King Safety

I've spent a great deal of time playing against Chessmaster: Grandmaster Edition, because for a few years I found it to be a pretty good way to get my chess fix. I was playing at instantchess.com, but quickly grew dismayed by the number of people who would waste 25 minutes because they got themselves into a hopeless position and didn't want to resign.

Other than the various instructional series--and we could debate their usefulness ad nauseum--I have to say that the years playing against a chess engine did one thing, and did one thing better than any human players ever could have:

It gave me intuition for when my King is in danger and when he isn't. It's hard describe this. Because I sometimes end up in positions like:

...and yet, I'm 100% confident in that position. The postgame analysis shows that I played remarkably well for one of my Blitz games--no mistakes or blunders and only 7 inaccuracies. For one of my Blitz games, I promise that's good.

The post-game analysis continually recommends g4, but I only see that as a mistake. It allows Qh4+, putting Black's Queen behind my pawn lines. I just don't see how that's desirable. In fact, that's why I never made the move.

Black went wrong here by launching an attack long before his pieces were developed, and I defended against the attack by developing. We reached this position because I kept bringing new pieces into the fray, but Black's were still tied up.

Black also lost a Bishop because of what did look like a good idea.

I was not at all worried about this. A cursory glance has it looking like a free pawn, but his Queen is the only thing protecting that Bishop, and his Queen is under attack. She has to move.

I knew that Black wasn't going to surrender the Bishop immediately, but before I moved 16.Kf1, I knew it would be met with 16...Qc5 17.Rb5. It is now impossible for Black to protect that Bishop, and Black's attack totally runs out of steam because the Queen is forced to retreat further.

This puts my King on f2, but I feel pretty good about that. My h1 Rook is free, and my other Rook can capture on b7, but I'm not sure about that. I'm not sure I want to give his Rooks access to b2, and Rxb7 would be doing exactly that. Actually, no, it wouldn't, because my Rhb1 eliminates that problem. I didn't see that during the Blitz game, but Black gave me something else to do anyway.

18...Bg4?

For a moment, I wrestled with whether I wanted to capture at b7 or move Rb5. In the end, I went with Rd5 to protect the Bishop at d2. Besides, Rb1 would always allow me to threaten b7, though also will allow b6. It won't matter, because Black's position is really bad anyway. The postgame analysis prefers Rxb7, but I really didn't want to give Black the open file. I didn't see how that was really conducive to my efforts, but I could see how it's certainly helpful for Black. It ties up both of my Rooks on the b-file.

Black soon lost on time, but I didn't set out in this blog to analyze that game anyway.

Let's look at another game.

Once again, I'm rather comfortable with the position. It looks really bad to have that opening on g2 and the doubled pawns on f2 and f3, but... I'm not feeling any pressure about it. In fact, I've done this plenty of times. If Black's light-square Bishop had still been on the board, I wouldn't have done this. And conventional pawn wisdom doesn't always factor into endgames.

More importantly, Black has no way to capitalize on the apparent weakness, and I haven't yet castled Kingside--though, realistically, Queenside isn't a viable option because my pawn structure is even worse over there.

Black grabbed the free pawn with 7...Qxd4, and I continued normally development. I tend to do that when positions get weird--I just develop normally. For the most part. Obviously, the best development for the c1 Bishop is to f4, but I chose e3 instead, to threaten Black's Queen. Obviously, I don't continue on as though nothing has happened, but I do quickly revert back into an opening mindset. I don't know if that's good or bad, but my almost universal response is to continue developing. The first game shows that pretty well, too.

8.Be3, naturally, pushing Black's Queen away. 8...Qd7? I have to question this one. 8...Qe5 looks like it would have been vastly superior, trapping my Bishop on e3 and protecting g5 from his Queen.

9.Bxc4. Hey, just because we've gotten far away from the Queen's Gambit Accepted doesn't mean I can't apply the lessons of QGA. 9...e6. A good opening move, because White has totally control of the board. White's development has gone unhindered through Black's Queen shenanigans and material exchanges. For sure, I have a weakness on g2, but how is Black supposed to seize that advantage and do something with it?

10.Bb5 O-O-O 11.e5 Nd5 12.O-O

Black's position is bad. Black's c6 Knight is pinned to protect his Queen from being taken by a Bishop. His dark-square can go to e7, which is decent and b4 which, which is questionable. White's pieces are active and threatening. White's King is prefectly safe.

