Still possible

[Lyudmil Tsvetkov]

Well, sorry indeed, I wanted to post a message, and as I actually did not very well know what topic I could pick up, I decided to make public ( ) a couple of games I just played against Stockfish 4. I must confess that this is the first time I try my hand at Stockfish 4, as for quite some time I was not quite sure which Stock version exactly I should have downloaded.

I must tell you that I am very happy with the way Stockfish 4 plays: it is kind of intriguing, straightforward chess with the occasional nasty surprise... I was not very sure I would get some draws, once Houdini 3 is defeated by this version, but it proved to still be possible to draw games. I lost many games, and won one (but do not publish it, as I had to retract a move in an already won position, but that is not counted). My problem is that I play blitz, and that certainly favours the engine. I might sometime try a longer TC match, but for the moment am happy with 5 or 6 drawn games, not so bad for a first try.

Below are 3 of the games I managed to draw (with adjudication by myself because of the fast TC in obviously drawn positions, as well as positions where the engine shuffles more than 5 moves in a row, meaning it does not see a convincing continuation; I understand very well such rules are a bit weird, but, bearing in mind the TC, and my desire to avoid dull play when I could play another interesting game instead, I think this is about fair).

[pgn][PlyCount "60"]
[Event "Blitz 2m+2s"]
[Site "Sofia"]
[Date "2013.09.12"]
[White "Tsvetkov, Lyudmil"]
[Black "Stockfish 4 64 SSE4.2"]
[Result "1/2-1/2"]
[ECO "D05"]
[TimeControl "120+2"]
[Annotator "Tsvetkov,Lyudmil"]
[MLNrOfMoves "30"]
[MLFlags "000100"]

{1024MB, Dell XPS 4Cores} 1. e3 {[%emt 0:00:00]} 1... e6 {[%emt 0:00:07]} 2. d4
{[%emt 0:00:02]} 2... c5 {[%emt 0:00:02]} 3. Nf3 {[%emt 0:00:02]} 3... Nf6
{[%emt 0:00:07]} 4. Bd3 {[%emt 0:00:04]} 4... d5 {[%emt 0:00:06]} 5. c3
{[%emt 0:00:02]} 5... Be7 {[%emt 0:00:09]} 6. O-O {[%emt 0:00:06]} 6... Nc6
{[%emt 0:00:01]} 7. Ne5 {[%emt 0: 00:03]} 7... Bd6 {[%emt 0:00:06]} 8. f4
{[%emt 0:00:02]} 8... O-O {[%emt 0:00:05]} 9. Nd2 {[%emt 0:00:02]} 9... Qc7
{[%emt 0:00:07]} 10. Qf3 {[%emt 0:00:05]} 10... Rd8 {[%emt 0:00: 06]} 11. g4
{[%emt 0:00:03]} 11... g6 {[%emt 0:00:09]} 12. g5 {[%emt 0:00:03]} 12... Nh5
{[%emt 0:00:03]} 13. Qh3 {[%emt 0:00:25]} 13... Nxe5 {[%emt 0:00:05]} 14. fxe5
{[%emt 0:00:02]} 14... Be7 {[%emt 0:00:09]} 15. Nf3 {[%emt 0:00:05]} 15... Bd7
{[%emt 0:00:04]} 16. Kf2 {[%emt 0:00:11]} 16... Qb6 {[%emt 0:00:05]} 17. Rg1
{[%emt 0:00:08]} 17... Ng7 {[%emt 0:00:04]} 18. Rg4 {[%emt 0:00:11]} 18... Bb5
{[%emt 0:00:07]} 19. Bc2 {[%emt 0: 00:30]} 19... Qa6 {[%emt 0:00:04]} 20. Rh4
{[%emt 0:00:08]} 20... Nh5 {[%emt 0:00:05]} 21. Rxh5 {[%emt 0:00:47]} 21... gxh5
{[%emt 0:00:01]} 22. Bxh7+ {[%emt 0:00:23]} 22... Kxh7 {[%emt 0:00:00]} 23.
Qxh5+ {[%emt 0:00:03]} 23... Kg7 {[%emt 0:00:06]} 24. Qh6+ {[%emt 0:00:02]}
24... Kg8 {[%emt 0:00:00]} 25. g6 {[%emt 0:00:07]} 25... fxg6 {[%emt 0:00: 00]}
26. Qxg6+ {[%emt 0:00:02]} 26... Kh8 {[%emt 0:00:01]} 27. Qh6+ {[%emt 0:00:02]}
27... Kg8 {[%emt 0:00:00]} 28. Qg6+ {[%emt 0:00:01]} 28... Kh8 {[%emt 0:00:04]}
29. Qh6+ {[%emt 0:00:01]} 29... Kg8 {[%emt 0:00:00]} 30. Qg6+ {[%emt 0:00:01]}
30... Kh8 {[%emt 0:00: 06]} 1/2-1/2

[PlyCount "94"]
[Event "Blitz 2m+2s"]
[Site "Sofia"]
[Date "2013.09.12"]
[White "Tsvetkov, Lyudmil"]
[Black "Stockfish 4 64 SSE4.2"]
[Result "1/2-1/2"]
[ECO "D40"]
[TimeControl "120+2"]
[Annotator "Tsvetkov,Lyudmil"]
[MLNrOfMoves "47"]
[MLFlags "000100"]

