Only Chess
05 Feb 11
07 Feb 11
Originally posted by VarenkaIf you're the one proposing a change in standards here, the onus would be on you to produce SOME evidence to support your assertion that a top CC GM would A) Likely produce higher matchup rates than CC GMs of the pre-engine age and Super OTB GMs of that and this and every other age; as well as B) That such top CC GM would be likely to play on RHP. You can't seem to do either, but also say you're not in favor of changing standards here. If you are not, what the hell are we discussing? Your attempt to nitpick my assertion?
Do you have evidence that every player on the ICCF uses computer assistance? They don't ban it; but they don't encourage it. Who's the highest rated ICCF player who chooses not to use engine assistance? Of course, your evidence isn't fast coming either...
07 Feb 11
Originally posted by VarenkaYes, since I've actually run analysis of recent WC matches and compared them to prior ones. No statistical difference.
Chess Cafe used to host an excellent article...
http://www.chesscafe.com/zip/misha.zip
Interview with Alexander Riazantsev (misha37)
Alexander Riazantsev. Born on September 12, 1985 in Moscow.
International grandmaster. 2006 Moscow champion. Current FIDE
rating – 2622.
MS: Has chess become more computer-like?
AR: With many strong GMs spe ...[text shortened]... ng players are learning this kind of chess.
----------
Do you know better than these guys?
07 Feb 11
1 edit
Originally posted by no1marauderSo a top Russian trainer and a 2600+ GM explain how they've seen the game evolve... but you know so much better....
Yes, since I've actually run analysis of recent WC matches and compared them to prior ones. No statistical difference.
I presented some evidence.
And let's see some of your actual results. I have some automated matchup software I could run to compare. Let's see *your* evidence.
07 Feb 11
1 edit
Originally posted by DanTriolaAm I being an idiot here ? If you match the engine it's a yes or no.. you either match it or you don't. The rest is counting.. how many times you match it out of the total number of moves you make.. once past a given point in the game. Where are all these statistical probability functions coming from ? And no.. I don't understand them.. but I can see yes and no and I can count.
No1,
I was not trying to be impressive. I was trying to simplify the explanation for your benefit. Obvious that it was still above your head though. Sorry.
The use of the term "Debunked" indicates a rigorous mathematical proof. If you cannot comment on the basic explanation which I've given other than to characterize it as "gobbledygook", then it is di tistics". Which is what you are doing.
And that was the point Twain was making, also.
Mike
07 Feb 11
Originally posted by VarenkaThey presented some opinions without any evidence.
So a top Russian trainer and a 2600+ GM explain how they've seen the game evolve... but you know so much better....
I presented some evidence.
And let's see some of your actual results. I have some automated matchup software I could run to compare. Let's see *your* evidence.
If you automated software, do it yourself. Try Kasparov-Karpov in the 80's and compare it to Kramnik-Topalov or Anand-Kramnik, etc. etc. etc. Tell me if the percentage of matchups increased significantly as you theorized it should.
07 Feb 11
Originally posted by no1marauderThey presented some opinions without any evidence.
I think their experience of being involved in chess at the top level is a bit more convincing than anything you've yet put forward.
If you automated software, do it yourself
I will. But as you know, at e.g. 30 seconds per move, it will take some time to run. But you stated that you've already done it - so let's see your results now.
07 Feb 11
A question I have been wondering: Is it realistically possible to clearly detect engine use by a strong player on RHP who uses the engine only for assistance such as with blunder checks and on critical moves only? Where he is not just blindly following the engine.
Avoid detection but say increase your RHP rating from 1850 (no engine) to 2100 (engine blunder check)?
- Joined
- 03 Sep 03
- Moves
- 87628
07 Feb 11
Originally posted by moon1969More difficult but not impossible. Those people will last longer, but I think would eventually get caught.
A question I have been wondering: Is it realistically possible to clearly detect engine use by a strong player on RHP who uses the engine only for assistance such as with blunder checks and on critical moves only? Where he is not just blindly following the engine.
Avoid detection but say increase your RHP rating from 1850 (no engine) to 2100 (engine blunder check)?
07 Feb 11
Originally posted by VarenkaSquelchbelch gave you the relevant data here: http://www.redhotpawn.com/board/showthread.php?threadid=114715&page=3#post_2286401
[b]They presented some opinions without any evidence.
I think their experience of being involved in chess at the top level is a bit more convincing than anything you've yet put forward.
If you automated software, do it yourself
I will. But as you know, at e.g. 30 seconds per move, it will take some time to run. But you stated that you've already done it - so let's see your results now.[/b]
You ignored it then; why wouldn't you ignore it again?
In truth, I've switched computers since then and the only relevant file I can find is an analysis of Kramnik-Kasparov. The games are analyzed but the results not quantified (I did that long ago but don't have them on hand any more). So, I'd have to go move by move through the file and check off which are matches and which are not a procedure that would take me hours. My clear recollection is that the results were consistent with other WC matches. Why isn't that good enough?
07 Feb 11
Originally posted by no1marauderI just looked at that thread, and I noticed Squelchbelch said:
Squelchbelch gave you the relevant data here: http://www.redhotpawn.com/board/showthread.php?threadid=114715&page=3#post_2286401
You ignored it then; why wouldn't you ignore it again?
In truth, I've switched computers since then and the only relevant file I can find is an analysis of Kramnik-Kasparov. The games are analy ...[text shortened]... tion is that the results were consistent with other WC matches. Why isn't that good enough?
"The already strong player who only uses Fritz to check the odd line or validity of a few moves in a few games won't really be caught by any detection method."
In a way an answer to my question in my previous post.
Seems like on RHP, one could increase their rating, for example, from 1850 (no engine use) to 2050 (mild engine use) without being detected?
- Joined
- 25 Jul 07
- Moves
- 27921
08 Feb 11
Originally posted by moon1969It would be very difficult to detect, but not necessarily impossible. I think you'd have to check for 'bad' moves, rather than top 3 match-ups, which will by definition be 'good' moves. It would be interesting to analyse a few games between good humans and see how many moves were bad by engine standards.
Seems like on RHP, one could increase their rating, for example, from 1850 (no engine use) to 2050 (mild engine use) without being detected?
Also, you have to consider the level of opposition your centaur would be facing. When playing 2k+ people, you can't get away with *any* sloppy moves. I know this from experience ... :-(. Someone switching on their engine only occasionally would have to be very lucky with their non-assisted moves to stand any chance at all.