30 May 2015

An Imperfect Understanding

Both game one of On a Losing Streak and game two, Passive vs. Active Play, were played on SchemingMind.com. The next three games were played on LSS. The biggest difference between the two sites is that engines are forbidden on SM, but allowed on LSS. That means the games on LSS are generally tougher and of higher quality.

For this next game I had Black in SP388 QBBRNNKR. The Queen is in the corner and the three diagonal pieces on adjacent files are aimed at the enemy King, While it might be advisable to get the King away from their influence by castling O-O-O, this is probably not going to happen. Opening the diagonals means moving Pawns, which means the Queenside will be too loose for the King.

The first moves of the game were 1.d4 d5 2.b3 c6 3.c4 Nf6 4.cxd5 cxd5 5.Ba3 Bd6 6.Bxd6 Rxd6 7.Ne3 Ne6 8.Nd3 O-O 9.O-O, leading to the position in the top diagram. After both players have castled, a chess960 game often starts to look like a position that could have arisen in traditional chess, and that is the case here. While I wasn't entirely comfortable with my position -- the pieces, especially the Rook, are not well coordinated -- I didn't feel that I was in any particular danger. My next few moves would be spent developing the rest of my pieces and coordinating them into some sort of a plan.

White has an awkward threat in Ne3-f5, so I played 9...g6, planning to reposition the Knight via g7, thereby opening the diagonal for the Bishop on c8. White played 10.f4, a move which I had underestimated. The hole on e4 is not really useful for Black, while White is planning to operate on the f-file with the possibility of switching a Rook to the h-file to support an attack on the King.

The game continued 10...Bd7 11.Ne5 Qd8 12.Qb2 Ng7 13.g4, leading to the bottom diagram.

Now I started to feel really uncomfortable. With every move White is building a Kingside attack, while Black has not yet latched on to a real plan. More pseudo-active moves for Black followed -- 13...Bc6 14.f5 Nd7 15.Qd2 Qb6 -- but after 16.Nf3, how does Black continue? I felt that my position was teetering on disaster and engine analysis confirmed it. White continued to whip up a terrific attack, I played on in an increasingly hopeless position, and finally resigned well after I should have.

Going back to the top diagram, I still can't find a better plan for Black. If that position is bad for Black, then I must have made a mistake in the moves preceding castling. But where? Was it a general problem of not finding an effective plan earlier in the game? If so, how to avoid this in the future? I'm afraid that there's something here that I'm not understanding.


HarryO said...

All I can say is that every move you played in the opening lacked initiative. The problem with the rook on d6 is that you need a knight on d6 not a rook. The other problem with a rook on d6 is that it blocks diagonal attacks on the king via b8-h2. I would say that the rook on d6 is the reason why white has so much kingside attack potential while still having a safe king. So perhaps reconsider the rook lift you made?

Notice too, that your queen was never contributing to anything because of that b7 pawn that was in the way (small point). White moved his b-pawn not only opening the queen up but being able to attack e7 with Ba3. That is pretty powerful!

Easy to say in hindsight! Anyway, it is one of the golden rules of openings not to play out a rook early, but this Chess960 position shows us why that is so in ways that Chess could not.


Note: only my two cents!

Mark Weeks said...

Hi Harry - Thanks for the comments. It's not often that Black can grab the initiative early in a game against a good player. That's a disadvantage of moving second. As for the Rook move, it was a recapture; the alternative spoils Black's Pawn structure.

Not moving the g-Pawn was a conscious choice to defer the decision on how to develop the corner Queen. If this was a mistake, then Black's path to a playable game is very narrow. - Mark

HarryO said...

You could be right about the path being narrow Mark. The warning sign for this SP is that e7 is undefended and white has two ways to attack it via Ba3 (which is benefited by b3 releasing the corner queen) and Nf5 (which is benefited by being supported by the b1 bishop). In hindsight, I think that is what needed to be seen.

In the future, I'll try to remember to experiment with undefended second rank squares from the perspective of two very good attack paths to it.

This SP is unusual because the undefended square is in the middle files (c-d-e-f)! How uncommon is that and do we have a new family of SPs that could be studied because of this feature?

And to people reading this from the future don't laugh at our ignorance at this SP. We are early pioneers of the primitive 21st century of Chess960 history!

Mark Weeks said...

Good observation. I hadn't noticed the undefended e-Pawn. I know that many players look for them as part of their 'pre-flight checklist', but I've always thought that it's better to treat them with the sort of routine tactical checking that's done at every move, the first move included.

I've been thinking about your original comment and am slowly coming to the conclusion that the slow development of the corner Queen had something to do with the later problem of not finding a good plan. I have seen other problems caused by corner Queens and there might be a general principle hidden there. - Mark

HarryO said...

This is good fun, talking Chess960 theory amongst friend! Why are not more people talking about it when it is so much fun to think without accepted conventions.

Ok, thinking about the queen in the corner, if it is generally true that the queen is always part of any chess960 opening plan without exception, then the player who gets the queen involved first, gets to implement their plan first and should always have the initiative (assuming perfect plan formation).

Isn't that a common-sense deduction?

The proviso would be that the player with the initiative would have to be at least two tempi ahead in plan. One tempi is not sufficient.

On another topic, this SP388 we are talking about has an undefended e-pawn. The only way that can happen is if these piece patterns exist behind that pawn:





uRBNu (where u = not N)

uNBRu (where u = not N)

Have I missed any?

That's at least six ways that an e or f-pawn can be undefended! Think about an undefended e-f pawn as a special family because after all they are the two primary center squares crucial to the formation of any plan. Why? Because the centre files e-f are logically involved in the initial phase of establishing control of the centre.

So to any statisticians out there, how many Chess960 position have an undefended e2 (e7) square and then double it to find how many positions have undefended e-or-f squares.

We have a new family of start positions? What do we call this family? The "empty nest" family of chess960 positions?

HarryO said...

I meant d-e as the center files, not e-f files.

Mark Weeks said...

I've done some work on this topic...

Undefended Pawns in Chess960 Start Positions [January 2009]

Naturally Weak Pawns [April 2010]

...so I went back to query my database of SPs. There are 32 SPs with an unprotected d-Pawn (and the same number with an unprotected e-Pawn). As you pointed out, they are not so unusual. - Mark