Thursday, February 7, 2013

Non-Random Chess960 Trial Game 8: SP868

The PGN database for these trials is here.

Our journey through the opening of SP868 has concluded. Incredibly on the second run through this opening black has found an absolutely perfect defensive setup despite all odds, and I as white was unable to crack it. Towards the end of the trial the position I was seeing on the board looked so powerful for white, but actually in the cold hard light of objectivity the position was equal. Here is the trial variation as we concluded. Check the database out for greater detail and comments.

Clearly white's moves were suboptimal. The most serious question marks are:
  • 7. O-O?! white decides to try the same thematic castling idea that worked so well on our first trial run (see database)
  • 9. Ba3?! a wasted tempo since black can match it with ...Bd6. White thought that ...Bxf5 where black drops the bishop pair would have to be good for white. However black didn't need both bishops to equalize.

SP868: Black's perfect defensive technique

An important heuristic in this fiendishly difficult SP:
Mark has come up with an excellent way of looking at this fiendishly difficult start, SP868. The idea is that black's apparent inability to castle is almost equivalent of being a pawn down. So black's strategy is to sacrifice a pawn and retain the ability to castle thus connecting the rooks and allowing more active play.

Sounds good! From what I am seeing of this SP, there is a serious game of cat and mouse at work. First strategy to rule out all together in my view, is for black to castle queenside, or to slide the king over to the queenside. Once white sees this strategy, they simply castle kingside and hit the queenside hard. Really hard. It might possibly work under special conditions as idea no.2 shows.

To me the strategy is for black to wait until white commits to the kingside, and then slide or castle the king to the kingside very carefully themselves. The combined QBB combination now tends to equalise, because both white and black's king will be the subject of their crossfire, with black's king a little less secure than white's.

Now that we are studying this SP in greater detail, here are a collection of ideas:

SP868 idea no.1: 1)...e6!?

The idea here is that the e7 square will become a temporary safe haven for black's king. It's worthy of some serious consideration!

SP868 idea no.2: ...2.Nh5!?

One of the fascinating things in the above variation is how the black king transfers over to the queen-side using the bishops and queen as protection rather than the traditional pawn shelter. (Note that 23.b5?! is met with  23...Nxe4!)

SP868 idea no.3: The Amazing Military Knight defenders!

One of the fascinating things about the above variation is how black manages to get the time to reorganise the Military Knights and how well they do as a cooperative unit. White still has an edge but black has a lot of play and the slightest slip from white and a lot of the advantage is gone.


  1. In the CCRL, there seems to be 99 games with only 54 wins for white after 1.Ng3. Interesting...


    ...White will rush attack the undefended g7 pawn at almost no cost, either disrupting black's ability to castle short and connect the rooks, or cause a serious disturbance in black's development harmony - Harry

  2. Subscribed! I want to look at the CCRL games before making a move, so give me a few days. - Mark

  3. Egads! Is SP868 the long sought SP that busts Fischer's idea?...

    1.Ng3 Nf6

    ...Makes space for the Rook to protect the g-Pawn. Over 90% of the CCRL games continue with this move. Black loses the right to O-O, but with all of the diagonal pieces bearing down on that side, it's not a great loss. Castling O-O-O will also be problematic. - Mark

  4. The snow storms arrived in New England!

    The reasoning above re "Black looses the right but..." is very true. There is beauty inside the most ugly of situations. Chess960 is more a philosophy than a game.

    1.Ng3 Nf6 2.Nf5 (if you play Rg8, I'll probably disappear for a few days to think what to play next...)

    ...Could be considered a partially free move but eventually black can play ...e6 driving the knight back, freeing the e7 square and taking control of d5 - Harry

  5. I have a lot of family in and near New England. They were hit pretty hard, but it's over now. Snow people know how to deal with snow...

    1.Ng3 Nf6 2.Nf5 Rg8

    ...I didn't really look at the alternative you suggested. Giving up a Pawn while losing the castling privilege isn't my style. It looks like White can force the Nf6 off the long diagonal, when the Knight escapes to h5.

    Call Black's plan, 'Shoveling out after the snow storm', because that's what I'm going to be doing for the next few moves.- Mark

  6. Sorry for the long delay Mark. I've been looking for the silver bullet in this SP but I don't think there is a tactical shot. Instead white builds up a gradual attack on black's stranded king.