{1024MB, Dell XPS 4Cores} 1. Nf3 {[%emt 0:00:00]} 1... Nf6 {[%emt 0:00:05]} 2.
d4 {[%emt 0:00:03]} 2... e6 {[%emt 0:00:01]} 3. e3 {[%emt 0:00:04]} 3... c5
{[%emt 0:00:06]} 4. c4 {[%emt 0:00:03]} 4... d5 {[%emt 0:00:04]} 5. Nc3
{[%emt 0:00:03]} 5... Nc6 {[%emt 0:00:07]} 6. Be2 {[%emt 0:00:07]} 6... cxd4
{[%emt 0:00:06]} 7. exd4 {[%emt 0:00:01]} 7... dxc4 {[%emt 0:00:05]} 8. Bxc4
{[%emt 0:00:02]} 8... Be7 {[%emt 0:00:15]} 9. O-O {[%emt 0:00:04]} 9... O-O
{[%emt 0:00:04]} 10. Be3 {[%emt 0:00:08]} 10... a6 {[%emt 0:00: 09]} 11. a3
{[%emt 0:00:08]} 11... b5 {[%emt 0:00:09]} 12. Ba2 {[%emt 0:00:07]} 12... Qc7
{[%emt 0:00:03]} 13. Qe2 {[%emt 0:00:12]} 13... b4 {[%emt 0:00:10]} 14. axb4
{[%emt 0: 00:15]} 14... Nxb4 {[%emt 0:00:00]} 15. Bc4 {[%emt 0:00:28]} 15... Bb7
{[%emt 0:00:05]} 16. Ne5 {[%emt 0:00:03]} 16... Bd6 {[%emt 0:00:01]} 17. Na2
{[%emt 0:00:21]} 17... Nxa2 {[%emt 0:00:05]} 18. Rxa2 {[%emt 0:00:03]} 18...
Bxe5 {[%emt 0:00:01]} 19. dxe5 {[%emt 0:00: 04]} 19... Qxe5 {[%emt 0:00:15]} 20.
Bd2 {[%emt 0:00:25]} 20... Qd6 {[%emt 0:00:05]} 21. Bc3 {[%emt 0:00:10]} 21...
Ne4 {[%emt 0:00:06]} 22. Ba5 {[%emt 0:00:16]} 22... Rfc8 {[%emt 0:00:04]} 23.
Bd3 {[%emt 0:00:18]} 23... Qd5 {[%emt 0:00:04]} 24. Bxe4 {[%emt 0:00: 08]} 24...
Qxa2 {[%emt 0:00:04]} 25. Bxb7 {[%emt 0:00:05]} 25... Qxa5 {[%emt 0:00:00]} 26.
Bxa8 {[%emt 0:00:03]} 26... Rxa8 {[%emt 0:00:02]} 27. Rc1 {[%emt 0:00:12]} 27...
g6 {[%emt 0:00:00]} 28. g3 {[%emt 0:00:03]} 28... Rd8 {[%emt 0:00:01]} 29. b4
{[%emt 0:00:19]} 29... Qxb4 {[%emt 0:00:04]} 30. Qxa6 {[%emt 0:00:04]} 30... Qb2
{[%emt 0:00:02]} 31. Rc8 {[%emt 0:00:26]} 31... Rxc8 {[%emt 0:00:05]} 32. Qxc8+
{[%emt 0:00:02]} 32... Kg7 {[%emt 0: 00:00]} 33. Qc5 {[%emt 0:00:13]} 33... Qb1+
{[%emt 0:00:05]} 34. Kg2 {[%emt 0:00:02]} 34... Qe4+ {[%emt 0:00:03]} 35. Kg1
{[%emt 0:00:04]} 35... Kf6 {[%emt 0:00:04]} 36. Qc3+ {[%emt 0:00:10]} 36... e5
{[%emt 0:00:05]} 37. h4 {[%emt 0:00:24]} 37... Qd5 {[%emt 0:00:06]} 38. Qe3
{[%emt 0:00:06]} 38... Qd1+ {[%emt 0:00:06]} 39. Kg2 {[%emt 0:00:03]} 39... Ke6
{[%emt 0:00:02]} 40. Qb6+ {[%emt 0:00:19]} 40... Qd6 {[%emt 0:00:03]} 41. Qe3
{[%emt 0:00:06]} 41... Qd4 {[%emt 0:00:05]} 42. Qe2 {[%emt 0:00:14]} 42... e4
{[%emt 0:00:00]} 43. Qa6+ {[%emt 0:00:10]} 43... Ke5 {[%emt 0:00:00]} 44. Qb7
{[%emt 0:00:09]} 44... Qd5 {[%emt 0:00:06]} 45. Qe7+ {[%emt 0:00:20]} 45... Kf5
{[%emt 0:00:04]} 46. Qg5+ {[%emt 0:00: 30]} 46... Ke6 {[%emt 0:00:00]} 47. Qe3
{[%emt 0:00:06]} 47... Qb7 {[%emt 0:00:00]} 1/2-1/2

[PlyCount "94"]
[Event "Blitz 2m+2s"]
[Site "Sofia"]
[Date "2013.09.12"]
[White "Tsvetkov, Lyudmil"]
[Black "Stockfish 4 64 SSE4.2"]
[Result "1/2-1/2"]
[ECO "D05"]
[TimeControl "120+2"]
[Annotator "Tsvetkov,Lyudmil"]
[MLNrOfMoves "47"]
[MLFlags "000100"]

