Greatest Player according Rybka 3, 3 min. quad

Hello! I did the same as with Rybka 2.3.2a at 14 ply forwards(posted in this site), but this time using 3 minutes per move with Rybka 3. Resuts are more logic, but Kasparov is still just at 8th place and Capablanca in 26th place. 6 of the top 10 still active players, 1 active player of the bottom 10. Most ahead of their time, Morphy, Fischer and Lasker. Promising new star: Karjakin. The top 8 have been World Chess Champions (Svidler of chess 960, 2nd at WCC 2005 ), the 9th Morphy is consider an unofficial chess champion, the 10th Leko is still an active player (World Chess 960 Champion, 2nd in WCC 2004, won Linares, Corus, Dortmund). More than 1000 moves consider except for Greco, Philidor , Eichborn. Players were selected considering results obteined in previous analysis of concur with Rybka 2.3.2a at 14 ply, and adding most famous players. Fischer, Topalov, Tal , Karjakin , Kasparov, Botvinnik, Petrosian and Leko most significative advances compare with previous study. Matulovic, Uhlmann and Ivkov are still a surprise, did they were that good?, or is a problem with this very simple method, or both? You can compare results with what Deka did , Bratko & Guid , Charles Sullivan and others. Similar results for Fischer at Top with Deka (nimh) and Charles Sullivan, with Kramnik at Top with Bratko & Guid.

1. Fischer 65,86
2. Kramnik 65,85
3. Svidler 64.98
4. Topalov 64.77
5. Karjakin 64.47
6. Tal 64,45
7. Anand 64,11
8. Kasparov 63,94
9. Morphy 63.75
10. Leko 63.50
11. Shirov 63.09
12. Ivanchuk 63.09
13. Lasker 62,92
14. Euwe 62.75
15. Beliavsky 62.65
16. Khalifman 62,64
17. Matulovic 62.42
18. Uhlmann 62.33
19. Pillsbury 61.95
20. Karpov 61,83
21. Alekhine 61,77
22. Ponomariov 61.62
23. Aronian 61.11
24. Botvinnik 60,91
25. Smyslov 60,85
26. Capablanca 60,84
27. Ivkov 60.58
28. Kasimdzhanov 60.53
29. Spassky 60.20
30. Geller 59.75
31. Petrosian 59.71
32. Rubinstein 59.33
33. Kolisch 58.69
34. Steinitz 58.4
35. Korchnoi 57.44
36. Eichborn 56.27
37. Stein 56.04
38. Zukertort 55.44
39. Anderssen 55.34
40. Greco 54.10
41. LaBourdonnais 53.73
42. Staunton 50.35
43. Philidor 46.20

wat

It's a ranking of how much Rybka 3 agrees with those players' moves. Apparently it agreed with Fischer 65.86% of the time.

The thing is, playing in a match against a human is a wholly different proposition than running a computer programme. Tal, for example, was so dangerous because he could conjure up positions that his opponent couldn't cope with, or induce such pressure through complications and threats that his opponent didn't find the accurate defence in time. Also, I'd be willing to bet on Capablanca in a match against Svidler.

I would be skeptical of this analysis at best. One does not need to be able to see ahead 14 moves in all combinations (like the computer can) to be a good chessplayer, so why should we use these criterion to rank players?

huh!!!

Fischer clearly used Rybka in all of his games!

yup

I think it would be interesting is to see what the match ups of ordinary masters would be. Take older events, say one each from pre-1900, mid-1900 and a modern tmt. just to see how the ordinary players compared.

I don't think anyone means that this is a true ranking of these players. I think it's just something interesting to think about. What it does show is that Fischer, Kramnik and Svidler are/were extremely accurate players. That doesn't mean they win more games than everyone else, it just means the moves they make tend to be very wise.

What Gabrielconroy said about Tal is a great point. It means nothing to come up with the accurate response to an extremely complicated move if you can't do it in an efficient amount of time. The clock and psychology are very important parts of chess that Rybka doesn't necessarily consider here (the clock is somewhat included through the existence of the games, I suppose).

Really cool analysis, anyway. Not wildly practical, but I like it.

I was always under the impression that capablanca was the most inline with computer programs....ohhhh maybe he is more of a fritz type player (hahaha)

Yes, good point. I didn't think about that.

That was before Rybka. Rybka 3 has terrific positional knowledge.

A great list by the way. Any links to other researches like these?

3 things surprised me:

1)Tal being so high up in the list. If I had limited words to describe Tal's style, "accurate" wouldn't be one of them, I'd prefer wild, complicating, etc.

2)Capablanca. I would expect him to be in top 5 or something. very interesting.

3)Karpov being so down! How did this happen? I thought he had a similar style with Kramnik.

and I must say, I wasn't at all surprised by Kramnik being one of the highest, I think he is the best player of all times.

very interesting research, thanks.

I'd prefer to see at least 20 plydepth and Multi PV - comparing played moves to only  the best move as proposed by chess engine (Rybka, Houdini or whichever) is not enough! In chess engines many moves score very closely to the "best move", especially with low plydepth! Also, the best move at plydepth 16 can often become only 3rd or worse at plydepth 20.

 I've seen this by analyzing one of the greatest chess games of all times - Kasparov vs Topalov (1999) - at plydepth of 22 (on average it took my PC 30 mins per move), best moves changed quite often compared to low plydepth. 

To compare many games plydept of 22 is surely too time consuming, but plydepth of 20 might be just appropriate ;)