    1.Ng3 Nf6 2.Nf5 Rg8 3.b3

    ...Restricting the f6 knight and setting the stage for a favorable fight in the center if that is what black wants. The irony of this move is that white does not immediately play out a center pawn because I think whoever plays out into the center first get's chopped off somewhat! White will try to open up lines in the center and c-side against the stranded king and can even drop a pawn if it helps clear out more lines. - Harry

    Check out the database if you want to see some of the lines I've calculated and documented. Enjoy the next ten moves of what will be something akin to mild torture!

  7. Yes, Black's main concern for the opening in this game is what to do with the King...

    2.Nf5 Rg8 3.b3 Ng6

    ...Playing the only obvious move and waiting to see how White continues. - Mark

  8. If you can deal with this SP, you will be in good stead for the rest of your Chess960 play!

    2.Nf5 Rg8 3.b3 Ng6 4.Nf3

    Volleying the ball back at black. Any opening of lines to black's king are dangerous and so white waits for black to commit to it. Also prepares to connect the rooks by castling which prevents black from being able to open the d-file since an exchange of rooks would leave black's king even more exposed - Harry

  9. Black can also interfere with White's castling...

    3.b3 Ng6 4.Nf3 Nf4

    ...Defers any Queenside decisions for another move. - Mark

  10. You might disagree with the move exclamation so remove it please. But I've never seen anything like this position before in my life! Looks like this SP might make it to 20 moves with a good contest lets see!

    3.b3 Ng6 4.Nf3 Nf4 5.O-O!?

    ...Prioritizing connecting the rooks over a more complicated scenario if Ne3. Ignoring the castling problem, white is at least a tempo up in development - Harry

  11. That's probably the best move you've made in our little series. I didn't see it coming...

    4.Nf3 Nf4 5.O-O! d6

    ...Grabbing the Pawn is suicide, so Black must scramble to find counterplay. Logically, it should be against the castled King. - Mark

  12. It's interesting how we are blind to certain moves but not others.

    Chess960 is wonderful training for knight moves as we are seeing in this game. As a beginner the most difficult thing for me was seeing backward (retrograde) moves of the knight. In this game, there might have been a sniff of a chance to grab the pawn with Nxe2 then quickly fallback with Nf4/Ng6, but in any case that is the kind of knight move training that is common in Chess960 in the first 100 and last 100 SP's of the series.

    So we are learning another characteristic of 960. I've revealed my castling move and now you have a clear plan. But you must execute fantastic defensive skills to pull it off....

    In some 960 starts we see very little ambiguity in plan but very great technical skills. In other SP's we see huge ambiguity for want of a plan (greater than SP518), with much less technical challenges at the start (that is debatable but anyway!)

    So it is possible that SP518 is a classic start because it blends plan ambiguity and opening technical skills very well.

    4.Nf3 Nf4 5.O-O! d6 6.Ng3

    ...The difference in relative king safety after only six moves is astonishing - Harry

  13. I might eventually pull the plug on this and take back 4...Nf4, but let's give it a whirl for now...

    5.O-O! d6 6.Ng3 b5

    ...Looks like the only move to give Black counterplay. I'm not as concerned about the safety of the Black King as I am about connecting the Rooks. - Mark

  14. Sure no problems glad that you are willing to take back and try again if need be.

    However the move you have played leads to some spectacular defensive chess the likes I have never seen this early in the game! In case you don't check the database (which has the moves), here is the line 7.c4 h5!? 8.Rfe1 h4!? 9.Nxh4 bxc4 10.bxc4 g5!? 11.Nf3 Nxg2!? 12.Kxg2 g4 13.d4 gxf3+ 14.exf3 d5 15.c5 c6 16.Qb2 Bc7 17.Kh1 Qb7 18.Nf5 e6 {black finds a way to connect the rooks via the Kd7 square!}

    So black begins an incredibly active idea with h5/h4/g5/g4. First open up white's king and then build a fortress using d7 as a safe square for black's king.


    Back to real life:
    5.O-O! d6 6.Ng3 b5 7.c4

    ...White is happy to open files (any will do) since the rooks are connected - Harry

    All I can say so far is that this SP is an amazing challenge for black....and white. We should remember that it has been really challenging for white to find the most efficient plan here as well. One less than accurate move and white looses much of the edge.

    Again, happy to rewind if you think it is necessary.

  15. There is no reason to rush into tactics. Better to develop slowly against the White King, who isn't going anywhere else...

    6.Ng3 b5 7.c4 bxc4

    ...Black wants a fluid Pawn structure on the Queenside, keeping the diagonals as open as possible. - Mark

  16. Good point. Tactics are risky and in the last variation I found a surprising move for white that spoils the variation for black. (15.h4!)