{1024MB, Dell XPS 4Cores} 1. d4 {[%emt 0:00:00]} 1... Nf6 {[%emt 0:00:04]} 2.
Nf3 {[%emt 0:00:03]} 2... e6 {[%emt 0:00:03]} 3. e3 {[%emt 0:00:02]} 3... d5
{[%emt 0:00:04]} 4. Bd3 {[%emt 0:00:03]} 4... Be7 {[%emt 0:00:06]} 5. O-O
{[%emt 0:00:05]} 5... c5 {[%emt 0:00:00]} 6. c3 {[%emt 0:00:03]} 6... O-O
{[%emt 0:00:09]} 7. Nbd2 {[%emt 0:00:04]} 7... b6 {[%emt 0:00:08]} 8. Qe2
{[%emt 0:00:06]} 8... Nc6 {[%emt 0:00:04]} 9. Ne5 {[%emt 0:00:02]} 9... Nxe5
{[%emt 0:00:07]} 10. dxe5 {[%emt 0:00:02]} 10... Nd7 {[%emt 0:00:04]} 11. f4
{[%emt 0:00:02]} 11... c4 {[%emt 0:00:06]} 12. Bc2 {[%emt 0:00:03]} 12... Ba6
{[%emt 0:00:21]} 13. Qh5 {[%emt 0:00:08]} 13... g6 {[%emt 0:00:00]} 14. Qh6
{[%emt 0: 00:13]} 14... Nc5 {[%emt 0:00:00]} 15. Nf3 {[%emt 0:00:12]} 15... Bb5
{[%emt 0:00:11]} 16. h4 {[%emt 0:00:26]} 16... f5 {[%emt 0:00:04]} 17. Ng5
{[%emt 0:00:15]} 17... Bxg5 {[%emt 0: 00:02]} 18. hxg5 {[%emt 0:00:01]} 18...
Ba4 {[%emt 0:00:03]} 19. Bb1 {[%emt 0:00:07]} 19... Rf7 {[%emt 0:00:20]} 20. Bd2
{[%emt 0:00:15]} 20... Qd7 {[%emt 0:00:06]} 21. Kf2 {[%emt 0:00:39]} 21... Ne4+
{[%emt 0:00:04]} 22. Ke1 {[%emt 0:00:07]} 22... b5 {[%emt 0:00: 03]} 23. a3
{[%emt 0:00:10]} 23... Qb7 {[%emt 0:00:05]} 24. Rf3 {[%emt 0:00:29]} 24... a5
{[%emt 0:00:04]} 25. Bxe4 {[%emt 0:00:17]} 25... dxe4 {[%emt 0:00:02]} 26. Rf2
{[%emt 0:00:06]} 26... Rd8 {[%emt 0:00:04]} 27. g3 {[%emt 0:00:02]} 27... Rfd7
{[%emt 0:00:03]} 28. Qh2 {[%emt 0:00:05]} 28... Rd3 {[%emt 0:00:03]} 29. Re2
{[%emt 0:00:19]} 29... Qd5 {[%emt 0:00:00]} 30. Ra2 {[%emt 0:00:28]} 30... Rd7
{[%emt 0:00:04]} 31. Ra1 {[%emt 0: 00:02]} 31... Kf8 {[%emt 0:00:03]} 32. Qg2
{[%emt 0:00:10]} 32... Ke7 {[%emt 0:00:03]} 33. Qh2 {[%emt 0:00:02]} 33... Kd8
{[%emt 0:00:05]} 34. Qg2 {[%emt 0:00:05]} 34... Bc2 {[%emt 0:00:00]} 35. Qh2
{[%emt 0:00:04]} 35... Kc8 {[%emt 0:00:04]} 36. Qg2 {[%emt 0:00:03]} 36... Kb8
{[%emt 0:00:02]} 37. Qh2 {[%emt 0:00:02]} 37... Ka7 {[%emt 0:00:03]} 38. Qg2
{[%emt 0:00:02]} 38... Rd8 {[%emt 0:00:03]} 39. Qh2 {[%emt 0:00:02]} 39... Kb8
{[%emt 0:00: 07]} 40. Qg2 {[%emt 0:00:02]} 40... Kb7 {[%emt 0:00:04]} 41. Qh2
{[%emt 0:00:02]} 41... Rd7 {[%emt 0:00:03]} 42. Qg2 {[%emt 0:00:02]} 42... Bb3
{[%emt 0:00:03]} 43. Qh2 {[%emt 0:00:04]} 43... Kb8 {[%emt 0:00:03]} 44. Qg2
{[%emt 0:00:02]} 44... Ka7 {[%emt 0:00:02]} 45. Qh2 {[%emt 0:00:02]} 45... Kb7
{[%emt 0:00:01]} 46. Qg2 {[%emt 0:00:02]} 46... Kc6 {[%emt 0:00:01]} 47. Qh2
{[%emt 0:00:02]} 47... Ba4 {[%emt 0:00:01]} 1/2-1/2
[/pgn]

Sorry again for the very stupid post, but, as I wanted to post a message, even a stupid post is something... Also some people might be interested to learn how not to play against engines.

And a single diagram:

[d]8/3r3p/2k1p1p1/pp1qPpP1/b1p1pP2/P1PrP1P1/1P1BR2Q/R3K3 w - - 0 48

Stockfish presses me hard, but I will not surrender.

Next time, when I am in a better form, better games and better posts.

Best, Lyudmil

PS. And, of course, many many thanks again for the possibility to have a free super-grandmaster at home. Many tennis enthusiasts would have to pay enormous amount of money to spar for half an hour or even minutes against the best world players; football enthusiasts would have to only dream to be part of a sparring game with the best world players, but we, humble chess enthusiasts, are really blessed to have what no other persons have: the ability to challenge for free even the world champions!!

[carldaman]

Interesting experiment, Lyudmil. If I may suggest something for your next match, for the sake of comparison, you could adjust SF's contempt setting. I am assuming that a positive contempt value will make it avoid draws (I know some engines require a negative contempt value for this).

Cheers,
CL

[Lyudmil Tsvetkov]

Interesting experiment, Lyudmil. If I may suggest something for your next match, for the sake of comparison, you could adjust SF's contempt setting. I am assuming that a positive contempt value will make it avoid draws (I know some engines require a negative contempt value for this).

Cheers,
CL

Hi Carl.
Thanks for the post.
Still thinking on your storming pawns post, think will need about a month to analyse it.

So you would like for Stockfish to be contemptuous of me
My problem is that I do not understand neither positive, nor negative contempt. What is your take: that Stockfish will play better or worse with contempt against me?

[carldaman]

Interesting experiment, Lyudmil. If I may suggest something for your next match, for the sake of comparison, you could adjust SF's contempt setting. I am assuming that a positive contempt value will make it avoid draws (I know some engines require a negative contempt value for this).

Cheers,
CL

Hi Carl.
Thanks for the post.
Still thinking on your storming pawns post, think will need about a month to analyse it.

So you would like for Stockfish to be contemptuous of me
My problem is that I do not understand neither positive, nor negative contempt. What is your take: that Stockfish will play better or worse with contempt against me?

Hah! I think it may well avoid draws better, if it works as expected, but beyond that we'll have to see if it can play or score better. This is what experiments are for

I haven't heard much discussion about SF's contempt. Houdini, on the other hand, is a different engine with its contempt changed.

Re: the storming pawn analysis, I think the lines I analyzed are promising for White, and some improvements may even be possible. It's sad that only an old engine like Deep Fritz 10 would consider the correct and thematic plans based on advancing White's b-pawn. Makes me wonder if we're really seeing positional progress in today's engines. I will be more impressed with the newest Stockfish and whatnot, etc. when I see them playing for the pawn storm with b4, or when they respect the opponents' pawn storm potential, and not play junk moves like Qd1-h5-d1 as it did in my tests (and described in my other older post).

CL

[Lyudmil Tsvetkov]

Interesting experiment, Lyudmil. If I may suggest something for your next match, for the sake of comparison, you could adjust SF's contempt setting. I am assuming that a positive contempt value will make it avoid draws (I know some engines require a negative contempt value for this).

Cheers,
CL

Hi Carl.
Thanks for the post.
Still thinking on your storming pawns post, think will need about a month to analyse it.

So you would like for Stockfish to be contemptuous of me
My problem is that I do not understand neither positive, nor negative contempt. What is your take: that Stockfish will play better or worse with contempt against me?