12...Nxc3?? 13.bxc3

Black gets rid of his only active piece to further damage my pawn structure. I'm not really concerned about my Queenside pawn skeleton at this point, because it doesn't look like it's going to factor into the game. I don't care how disjointed they are when I start pushing them to break open Black's Queenside castle. As long as they're not tripled, they'll do their job satisfactorily. Meanwhile, Black's out of steam and paralyzed. Black's Bishop only has e7 now. Even the questionable Bb4? is covered by cxb4. Black's Kingside Rook might as well not be in the game, and his Knight isn't going anywhere until he moves his Queen. But the only square his Queen can go to is e7, which blocks his Bishop--e8, of course, does nothing to free the Knight.

And my King is perfectly safe, over there with doubled pawns and a Rook that can't actually protect him because of the f2 pawn.

13...h5?? I tend to think this move is the result of some kind of maxim I'm not aware of, that when you're out of options you start pushing pawns and hope for the best...? That sounds more snide than I mean for it to; what I mean is that I think my opponent has misunderstood the concept of pushing pawns to break open a position. Black should be developing his pieces, not worrying about my King.

14.Rad1. Covering more territory and forcing his Queen away. Although the d-file already belonged to me because of Pawn and Bishop positioning, this keeps a tighter grip. 14...Qe8? not Qe7...?

Oh, well. Time to break open this position.

15.Bxc6 Qxc6 16.Qxa7

Black surely expected 16.Qxc6 bxc6, but I had a different plan exchange pawns. Kinda. Black can't actually capture the f3 pawn, because 16...Qxf3 leads to 17.Qa8#. That Rook isn't on d1 because he likes the view. But he DOES, in fact, like the view...

16...Bd6?? Black panics, and the rest is in the bag. It seems that 16...b5 was Black's best move, though he's still down quite a bit according to the analysis. I think I've seen similar moves in the past, but I didn't see b5 just now without the computer's assistance. In fact, I was getting ready to say that 16...Bd6 was Black's only option.

17.exd6 cxd6 18.Rd2

I knew when moving Qxa7 that Black's Queen would likely recapture on f3, but... again, wasn't worried about it. Black can't actually do anything over there. My Bishop prevents his Kingside Rook from getting involved, and his other rook is tied up keeping my Rook from capturing at d6. Black simply doesn't have enough active material to be a threat.

19.Rd4?!

Although I didn't want to move the Rook to d2 and then to d4, because I don't like shuffling pieces around the board indefinitely, I hadn't anticiapted Qxf3 to happen so quickly.

19...Kd7 prevents 20.Rc4+ Kd7, but that wouldn't have really helped me any. 20.Rf4! and now my King is perfectly safe--he's the safest piece on the board, except that pawn at a2 who everyone is just ignoring. 20...Qc6, Black realizes he has no attack and retreats.

But I have an attack...

21.Rxf7+

I tend to think Black simply didn't consider this as a possibility, but I'd been eyeing that pawn for a while, because it was really weak. A quick glance of the board reveals that f pawn was Black's greatest weakness.

A little half-windmill goes into action, and I mop up two more of Black's pawns at b7 and g7 before he continues pushing that pawn on the h-file and leaves himself vulnerable to Qe7#. Clearly, Black's plan was to get the pawn to h3 and Qg2#, but... Black should have been focusing on his defenses at this point, not his offense.

Besides, after h3, my own f3 nips all of Black's attacks in the bud. And if that wasn't enough, Rxg7 certainly uprooted the attack and turned it on its head, since after Qg2+?? Rxg2 hxg2 Kxg2, and Black is left with:

...and Black is still paralyzed. Neither of his Rooks can leave without losing the other, and Black's King is restricted to only two squares, one of which is easily covered by my Bishop with Bh6.

King safety is a difficult thing to get a feel for, and I'm not always accurate with that intuition--I apologize if I've made it seem that I am. I have good intuition for King safety, not perfect. To that end, I strongly recommend playing against engines repeatedly. Be receptive to what happens and learn from the games; you'll quickly* get a feel for when your King is safe and when your King is not. It's rarely as simple as "I've castled; King is safe." It's not always bad to wreck your own pawn skeleton, especially if you can cripple the opponent's development by doing so.

Chess is a game of subtlety, and it's not a good fit with hard rules that are unyielding.

When I was an active musician, I got into a lot of arguments with people about learning scales, keys, and timings. "But we play rock--it's all about breaking the rules!" they would say ignorantly. And yes, there is some truth to that, but before you can break the rules you must know the rules. Otherwise you're just a baboon smacking pieces around the board, or a baboon smacking an instrument. You'd definitely be breaking the rules to open with a3, h3, a4, h4, Ra3, but that isn't to say you'd be playing a good game. So I don't mean to imply that you should ignore the maxims you've always heard. It's... complicated, and the distinction is subtle. When should you break your pawn skeleton? When should you not? I can't answer that. I'm not right half the time. At least half the time.

* "Quickly" in chess... = 2-3 years