    Here we are in a philosophical disagreement which is interesting and good! You think opening up the queenside is a good idea considering the difficulties, and I think it is not.

    6.Ng3 b5 7.c4 bxc4 8.bxc4

    ...Opening up the queenside means more access to the king...hopefully - Harry

  17. Here I had planned to play 8...h5 9.d4 g5 {P1} 10.d5 h4, but then I discovered 11.Bb2!, and Black is busted. Then I saw that 8.d4 g5 9.bxc4 h5 {P1 again} forced the same position. Your 8.bxc4 also gives White an excellent game, but 8.d4 appears to win by force.

    My 4...Nf4 was meant to disrupt White's game, but it turned out to be a poor move because of 5.O-O!. I'm going to abandon the current variation and look for a better 4th move for Black.

    Congratulations on a well played game! Please give me a few days to find a different 4th move. The first time I looked at that position, I couldn't find anything particularly attractive, but hopefully I understand it better now. - Mark

  18. No probs this is fiendishly difficult. Will have a look at the 4th move afresh myself. Boy that move 11.Bb2! is a beauty isn't it? In hindsight (only), it is quite easy to see. Surprising that it takes Houdini-3 so long to find it.

    Looks like Bb2 will be one of the main themes in this start because of a potentially weak f6 square so close to black's king.

    I think the d7 square could be a way to build a king fortress. After all, it should be possible to put the king on d7 and develop the light bishop on b7/a6 still aiming the queen and bishops at white's kingside.

    Will we have to rewind the moves any further?

    I wonder if you missing 5.O-O! was because of a weakness of perceiving castling moves in Chess960 you have?

  19. Second verse, better than the first...

    3.b3 Ng6 4.Nf3 c5

    ...Rolling the position back hasn't made it any easier. While I was working on it I realized that, since Black will never castle, Black is essentially a Pawn down, at least according to chess theory. This means that first moves where Black sacrifices a Pawn are no worse than giving up the castling O-O privilege, as I did and as most of the CCRL engines have done.

    I went back to the first move and found a new idea involving a Pawn sac. Maybe after we test 4...c5, we can give the other idea a try. - Mark

  20. Mark I'm totally inspired by your little heuristic there. Thinking in this way totally opens the mind up to possibilities! I think we tend to get negative without a heuristic like this. Well done! I think it's a very important idea for Chess960 theory and it applies to more SP's than this I think.

    That said, I'm delayed in my move because I'm playing through a few different ideas and posting them on my main blog as they come. Could you take a look at them when you have the time as I post more?

    I'll play my move in the original variation soon.

  21. Re 'first strategy to rule out all together, is for Black to castle Queenside', I agree. To develop the diagonal pieces, both players have to move some Queenside Pawns. This will render that wing too airy for a King. - Mark

  22. Well, your last move 4...c5 seems pretty strong to me! It put me into a deep think between c4 or d4? At this rate I'm not sure we will need to try another opening for black. But if we did....I place a bet that you are thinking 1.Ng3...e6!? That's the most lateral alternative I could find!

    3.b3 Ng6 4.Nf3 c5 5.d4

    ...Wasting a tempo in a strict sense. Tries to take command of the center more directly - Harry

  23. I was looking at 1.Ng3 g6. Black's main tactical problem is that ...e6 always leaves the Nf6 unprotected after Nf5xg7+. it might be better to prevent Ng3-f5 altogether. I'll take another look at 1...e6 when I get a chance...

    4.Nf3 c5 5.d4 b6

    ...I think 5.c4 and 5.d4 mostly steer into the same lines after 5...b6. Why do you think 5.d4 wastes a tempo?

    Another thing to examine is the same SP with the Qa1 and Bc1 switched. That might also be problematic for Black. - Mark

  24. Oh dear you have a point. After all the dark bishop just loves to be on the long diagonal in this SP, and the queen can go to g5 attacking the weak king-side! Put that on the list. Also we must not forget to check SP941 the classic Etienne Bacrot blunder that we have both blogged on. That SP is not settled yet either.

    5.d4 is only a wasted tempo in a strict sense because the c1 bishop can already develop to b2 or a3. If d4 were an undisputed candidate move like it is in SP518, then the tempo would not be wasted. But in this SP, I am torn to keep the long diagonal open and delay d4 for later.

    This SP is really creative I actually enjoy it now that I am over the negativity of it. You must have a royal headache playing black! The good news is that all your moves are principled and so if it turns out that your opening here is solid, it will be easy to commit the black moves to memory for future Chess960 players if there ever will be any.