Hah! I think it may well avoid draws better, if it works as expected, but beyond that we'll have to see if it can play or score better. This is what experiments are for

I haven't heard much discussion about SF's contempt. Houdini, on the other hand, is a different engine with its contempt changed.

Re: the storming pawn analysis, I think the lines I analyzed are promising for White, and some improvements may even be possible. It's sad that only an old engine like Deep Fritz 10 would consider the correct and thematic plans based on advancing White's b-pawn. Makes me wonder if we're really seeing positional progress in today's engines. I will be more impressed with the newest Stockfish and whatnot, etc. when I see them playing for the pawn storm with b4, or when they respect the opponents' pawn storm potential, and not play junk moves like Qd1-h5-d1 as it did in my tests (and described in my other older post).

CL

Hi Carl.

Please, do not speak of junk moves, because Marco may well be reading this...

I played a game with contempt set to 50 (highest possible), that I lost, coming to the important conclusion in the process that Stockfish with contempt set to 50 plays exactly like Houdini.

I do not believe in contempt. Thorough tests have shown that this is a worse setting than the default. It might wotk better, but only under certain conditions, for example against humans when they have little time, because the arising positions require many calculations. With normal time control this should be a bad setting even against humans, I am sure, at least suboptimal. Why play suboptimal? Certan conditions are an artificial concept. The default is always the best.

But I understand, you want the engine to beat me easily, but then I might change the TC to long and very long

[carldaman]

Interesting experiment, Lyudmil. If I may suggest something for your next match, for the sake of comparison, you could adjust SF's contempt setting. I am assuming that a positive contempt value will make it avoid draws (I know some engines require a negative contempt value for this).

Cheers,
CL

Hi Carl.
Thanks for the post.
Still thinking on your storming pawns post, think will need about a month to analyse it.

So you would like for Stockfish to be contemptuous of me
My problem is that I do not understand neither positive, nor negative contempt. What is your take: that Stockfish will play better or worse with contempt against me?

Hah! I think it may well avoid draws better, if it works as expected, but beyond that we'll have to see if it can play or score better. This is what experiments are for

I haven't heard much discussion about SF's contempt. Houdini, on the other hand, is a different engine with its contempt changed.

Re: the storming pawn analysis, I think the lines I analyzed are promising for White, and some improvements may even be possible. It's sad that only an old engine like Deep Fritz 10 would consider the correct and thematic plans based on advancing White's b-pawn. Makes me wonder if we're really seeing positional progress in today's engines. I will be more impressed with the newest Stockfish and whatnot, etc. when I see them playing for the pawn storm with b4, or when they respect the opponents' pawn storm potential, and not play junk moves like Qd1-h5-d1 as it did in my tests (and described in my other older post).

CL

Hi Carl.

Please, do not speak of junk moves, because Marco may well be reading this...

I played a game with contempt set to 50 (highest possible), that I lost, coming to the important conclusion in the process that Stockfish with contempt set to 50 plays exactly like Houdini.

I do not believe in contempt. Thorough tests have shown that this is a worse setting than the default. It might wotk better, but only under certain conditions, for example against humans when they have little time, because the arising positions require many calculations. With normal time control this should be a bad setting even against humans, I am sure, at least suboptimal. Why play suboptimal? Certan conditions are an artificial concept. The default is always the best.

But I understand, you want the engine to beat me easily, but then I might change the TC to long and very long

I think it is generally accepted that a certain level of contempt is beneficial (to the engine's performance) when playing against lower rated opponents (say, 100-200 points or more), but not when playing close to equal opposition. The (positive) contempt setting generally forces the engine to take more risks to skirt around potential draws. Taking more risks tends to work less against roughly equal opposition, so as you say, the play would seem to become suboptimal, but only against such equal opponents.

This is also noticed with Houdini, which now suffers when facing SF or Komodo, with its default contempt, which is set at =1 (not zero). However, the rating of H3 has been built on the benefits of C=1, which works very well when facing any other engine than these two.

This makes default a somewhat controversial topic. As a human, I tend to favor having contempt on, since I'm much weaker and the engine will give me a truer/harder test this way, and that is what I value most. As for testers of engine vs. engine, this issue is a bit of a dilemma, however. Do we change Houdini's default when playing equal engines?! It seems that H3 with C=0, or C=2 is a different entity. H3 C=0 does well vs latest SF, but H3 C=1 (default) now struggles vs SF. Which is the real Houdini now ?!

Maybe what is default can be determined differently in the future, with a way to query the strength of the opponent, but we're not there yet.

[Lyudmil Tsvetkov]

I played some more games against Stockfish 4, but obviously am in a bad form, as still have been unable to win a proper game. Draws are frequent though.

Below 2 more, probably not entirely uninteresting ones.

[pgn][PlyCount "85"]
[Event "Blitz 2m+2s"]
[Site "Sofia"]
[Date "2013.09.13"]
[White "Stockfish 4 64 SSE4.2"]
[Black "Tsvetkov, Lyudmil"]
[Result "1/2-1/2"]
[ECO "B11"]
[TimeControl "120+2"]
[Annotator "Tsvetkov,Lyudmil"]
[MLNrOfMoves "42"]
[MLFlags "000100"]

