fair enough, thanks.
i guess if you have the motivation, then go for it. i don't think i could personally be bothered going through that evidence-gathering procedure, which looks like it requires a considerable investment of time and effort, but then, as you say, i've never been knocked out of the final round of a tournament by a blatant engine user.
good luck with it.
I know you have to go over a sample of games but what about,
say 21 top choice Fritz moves on the trot in a complicated position.
Does that not indicate something is odd.
By the way, I don't think you are obsessed, it's just a healthy hobby.
Also if you fine tune your method then you may be called upon to
prove or disprove claims that an OTB player has cheated.
Originally posted by Blackampeverybody starts off naive and in certain ways tolerant of cheating. the first cheater you face won't even matter that much, you just shrug and go on. but the bad experiences keep piling up, and quite soon wasting months of analysis against engine users start really pissing you off. you also find time after time that people who you thought were beyond suspicion, were actually cheating all the time. and in the end you check at least superficially every suspicious opponent before agreeing to a game (or setting him up against someone in your clan etc). because you don't want to waste your time against them.
fair enough, thanks.
i guess if you have the motivation, then go for it. i don't think i could personally be bothered going through that evidence-gathering procedure, which looks like it requires a considerable investment of time and effort, but then, as you say, i've never been knocked out of the final round of a tournament by a blatant engine user.
good luck with it.
squelch, you're the man. recced.
Originally posted by greenpawn34Despite having a degree in maths, I was never very interested in statistics.
I know you have to go over a sample of games but what about,
say 21 top choice Fritz moves on the trot in a complicated position.
Does that not indicate something is odd.
However, if you consider that a very strong player has, say, a 0.65 chance of choosing a move which is Fritz' first choice in a particular position, then the probability of this player choosing Fritz' #1 move for the last 21 moves in a particular game would be 0.00012. I.e. you might expect to see this happen 1 game in every 8000.
What I wonder about is how a players style, or should I say strengths and weaknesses, affect the data. Is it not possible that a player could naturally play like a computer?
Say a strong junior who has a gift for tactics but very little positional understanding or endgame knowledge. Is there not a danger that a player like this could be unfairly punished by these witch-hunt?
Originally posted by peacedogno, it's not possible. engines solve positions in a fundamentally different way, which is impossible for a human to replicate.
What I wonder about is how a players style, or should I say strengths and weaknesses, affect the data. Is it not possible that a player could naturally play like a computer?
Say a strong junior who has a gift for tactics but very little positional understanding or endgame knowledge. Is there not a danger that a player like this could be unfairly punished by these witch-hunt?
to back that up, there hasn't been a single past nor contemporary GM who played like that. it simply doesn't happen. people play like people, engines play like engines. and while their strength might in some circumstances be similar, they still play very differently.
engines solve positions in a fundamentally different way, which is impossible for a human to replicate. There hasn't been a single past nor contemporary GM who played like that
This is true enough. But if one looks these days at super-GM games, they are significantly different in dynamic conception and structural characteristics from GM games even 20 years ago. That is the influence of machines, most notably on openings. That said, machines simply calculate better, and are unimpeded by 'principle' or the lessons of experience. They simply propose what works best in mechanical terms here and now. So a machine never sacs merely for pressure or initiative, only for measurable advantage. Humans of course play all sorts of intuitive and psychological tricks, pointless against a machine.
How to win, or more likely hold the line, against a machine? Reach positions machines don't handle well if you're good enough. Blocked middle-game positions that contain a distant dynamic (outside passed pawn in due course, for example) that can be worked for by manoeuvre. Or rook endings. I've found machines (or suspected machines) are pretty clueless in endings of this kind. Everything is relative of course - you've got to know theory very well yourself to get or keep an advantage. Minor pieces endings are not handled well either.
But in the end, one machine or another is going to crunch you down, even if you're a strong player. So offer a half while you still can 🙂
I think I just found a new GM playing on RHP:
I just checked one game, and the match up with fritz is
best move: 79%
+second best: 89%
+ third : idem.
What do you think of that? 80 % of fritz best line...
we may even go further, those moves who are not best line are : -either when second move is very close to first one; - at the end of the game, when he has completely won (the 11% left, not best N/A moves...)
given that the guy used to be around 1700, and is now much higher...
Originally posted by Macpoone game can give any results up to 100% 1st choice. and it's just good technique to make suboptimal but uncomplicated moves in a won position.
I think I just found a new GM playing on RHP:
I just checked one game, and the match up with fritz is
best move: 79%
+second best: 89%
+ third : idem.
What do you think of that? 80 % of fritz best line...
we may even go further, those moves who are not best line are : -either when second move is very close to first one; - at the end of the game, ...[text shortened]... ft, not best N/A moves...)
given that the guy used to be around 1700, and is now much higher...
on the 1700 thing, it's crucial how fast the improvement occurred (and whether the 1700 was reliable to begin with). if you look at my graph, you'll see it starts from 1500. if you look at my first games it's obvious I wasn't even nearly that good, but more like 1200-1300. the slope up to 2000 is very steep, and I'm sure I've been reported many times because of it (would be fun to know how many).
but if you look at the dates, you'll realize the first games were 4 years ago. there never was any instant changes, but instead slow and gradual improvement. so such things need to be taken into account as well.
but if it quacks like a duck etc, and there are no easy explanations you can think of, just report the player.
It's sometimes good to 'bump' a post if it has, as is the case here, relevant
and excellent information in it.
I'm positive there is a good chess book waiting to be written using this data
crunching method.
A scan could be run over every World Champion using the one match that won
them the World Title. This would give an indication who played the best chess
according to F12 or R3.
It would be interesting to see Botvinnik's figures as he won it 3 times.
Did he maintiain a steady figure or was there a particular match where he
raised his game.
How good was Capablanca?
Was Euwe's win a fluke?
Who has the best score from the 3 K's?
Who do you think would get the lowest match up?
Petrosian? Not that he was bad, the opposite infact. But some of his
deep 'kill the idea before my opponent sees them' moves would simply not compute.
Lasker? Some of his trickery was proved unsound by analysis carried out hours
after the game had finished. The box would have a field day jumping on his moves.
But thinking about this, perhaps not. After all Lasker gambled only on one move.
The moves played after that were perfect.
With all the interest in computers and the eternal debate about who was the
greatest ever pawn pusher this book is simply wating to be written.
A good bump.
I agree. It was a good post. But I have two issues:
If cheats are to be found using a standardised method such as that proposed by Squelchbelch, doesn't making that method public (by posting it here) make it easy to dodge?
Suspect 99.9% of cheats are of the occasional variety anyway. There is absolutely no rationale behind using an engine for every move. But the occasional "haven't got time to analyse this it's really complex", "just this once". "really want to win this tournament", type of cheat is understandable, though not of course justifiable.
Of course it's good to get rid of that 0.1% who are stupid enough to use their engine on every move...but it's an awful lot of work for a small benefit