computer era human games, both OTB and correspondence. In fact, if SB's numbers are accurate (and they sound about right), then they err quite a bit on the conservative side.
02 Dec 08
Originally posted by no1marauderThe average subscriber doesn't seem to come anywhere near the forums. Most of them play chess and ignore the antics in here. I also suspect the defence was robust enough to reassure anyone who did bother to follow what was going on. I think that if a real problem with the system was discovered, the same forces that drove the admins to set up the game mod system and have made sure that the problem of cheating continues to be dealt with aggressively would ensure that the admins either fixed the problem or found another way to deal with cheating.
Pressure from users here is the driving force behind the Site Admins taking a firm stand against engine use. There would be no Game Mods if there hadn't been a huge outcry in the forums almost 4 years ago (much of it fueled by accurate cheating accusations in the forums). Reduce the confidence that the average subscriber has in the efficiency of the syst ...[text shortened]... nd there will be less incentive for the Site Admins to act aggressively to address the problem.