{1024MB, Dell XPS 4Cores} 1. e4 {[%emt 0:00:04]} 1... c6 {[%emt 0:00:04]} 2.
Nf3 {[%emt 0:00:08]} 2... d5 {[%emt 0:00:02]} 3. Nc3 {[%emt 0:00:03]} 3... Bg4
{[%emt 0:00:05]} 4. d4 {[%emt 0:00:10]} 4... e6 {[%emt 0:00:05]} 5. Bd3
{[%emt 0:00:07]} 5... Nf6 {[%emt 0:00:02]} 6. O-O {[%emt 0:00:03]} 6... Be7
{[%emt 0:00:02]} 7. Bf4 {[%emt 0: 00:04]} 7... O-O {[%emt 0:00:03]} 8. h3
{[%emt 0:00:07]} 8... Bh5 {[%emt 0:00:07]} 9. a3 {[%emt 0:00:01]} 9... Nbd7
{[%emt 0:00:03]} 10. e5 {[%emt 0:00:05]} 10... Ne8 {[%emt 0:00: 03]} 11. Be2
{[%emt 0:00:03]} 11... a6 {[%emt 0:00:32]} 12. Qd2 {[%emt 0:00:04]} 12... Rc8
{[%emt 0:00:07]} 13. Be3 {[%emt 0:00:08]} 13... c5 {[%emt 0:00:06]} 14. Rfd1
{[%emt 0:00:00]} 14... c4 {[%emt 0:00:03]} 15. Bg5 {[%emt 0:00:05]} 15... Bxg5
{[%emt 0:00:25]} 16. Qxg5 {[%emt 0:00:09]} 16... Bxf3 {[%emt 0:00:33]} 17. Qxd8
{[%emt 0:00:03]} 17... Rxd8 {[%emt 0:00:02]} 18. Bxf3 {[%emt 0:00:04]} 18... b5
{[%emt 0:00:02]} 19. g3 {[%emt 0: 00:05]} 19... Nb6 {[%emt 0:00:02]} 20. Kg2
{[%emt 0:00:06]} 20... Nc7 {[%emt 0:00:06]} 21. Rdb1 {[%emt 0:00:03]} 21... Rb8
{[%emt 0:00:12]} 22. h4 {[%emt 0:00:00]} 22... f5 {[%emt 0: 00:12]} 23. Ne2
{[%emt 0:00:05]} 23... g6 {[%emt 0:00:07]} 24. b3 {[%emt 0:00:06]} 24... Kf7
{[%emt 0:00:09]} 25. Nf4 {[%emt 0:00:00]} 25... Ke7 {[%emt 0:00:19]} 26. Rh1
{[%emt 0:00:04]} 26... a5 {[%emt 0:00:13]} 27. Nh3 {[%emt 0:00:00]} 27... h6
{[%emt 0:00:17]} 28. Nf4 {[%emt 0:00:06]} 28... Rg8 {[%emt 0:00:13]} 29. Be2
{[%emt 0:00:00]} 29... a4 {[%emt 0:00:15]} 30. b4 {[%emt 0:00:07]} 30... Nd7
{[%emt 0:00:10]} 31. c3 {[%emt 0: 00:06]} 31... Rbf8 {[%emt 0:00:02]} 32. h5
{[%emt 0:00:04]} 32... g5 {[%emt 0:00:04]} 33. Ng6+ {[%emt 0:00:00]} 33... Kf7
{[%emt 0:00:03]} 34. Nxf8 {[%emt 0:00:07]} 34... Kxf8 {[%emt 0:00:04]} 35. f3
{[%emt 0:00:10]} 35... Ke7 {[%emt 0:00:09]} 36. Rae1 {[%emt 0: 00:00]} 36... Kf7
{[%emt 0:00:02]} 37. Bd1 {[%emt 0:00:04]} 37... Ke7 {[%emt 0:00:04]} 38. g4
{[%emt 0:00:05]} 38... f4 {[%emt 0:00:04]} 39. Bc2 {[%emt 0:00:00]} 39... Kd8
{[%emt 0: 00:05]} 40. Bh7 {[%emt 0:00:04]} 40... Rf8 {[%emt 0:00:03]} 41. Bg6
{[%emt 0:00:01]} 41... Ke7 {[%emt 0:00:03]} 42. Rb1 {[%emt 0:00:05]} 42... Rb8
{[%emt 0:00:05]} 43. Kf2 {[%emt 0:00:07]} 1/2-1/2

[PlyCount "76"]
[Event "Blitz 2m+2s"]
[Site "Sofia"]
[Date "2013.09.13"]
[White "Tsvetkov, Lyudmil"]
[Black "Stockfish 4 64 SSE4.2"]
[Result "1/2-1/2"]
[ECO "D00"]
[TimeControl "120+2"]
[Annotator "Tsvetkov,Lyudmil"]
[MLNrOfMoves "38"]
[MLFlags "000100"]

{1024MB, Dell XPS 4Cores} 1. e3 {[%emt 0:00:00]} 1... e6 {[%emt 0:00:08]} 2. d4
{[%emt 0:00:02]} 2... c5 {[%emt 0:00:04]} 3. Bd3 {[%emt 0:00:02]} 3... d5
{[%emt 0:00:04]} 4. c3 {[%emt 0:00:02]} 4... Nf6 {[%emt 0:00:06]} 5. f4
{[%emt 0:00:02]} 5... Bd6 {[%emt 0:00:07]} 6. Nf3 {[%emt 0:00:02]} 6... O-O
{[%emt 0:00:03]} 7. O-O {[%emt 0:00:03]} 7... b6 {[%emt 0:00:03]} 8. Qe2
{[%emt 0:00:06]} 8... Nc6 {[%emt 0:00:04]} 9. Nbd2 {[%emt 0:00:07]} 9... Bb7
{[%emt 0:00:13]} 10. Ne5 {[%emt 0:00:02]} 10... Qc7 {[%emt 0:00: 04]} 11. g4
{[%emt 0:00:05]} 11... Ne7 {[%emt 0:00:03]} 12. Qf3 {[%emt 0:00:07]} 12... h6
{[%emt 0:00:07]} 13. h4 {[%emt 0:00:05]} 13... h5 {[%emt 0:00:04]} 14. g5
{[%emt 0:00: 22]} 14... Bxe5 {[%emt 0:00:05]} 15. fxe5 {[%emt 0:00:17]} 15...
Ng4 {[%emt 0:00:00]} 16. Qg3 {[%emt 0:00:12]} 16... Qd7 {[%emt 0:00:07]} 17. Nf3
{[%emt 0:00:10]} 17... Qc8 {[%emt 0:00:07]} 18. Nh2 {[%emt 0:00:21]} 18... Nxh2
{[%emt 0:00:00]} 19. Qxh2 {[%emt 0:00: 09]} 19... Ba6 {[%emt 0:00:00]} 20. Qe2
{[%emt 0:00:05]} 20... Bxd3 {[%emt 0:00:05]} 21. Qxd3 {[%emt 0:00:01]} 21... Nf5
{[%emt 0:00:06]} 22. Rf4 {[%emt 0:00:03]} 22... Qc6 {[%emt 0:00:04]} 23. Bd2
{[%emt 0:00:04]} 23... c4 {[%emt 0:00:05]} 24. Qc2 {[%emt 0:00:04]} 24... Ng3
{[%emt 0:00:01]} 25. Kg2 {[%emt 0:00:13]} 25... Ne4 {[%emt 0:00:00]} 26. Raf1
{[%emt 0:00:18]} 26... Rab8 {[%emt 0:00:00]} 27. Be1 {[%emt 0:00:04]} 27... b5
{[%emt 0:00: 02]} 28. a3 {[%emt 0:00:03]} 28... g6 {[%emt 0:00:04]} 29. Kh1
{[%emt 0:00:21]} 29... Qb7 {[%emt 0:00:05]} 30. Kg2 {[%emt 0:00:03]} 30... Rbc8
{[%emt 0:00:04]} 31. R4f3 {[%emt 0:00:06]} 31... a6 {[%emt 0:00:08]} 32. Rf4
{[%emt 0:00:01]} 32... Kh7 {[%emt 0:00:16]} 33. R4f3 {[%emt 0:00:08]} 33... Rb8
{[%emt 0:00:06]} 34. Rf4 {[%emt 0:00:02]} 34... Kg8 {[%emt 0:00:02]} 35. R4f3
{[%emt 0:00:02]} 35... Rfc8 {[%emt 0:00:06]} 36. Rf4 {[%emt 0:00: 24]} 36... Rc7
{[%emt 0:00:05]} 37. R1f3 {[%emt 0:00:04]} 37... Rbc8 {[%emt 0:00:03]} 38. Rf1
{[%emt 0:00:02]} 38... Rd7 {[%emt 0:00:03]} 1/2-1/2
[/pgn]