    4.Nf3 c5 5.d4 b6 6.c4

    ...Pushing ahead with development inviting capture on d4 - Harry

    (I had to type this entire post twice because my Google browser crashed for the first time).

  25. I got whacked with food poisoning this weekend and lost a couple of days time in all my games. I'll be back as soon as soon as I have a decent move. It's still a tough position for Black to play. - Mark

  26. Finally catching up with my games...

    5.d4 b6 6.c4 Nf4

    ...Analyzing this move for the first game gave me the idea to play it on the 4th move. We know how that worked out! Second time lucky?

    None of the alternatives appeal to me. At least I have counterplay here. - Mark

  27. Glad you're over it Mark not much fun. Was it salami? Are your other games 960 or 518?

    I think Nf4 is still highly principled. After all, black has done about as much developing as can be done right now and Nf4 is a pheno-typical weapon in this SP (did you like the big word? Actually "phenotypes" pretty much sums it up to me and is a good way of understanding Chess960).

    5.d4 b6 6.c4 Nf4 7.Ne3

    ...The knight doesn't block the c1 bishop since that bishop will develop to b2/a3. The problem for white will be how to unblock the e-pawn. It just seems that no matter what, white's position is always nicely organised and black will struggle to unbalance it. - Harry

  28. I think it was chicken as a pizza topping. In the past it's always been some kind of shellfish, which is even worse...

    6.c4 Nf4 7.Ne3 cxd4

    ...I was more concerned about 7.O-O, but it wouldn't have been as awkward as the previous game. The move I played avoids d4-d5 and keeps lines open, still searching for counterplay.

    Phenotypes make for an interesting concept when applied to chess960. Re your question, I'm currently playing 11 other server games, two of which are SP518. I have another eight SP518 games starting in the next few weeks. My usual load is 10-15 games.

    Do you still use your gmail address? I need to contact you privately. - Mark

  29. Yep the gmail address get's me pretty much straight away.

    This is so complicated! I have a request.

    I think I thematically stuffed up with 7.Ne3. I thought that the problem with ...Nf4 was when the f6 square is forcibly loosened and that is what kills black as we found out before. So I withdrew the knight to e3 because in this situation the problem is not your f6 square.

    Now that you have once again brought up 7.O-O!? I've taken another look at it and I think it is once again a winning move but instead for the reason that your king is relatively more stranded than white's, and you cannot coordinate the rooks while I can....

    So my request is that we explore 7.O-O!? once again, because I think that if it works in 100% of cases, we may have to scrub the entire opening you have played (not sure).

    So can we rewind to 7.O-O!? then? If you think it is not necessary that is fine. We are trying to see how black survives this really difficult SP. - Harry

  30. It's OK with me. The previous moves become...

    6.c4 Nf4 7.O-O

    ...I'll now take some time to reset my working files. - Mark

  31. I'm ready to rumble...

    6.c4 Nf4 7.O-O Nxe2+ 8.Kh1 d5

    ...According to my calculations, the move 8...Nxc1 loses by force. The main alternative 8...d5 looks antipositional, but Black seems to hold. - Mark

  32. This was a difficult choice between 9.Ba3 and 9.Rfe1. I also felt that 8...Nxc1 was loosing by force, but it would have been a very chaotic fight for both sides from what I could see.

    7.O-O Nxe2+ 8.Kh1 d5 9.Ba3

    ...Developing the bishop to it's natural square. Note how few opportunities exist to exploit the c1-h6 diagonal for the said piece - Harry

  33. Watching the Candidates tournament is taking time away from my own games, but it will be over in a few days...

    8.Kh1 d5 9.Ba3 Bxf5; if 10.Bxf5 Nxd4

    ...The alternative was first to swap Pawns on c4. Black gives up the Bishop pair to get rid of the dangerous Knight.

    I suggest we continue until Black has resolved the issue of King safety. - Mark

  34. It has been a great tournament and must be giving you a blogging headache having to keep up with it :-) I was shocked yesterday at Aronians disposal of Ivanchuk yesterday.

    I actually cannot believe that white has not a completely winning game here. Everything points to it surely.

