...and I will analyse their games to see how accurately they play according to Houdini.
http://www.playtheimmortalgame.com/players/playertable.php?cbqsid=5258
It will be interesting to see how gifted the players at the top of the player tables are these days.
Please post the name of your chosen player in response and in 48 hours I will analyse the one who gets the most votes.
For purposes of transparency, the criteria for game selection will be as follows:
20 of the most recently completed non-thematic games vs 2000+ rateds, each of which have 35+ moves.
I will post full ply-by-ply analyses in this thread.
- Joined
- 08 Apr 09
- Moves
- 20004
A few of the top 20 players have hardly any losses (DonQ, Ian Maver, Solving In Style), which indicates they play very accurately and consistently at a high-level. Since DonQ is no longer active apparently, I'd pick Ian Maver as he is the highest rated of the two others. Also, he plays stronger opponents than Solving In Style.
sbelanoff - User 560746 - interests me. He recently beat carpmaniac71, who is an obvious engine user, and I'd like to know if that game was on the level or not - Game 10374658.
sbelanoff is in the top 25 on another correspondence site which explicitly prohibits the use of engines during play:
http://www.chessbymail.com/top25.htm
There are several well known and respected correspondence players on that list, which suggests to me that chessbymail.com takes cheating more seriously than this site does.
For your edification and enjoyment, I'd like to present another game by sbelanoff: Game 7737828.
Originally posted by Fat LadyOh, come on now! He clearly forgot to use his Analyze Board
For your edification and enjoyment, I'd like to present another game by sbelanoff: Game 7737828.
Originally posted by tvochessWell, i think it's pretty clear that Solving in style uses multiple accounts, which is something i have raised with admin already..
A few of the top 20 players have hardly any losses (DonQ, Ian Maver, Solving In Style), which indicates they play very accurately and consistently at a high-level. Since DonQ is no longer active apparently, I'd pick Ian Maver as he is the highest rated of the two others. Also, he plays stronger opponents than Solving In Style.
Game 10266655
Game 10272260
Game 10272269
Game 10271216
Find me a 2000 player who falls for an f7 mate against the same person four times and i'll eat my hat! Whether that means they are really cheating or not, who knows..
Originally posted by Fat LadyLots of the recent games are far too short for my purposes.
sbelanoff - User 560746 - interests me. He recently beat carpmaniac71, who is an obvious engine user, and I'd like to know if that game was on the level or not - Game 10374658.
sbelanoff is in the top 25 on another correspondence site which explicitly prohibits the use of engines during play:
http://www.chessbymail.com/top25.htm
There are ...[text shortened]... ur edification and enjoyment, I'd like to present another game by sbelanoff: Game 7737828.
I'll only be analysing ones with at least 35 moves vs 2000+ rateds.
Originally posted by Fat LadyThe current game of sbelanoff against david Tebb is at 53 moves...
Fair enough. How about David Tebb then?
You could do the analysis of all 35+ move games of the current round of the site championship Tournament 16990. That would surely be interesting
Originally posted by Fat LadyI analysed David's RHP games in Nov 2012.
Fair enough. How about David Tebb then?
Back then his return vs 2000+ rateds was:
{ David Tebb (Games: 20) }
{ Top 1 Match: 411/695 ( 59.1% ) Opponents: 412/687 ( 60.0% )
{ Top 2 Match: 552/695 ( 79.4% ) Opponents: 538/687 ( 78.3% )
{ Top 3 Match: 604/695 ( 86.9% ) Opponents: 584/687 ( 85.0% )
{ Top 4 Match: 640/695 ( 92.1% ) Opponents: 609/687 ( 88.6% )
Which is about what you'd expect the absolute best Super GM's to return OTB.
eg:
A few months ago I went to chessgames.com & looked at Carlsen's games & selected games under the following criteria:
20 most recently completed vs top players (lowest rated being Gawain Jones FIDE 2632) which all have at least 20 non-theory moves. To be fair, I avoided all the blitz games.
The match rates didn't surprise me. They were consistent with what I'd expect the best unassisted OTB player to achieve.
The analysis of non-theory moves was done using the usual method to create benchmarks & also find cheats:
Houdini 1.5a x64 Hash Table:256Mb Time:30s per ply Max Depth:20ply
Using the system:
AMD Phenom x4 2.30 Ghz
4GB RAM
Carlsen {20 games}
{ Top 1 Match: 477/828 ( 57.6% ) Opponents: 452/821 ( 55.1% )
{ Top 2 Match: 617/828 ( 74.5% ) Opponents: 594/821 ( 72.4% )
{ Top 3 Match: 690/828 ( 83.3% ) Opponents: 667/821 ( 81.2% )
{ Top 4 Match: 732/828 ( 88.4% ) Opponents: 709/821 ( 86.4% )
Oh. To say that's disappointing would be an understatement on par with someone stating that Greenpawn quite likes a game of chess every now and then. I suppose I was a little naive to think that an unassisted human could hold his own against the many cheats on this site.
Originally posted by Fat LadyWell if we compare we should compare with Correspondence masters. Someone did that years ago.
Oh. To say that's disappointing would be an understatement on par with someone stating that Greenpawn quite likes a game of chess every now and then. I suppose I was a little naive to think that an unassisted human could hold his own against the many cheats on this site.
- Joined
- 25 Jul 07
- Moves
- 27921
Originally posted by PonderableI can do that if someone can provide some games. Thanks to thaughbaer's Perl scripts, I have the process pretty much automated now.
Well if we compare we should compare with Correspondence masters. Someone did that years ago.
My results for David Tebb were:
David Tebb
Number of games 15
Number of moves 442
Number of mistakes 22
Mistakes per move 0.0497737556561
Top 1 Matches 0.563656621095
Top 2 Matches 0.75746835376
Top 3 Matches 0.843934782085
A mistake is defined as a move that lowers the engine evaluation by more than 50 centipawns.
For comparison:
TheBigKat
Number of games 25
Number of moves 730
Number of mistakes 13
Mistakes per move 0.0178082191781
Top 1 Matches 0.63446022674
Top 2 Matches 0.795293979111
Top 3 Matches 0.876121289092
Carlsen, Magnus
Number of games 9
Number of moves 380
Number of mistakes 13
Mistakes per move 0.0342105263158
Top 1 Matches 0.592904469797
Top 2 Matches 0.766378459551
Top 3 Matches 0.850189708905
.. but, this isn't entirely fair as correspondence is very different to OTB. My RHP matches would be in a different league to my OTB matches. So I'd be interested in running some pre-engine correspondence games if anyone has any.
There are another couple of points to consider.
Firstly, assuming these players aren't using engines, do you think that the longer such a player looks at a position the more engine-like their moves necessarily become?
The pre-1990's correspondence World Championship finalists had match rates which were very similar to the best OTB Super GM's. These days OTB players like Kramnik & Carlsen have teams preparing their middlegame play using the strongest engines available, yet still the match rates all hover around the same level when you analyse a large batch of games.
I think that instead of becoming more engine-like, their play becomes less error-prone given longer time controls.
Secondly, let's assume David Tebb isn't using an engine.
He's doing remarkably well holding his position in the top 20 if it's a case of man vs machine, unless you think most of his 2200+ opponents are also naturally gifted correspondence players. Most engines are upwards of 3000 Elo rated now & a patzer using one in combination with a Megabase can easily beat an unassisted GM, especially if the GM was unaware that an engine was being used.