[d]1r6/2nnk3/4p1Bp/1p1pP1pP/pPpP1pP1/P1P2P2/5K2/1R5R b - - 0 43

Interesting if Stockfish recognizes this positio as a straightforward draw. Do I have to play 50 more completely meaningless moves if there is not any kind of adjudication in similar positions?

[carldaman]

Lyudmil,
I'm glad you're running these tests, because they uncover an age-old weakness of chess engines. OK, maybe now they're harder to beat, but getting a draw is not so difficult if you play less ambitiously, just to lock up the position and trade pieces off.

You have to wonder how a 2700+ GM (or 2500+ even) would have a field day playing against these weaknesses, but we don't see too many contests between super-GMs and top engines. It seems that most advancements have been in the area of faster search, and better eval in open and semi-open middlegames, where humans are no match for the engine, and also in the endgame, where even top engines used to be weak, but not any longer.

Unfortunately, closed and semi-closed positions are still just as difficult for engines to deal with, something people forget about, since most engine games are played against other engines, and these weaknesses don't get punished.

It seems that you can draw against SF regularly with this strategy, so a human GM amenable to the same method would probably be able to draw almost at will. This would have to call into question the stratospheric ratings of these top engines. I have a tough time understanding how the SSDF could enlist Swedish players, many of them masters, to play routinely against computer opponents 20 years ago, but now that is suddenly 'impossible'. If there's a will, there ought to be a way to better testing, and that should include human vs engine testing. I believe that is a must, if there is to be any relevancy to engine ratings.

If (hypothetically) Engine A does poorly against humans, but Engine B destroys humans regardless of their anti-computer strategy, then Engine B would be very valuable to have (for both sparring and analysis), regardless of whether Engine A is higher rated than Engine B. Most people don't get this.

Engine authors have a real opportunity here. Instead of seeing the end of commercial engine development, they should strive to create engines that play good anti-human chess. The claim is that they do already, but it is because they hide behind opening books that help them in avoiding troublesome positions. This is a sham, in my view. Weaknesses should be dealt with, not covered up!


Regarding your last question, I think only a few programs may be able to detect the drawishness of the position. I haven't had time to check, but Shredder and Chiron come to mind, though they're "weaker" engines, ya know!

[Lyudmil Tsvetkov]

Lyudmil,
I'm glad you're running these tests, because they uncover an age-old weakness of chess engines. OK, maybe now they're harder to beat, but getting a draw is not so difficult if you play less ambitiously, just to lock up the position and trade pieces off.

You have to wonder how a 2700+ GM (or 2500+ even) would have a field day playing against these weaknesses, but we don't see too many contests between super-GMs and top engines. It seems that most advancements have been in the area of faster search, and better eval in open and semi-open middlegames, where humans are no match for the engine, and also in the endgame, where even top engines used to be weak, but not any longer.

Unfortunately, closed and semi-closed positions are still just as difficult for engines to deal with, something people forget about, since most engine games are played against other engines, and these weaknesses don't get punished.

It seems that you can draw against SF regularly with this strategy, so a human GM amenable to the same method would probably be able to draw almost at will. This would have to call into question the stratospheric ratings of these top engines. I have a tough time understanding how the SSDF could enlist Swedish players, many of them masters, to play routinely against computer opponents 20 years ago, but now that is suddenly 'impossible'. If there's a will, there ought to be a way to better testing, and that should include human vs engine testing. I believe that is a must, if there is to be any relevancy to engine ratings.

If (hypothetically) Engine A does poorly against humans, but Engine B destroys humans regardless of their anti-computer strategy, then Engine B would be very valuable to have (for both sparring and analysis), regardless of whether Engine A is higher rated than Engine B. Most people don't get this.

Engine authors have a real opportunity here. Instead of seeing the end of commercial engine development, they should strive to create engines that play good anti-human chess. The claim is that they do already, but it is because they hide behind opening books that help them in avoiding troublesome positions. This is a sham, in my view. Weaknesses should be dealt with, not covered up!


Regarding your last question, I think only a few programs may be able to detect the drawishness of the position. I haven't had time to check, but Shredder and Chiron come to mind, though they're "weaker" engines, ya know!

Hi Carl.
Thanks for the answer.

I am weak, but not such a weak player after all. My best FIDE rating was a bit over 2100 (but only because I played too few FIDE rated games), my Bulgarian rating a bit over 2200 (there is basically no distinction between Bulgarian and FIDE rating, as with USCF ratings, with Bulgarian only the time control was without increment, say 1 hour per the entire game; once they did not have digital clocks att all tournaments ), but those ratings are based just on a couple of years' record, very few tournaments, playing after work hours, etc.
My last 2 complete tournaments (for Bulgarian rating) end 2005/beginning 2006 exhibited a performance rating of some 2400 points, with wins against a couple of GMs, IMs, etc. If you are able to do this once or twice, you should be able to do it on a regular basis, but the fact is I stopped playing competitive chess after that. In the meanwhile, I have trained quite seriously (but only privately), first with Rybka, and then especially with the latest generation of chess engines, meaning I could have added a couple of 'unregistered' elos in the process.

But that is completely unimportant for me, as I do not compete and do not have the wish to do so in the future. My only point was to suggest that I am not so weak after all.