    9.Ba3 Bxf5; 10.Bxf5 Nxd4 11.Nxd4

    ...The a3-f8 diagonal will open, the white queen will centralise, white has the bishop pair on an open board and black cannot castle nor connect rooks. What more is needed to win this game! - Harry

  35. All very true and if Black didn't have an extra Pawn as compensation, the situation would be grim...

    10.Bxf5 Nxd4 11.Nxd4 cxd4

    ...If Black manages to consolidate the position, the game will be equal, perhaps even better for Black. - Mark

  36. Response time is slowing as I watch the replays of the candidates tournaments (cannot watch in real time living in the southern hemisphere). Tiredness of all the players is really becoming evident.

    I've been really fascinated at the way Kramnik has morphed his playing style over the years as one of the last from the school of the Botvinnicks. With the help of computer analysis his creativity has been unleashed. In the late 90's he was exploring the solidity of opening systems for endgame advantage (such as the Berlin), then he went on a deep journey during 2000-2010 of the art of queening the pawn, and now he appears to be on a journey of building dynamic balance in the presence of rich tactics.

    10.Bxf5 Nxd4 11.Nxd4 cxd4 12.Qxd4

    ...White's coverage is almost ideal, now just requiring Rfe1 to complete. Surely the positional compensation for white is much greater than the material loss? - Harry

  37. Black's plan is easy: neutralize White's positional advantages and keep the extra Pawn. White's plan is less obvious, because Black has no static weaknesses...

    11.Nxd4 cxd4 12.Qxd4 Bd6

    ...Will White lose a tempo to keep the Bishop pair? - Mark

  38. Good point Mark and thanks for the thoughts.

    I wonder if this exaggerated somewhat bizarre scenario is unique to Chess960? you would be the judge of that better than me. White has a highly coordinated army but no targets?

    You have one strategic weakness as far as I'm concerned. Your king-side is weak to me which could well expose your king in the center especially if I shift my queen over along the 4th rank to the king-side.

    1. White has two bishops on an open board black just one
    2. White will have at least one bishop on a long diagonal
    3. Both whites rooks will centralize and are connected
    4. White's queen is in the center of the board with no threat of attack while black's queen is in the corner and is not even coordinated with a bishop of the same colour.
    5. White's king is relatively safe while black's is stuck
    6. One of black's rook is effectively out of play

    And yet this is close to even score????

    No tell me this is not true?
    How in my mind could I evaluate this position objectively as white, if I had to hold this position in memory in a real life situation? That would take skill indeed!

    11.Nxd4 cxd4 12.Qxd4 Bd6 13.Bb2

  39. Re 'I wonder if this exaggerated somewhat bizarre scenario is unique to Chess960', are you familiar with the Sicilian Najdorf? Black falls behind in development, but has wonderful compensation...

    12.Qxd4 Bd6 13.Bb2 dxc4; if 13.Qc4 Kf8

    ...Black continues to eliminate various weaknesses. The extra pawn looms large. - Mark

  40. Hi Mark got your email. I've touched on the S.Najdorf and I can see that some of the theory applies here but to what extent? This position could well be the most extreme form of Najdorf theory possible! At least in the Najdorf black has the castling option!

    Your last move has thrown me completely!
    ...and so I still am thinking what best to play here.


  41. 13.Bb2 dxc4 14.Qc4 Kf8 15.Qh4

    ...Fakes an attack on the h7 pawn but more importantly, tries to stall black connecting the rooks via g7. White had at least 32 possible moves here! Black's defenses are perfect and there is nothing much white can do about it. An incredible situation - Harry

  42. Batteries recharged...

    14.Qxc4 Kf8 15.Qh4 h5

    ...Prepares ...g6 by preventing Qh6+. To recapture the missing Pawn, White must relinquish the Bishop pair.

    I'm ready to tackle a new SP. What do you think? - Mark

  43. We've proven that black can play this SP producing some really special defensive technique never seen before in the history of Chess so early in the game.

    Looks like the twist in the plot is that if black is careful there is no need to connect the rooks because the g8 rook is actually useful where it is (a potent counter-attacker against white's king).

    There were so many unexplored lines and I should not have suggested the thematic continuation a second time of 7.O-O?!

    Suggest we move onto the same SP except with Qa1 and Bc1 swapped. Thus it would be SP864=BBQRKRNN, another BBQ!

  44. SP864 BBQRKRNN is fine with me. Let me know with a comment on this post when you have it set up. Do you want to play Black for a change? - Mark

  45. Ok Mark, I will put myself onto the skillet for you to fry up - I will play black. At least I know that I should last more than eight moves I hope!

    The template for SP864 BBQRKRNN is up and ready.

    Thanks for this trial! Blacks defense was very instructive and hopefully I have learned something from it.