But of course, you got it right on many fronts. I think a strong GM with the necessary preparation (Kasparov, the 2004 Bilbao guys, Kramnik, etc., definitely almost did not have a clue how they should play against engines, and that was one of the reasons for the bad results) could be able to play much more successfully with engines. But in order to prepare, you need a couple of years of specific engine experience, for which most GMs do not have time or will to do.

I think the stronger the engine, the more difficult it is to play against, no matter how it plays. Non-standard entities might be sometimes more dangerous to humans, but they also allow a significant amount of opportunities, so basically it should be easier to win against those. Spark comes to my mind: sometimes it plays fantastic, but at other times, looking at its pawn structure, you see the engine does not have a clue about pawns, much easier to win against it in general as against any top-rated engine.

[Lyudmil Tsvetkov]

And another 2 games, I think very illustrative.

In the first one you see how easy it is to draw: just exchange pieces, then exchange more pieces, and it is a draw.

In the second one you see how incredibly easy the engine can crush you: you just need to make a small tactical mistake, and it is over, although the position might suggest a somewhat different story. Nothing to do there, the engines simply make very small number of bigger mistakes, while humans do a lot of bigger mistakes when short on time.

[pgn][PlyCount "113"]
[Event "Blitz 2m+2s"]
[Site "Sofia"]
[Date "2013.09.13"]
[White "Stockfish 4 64 SSE4.2"]
[Black "Tsvetkov, Lyudmil"]
[Result "1/2-1/2"]
[ECO "C01"]
[TimeControl "120+2"]
[Annotator "Tsvetkov,Lyudmil"]
[MLNrOfMoves "56"]
[MLFlags "000100"]

{1024MB, Dell XPS 4Cores} 1. e4 {[%emt 0:00:04]} 1... e6 {[%emt 0:00:04]} 2.
Nf3 {[%emt 0:00:04]} 2... d5 {[%emt 0:00:02]} 3. exd5 {[%emt 0:00:02]} 3... exd5
{[%emt 0:00: 01]} 4. d4 {[%emt 0:00:05]} 4... Nf6 {[%emt 0:00:01]} 5. Bd3
{[%emt 0:00:06]} 5... Bd6 {[%emt 0:00:02]} 6. O-O {[%emt 0:00:05]} 6... O-O
{[%emt 0:00:01]} 7. c4 {[%emt 0:00: 26]} 7... c6 {[%emt 0:00:03]} 8. c5
{[%emt 0:00:05]} 8... Bc7 {[%emt 0:00:02]} 9. Nc3 {[%emt 0:00:03]} 9... Bg4
{[%emt 0:00:03]} 10. h3 {[%emt 0:00:04]} 10... Bh5 {[%emt 0:00: 02]} 11. Re1
{[%emt 0:00:12]} 11... Nbd7 {[%emt 0:00:03]} 12. g4 {[%emt 0:00:04]} 12... Bg6
{[%emt 0:00:02]} 13. Bxg6 {[%emt 0:00:03]} 13... hxg6 {[%emt 0:00:02]} 14. Bg5
{[%emt 0:00:02]} 14... Re8 {[%emt 0:00:04]} 15. Qd3 {[%emt 0:00:05]} 15... Rxe1+
{[%emt 0: 00:12]} 16. Rxe1 {[%emt 0:00:04]} 16... Qf8 {[%emt 0:00:01]} 17. Kg2
{[%emt 0:00:13]} 17... Re8 {[%emt 0:00:05]} 18. Re3 {[%emt 0:00:00]} 18... Rxe3
{[%emt 0:00:22]} 19. Qxe3 {[%emt 0:00:00]} 19... Qe8 {[%emt 0:00:06]} 20. Qxe8+
{[%emt 0:00:00]} 20... Nxe8 {[%emt 0: 00:01]} 21. b4 {[%emt 0:00:13]} 21... a6
{[%emt 0:00:03]} 22. a3 {[%emt 0:00:04]} 22... Nef6 {[%emt 0:00:18]} 23. Ne2
{[%emt 0:00:04]} 23... Ne4 {[%emt 0:00:03]} 24. Be3 {[%emt 0:00:00]} 24... Nef6
{[%emt 0:00:24]} 25. Ng5 {[%emt 0:00:04]} 25... Nh7 {[%emt 0:00: 05]} 26. Nxh7
{[%emt 0:00:04]} 26... Kxh7 {[%emt 0:00:01]} 27. a4 {[%emt 0:00:04]} 27... Kg8
{[%emt 0:00:07]} 28. Kf3 {[%emt 0:00:03]} 28... f6 {[%emt 0:00:07]} 29. Bf4
{[%emt 0:00:04]} 29... Bd8 {[%emt 0:00:11]} 30. Bd6 {[%emt 0:00:00]} 30... g5
{[%emt 0:00: 04]} 31. Ke3 {[%emt 0:00:07]} 31... Kf7 {[%emt 0:00:02]} 32. b5
{[%emt 0:00:02]} 32... axb5 {[%emt 0:00:08]} 33. axb5 {[%emt 0:00:00]} 33... Ke6
{[%emt 0:00:18]} 34. Kd3 {[%emt 0:00:05]} 34... g6 {[%emt 0:00:06]} 35. Nc3
{[%emt 0:00:04]} 35... Ba5 {[%emt 0:00:10]} 36. bxc6 {[%emt 0:00:04]} 36... bxc6
{[%emt 0:00:02]} 37. f3 {[%emt 0:00:02]} 37... Kf7 {[%emt 0:01:11]} 38. Ne2
{[%emt 0:00:04]} 38... Ke6 {[%emt 0:00:05]} 39. f4 {[%emt 0:00:03]} 39... Kf7
{[%emt 0:00:10]} 40. fxg5 {[%emt 0:00:03]} 40... fxg5 {[%emt 0:00:02]} 41. Ng1
{[%emt 0:00:01]} 41... Bd8 {[%emt 0:00:20]} 42. Nf3 {[%emt 0:00:00]} 42... Bf6
{[%emt 0:00: 04]} 43. Kd2 {[%emt 0:00:00]} 43... Nf8 {[%emt 0:00:08]} 44. Be5
{[%emt 0:00:00]} 44... Bd8 {[%emt 0:00:10]} 45. Bh2 {[%emt 0:00:04]} 45... Bf6
{[%emt 0:00:06]} 46. Kd3 {[%emt 0:00:00]} 46... Ne6 {[%emt 0:00:02]} 47. Be5
{[%emt 0:00:03]} 47... Bd8 {[%emt 0:00:03]} 48. Ke2 {[%emt 0:00:05]} 48... Ke8
{[%emt 0:00:06]} 49. Bd6 {[%emt 0:00:04]} 49... Bf6 {[%emt 0:00:04]} 50. Ke3
{[%emt 0:00:00]} 50... Kf7 {[%emt 0:00:03]} 51. Be5 {[%emt 0: 00:01]} 51... Bd8
{[%emt 0:00:02]} 52. Bg3 {[%emt 0:00:02]} 52... Bf6 {[%emt 0:00:02]} 53. Bb8
{[%emt 0:00:03]} 53... Ke8 {[%emt 0:00:02]} 54. Bh2 {[%emt 0:00:04]} 54... Kf7
{[%emt 0:00:01]} 55. Bg3 {[%emt 0:00:10]} 55... Ke8 {[%emt 0:00:02]} 56. Be5
{[%emt 0:00:03]} 56... Bd8 {[%emt 0:00:04]} 57. Bh2 {[%emt 0:00:00]} 1/2-1/2

[PlyCount "62"]
[Event "Blitz 2m+2s"]
[Site "Sofia"]
[Date "2013.09.13"]
[White "Tsvetkov, Lyudmil"]
[Black "Stockfish 4 64 SSE4.2"]
[Result "0-1"]
[ECO "E11"]
[TimeControl "120+2"]
[Annotator "Tsvetkov,Lyudmil"]
[MLNrOfMoves "31"]
[MLFlags "100100"]

{1024MB, Dell XPS 4Cores} 1. d4 {[%emt 0:00:00]} 1... Nf6 {[%emt 0:00:05]} 2.
c4 {[%emt 0:00:02]} 2... e6 {[%emt 0:00:19]} 3. Nf3 {[%emt 0:00:02]} 3... Bb4+
{[%emt 0:00: 03]} 4. Bd2 {[%emt 0:00:02]} 4... Bxd2+ {[%emt 0:00:04]} 5. Nbxd2
{[%emt 0:00:06]} 5... O-O {[%emt 0:00:09]} 6. e3 {[%emt 0:00:02]} 6... b6
{[%emt 0:00:03]} 7. Be2 {[%emt 0:00:03]} 7... c5 {[%emt 0:00:05]} 8. O-O
{[%emt 0:00:05]} 8... Bb7 {[%emt 0:00:06]} 9. Rc1 {[%emt 0:00:12]} 9... d6
{[%emt 0:00:00]} 10. Qc2 {[%emt 0:00:04]} 10... Nc6 {[%emt 0: 00:05]} 11. Rfd1
{[%emt 0:00:05]} 11... cxd4 {[%emt 0:00:07]} 12. exd4 {[%emt 0:00: 03]} 12...
Rc8 {[%emt 0:00:14]} 13. Qb1 {[%emt 0:00:04]} 13... d5 {[%emt 0:00:07]} 14. c5
{[%emt 0:00:12]} 14... bxc5 {[%emt 0:00:04]} 15. dxc5 {[%emt 0:00:05]} 15... e5
{[%emt 0:00: 11]} 16. b4 {[%emt 0:00:05]} 16... e4 {[%emt 0:00:03]} 17. b5
{[%emt 0:00:17]} 17... exf3 {[%emt 0:00:04]} 18. Nxf3 {[%emt 0:00:02]} 18... Ne7
{[%emt 0:00:03]} 19. Qb4 {[%emt 0:00:28]} 19... Rc7 {[%emt 0:00:04]} 20. Nd4
{[%emt 0:00:03]} 20... Ng6 {[%emt 0:00:04]} 21. g3 {[%emt 0:00:12]} 21... Re8
{[%emt 0:00:10]} 22. a4 {[%emt 0:00:19]} 22... Qe7 {[%emt 0:00:05]} 23. Rc2
{[%emt 0:00:37]} 23... Ne4 {[%emt 0:00:05]} 24. a5 {[%emt 0: 00:09]} 24... Ne5
{[%emt 0:00:06]} 25. a6 {[%emt 0:00:08]} 25... Ba8 {[%emt 0:00:00]} 26. Nb3
{[%emt 0:00:19]} 26... Ng5 {[%emt 0:00:08]} 27. Nd4 {[%emt 0:00:10]} 27... Nc4
{[%emt 0:00:01]} 28. Bxc4 {[%emt 0:00:53]} 28... dxc4 {[%emt 0:00:06]} 29. Qxc4
{[%emt 0:00: 00]} 29... Qe1+ {[%emt 0:00:00]} 30. Rxe1 {[%emt 0:00:00]} 30...
Rxe1+ {[%emt 0:00:00]} 31. Qf1 {[%emt 0:00:02]} 31... Nh3# {[%emt 0:00:00]} 0-1
[/pgn]

In both games I used more time than usual, started thinking of something, then the second game was really extremely entertaining, but you need at least 3 to 5 minutes for each difficult move to solve it, tens of seconds will not help.

[d]2rq1rk1/pb3ppp/2n2n2/1PPp4/4p3/5N2/P2NBPPP/1QRR2K1 b - - 0 17

Difficult to make sense what happens here, I think white is slightly better.

[d]4r1k1/pbr1qppp/5nn1/1PPp4/PQ1N4/6P1/4BP1P/2RR2K1 w - - 0 23

Now, is Stockfish better here?; or a move earlier? I do not know.

[d]b3r1k1/p1r1qppp/P7/1PPpn3/1Q1Nn3/6P1/2R1BP1P/3R2K1 w - - 0 26

Here white already might be lost, but I lose it very quickly indeed. A single bigger mistake and it is over.

[d]b3r1k1/p1r2ppp/P7/1PP3n1/2QN4/6P1/2R2P1P/3Rq1K1 w - - 0 30

Well, you see, I do not see this. I do not know with longer time control, but Nc4 for black was great.

[d]b5k1/p1r2ppp/P7/1PP5/3N4/6Pn/2R2P1P/4rQK1 w - - 0 32

You say hello, and I say goodbye. Now that I see the way I have been humiliated by Stockfish, I am eager to score some wins, but I do not know when that will happen with my current form.

Maybe someone might give a try at the diagrammed positions, at least the first one or two, I think they are very interesting and deeper than what they seem, although engines might not think so.