I grew up and have played most of my chess in the San Francisco Bay Area, where I was fortunate enough to face (and frequently lose to) some of the country’s most talented chess phenoms. Some I grew up with, such as David Pruess, Vinay Bhat, Jordy Mont-Reynaud, and Dmitry Zilberstein. Then I got older and wiser and a little stronger, and a new generation of talented kids arrived on the Bay Area chess scene: Sam Shankland, Sam Sevian, Hans Niemann, and Daniel Naroditsky. Following the time-honored rule that you should always try to beat future top GMs before they start amassing international titles, managed to win games the first three kids (although not Daniel, who I didn’t face until he was over 2300).
I’ve known Sam Shankland the longest and best. We first met as teacher and student when I gave him chess lessons years and years and years ago. Back in those days Sam would stubbornly defend his ideas and then, after I beat him over and over again, would explain that the problem was not that the idea was bad but that I was just better at chess than he was. He probably wasn’t wrong, and pretty soon he was a lot better than I was.
I generally enjoy watching tournaments featuring any of the former Bay Area chess kids because they give me a natural rooting interest. Both of the Sams and Hans were playing in the Sharjah Masters a week or two ago, and Hans and Sam Shankland faced off in Round 7. It was a critical game, as either player could move into a tie for first with a win. The first half was unremarkable. Sam gained relatively easy equality with the black pieces and then both players seemed content to trade everything off. It’s not every day that you get a king and pawn ending as early as move 26!
The diagram below shows the starting point of the king and pawn ending (the last move pair was 26 Rd1xd8 Ke7xd8). Hans Niemann has the white pieces, Sam Shankland is playing black. Despite the fact that Hans traded the last pair of rooks voluntarily, Sam went on to win. I decided it was my job to figure out why.
Pretty much the only chess improvement work I’ve done in the past three years has been in the endgame, so this position seemed like a good place to test myself to see if I’d learned anything. I gave myself ten minutes to figure it out, pretending that I was playing in a tournament (although I was actually staring at a tiny digital board while eating lunch at work). Here’s a rough approximation of my thoughts, as I wrote them down:
Who’s better? I guess black is a little better (he won the game, after all), thanks to his more flexible pawn structure. His king also has better access to both sides of the board for the moment. The only thing white’s got going for him are his advanced queenside pawns, which could be a blessing or a weakness, depending on who outflanks whom.
What should each side be trying to accomplish? White should steer the game towards a draw as quickly as possible. Black should try to keep some winning chances alive.
How complicated is the position? It depends, is it possible to open up the queenside or not? Black could try a6 followed by c5 to try to open things up, but this doesn’t accomplish anything: white can just push past with b5 or respond to a6 with c5 himself. The best that black can do is to force white to play c5 to remove a defender of d5.
OK, so black’s only play is on the kingside with the goal of maybe getting in through d5 someday. Is it really hard to get to a drawn position? Probably not that hard. The basic technique is to lock out the opposing king by creating a fortress across the whole board. I’ve got a couple of examples from my own games as a reference point:
So what’s happening on the kingside? How should white try to lock things up? White can try a defensive set-up with Kf3 and g3 and then sit around and wait. If black plays g5 we’re playing g4 immediately and there’s no way through. Black’s only way to make progress is to play f5, with or without some shenanigans with the h-pawn. The general idea (with the white king on f3) is to take e4xf5 and then immediately play e3-e4 to reestablish control over d5 and f5. If the king isn’t on f3 (or d3, I suppose) when black plays f5, just move it back to f3 and capture next move.
What’s the conclusion, and how confident are you? The game should be drawn; Hans’ decision to trade rooks looks sound. I’m about 98% confident that if I were playing white this position would be my last significant think of the game and I’d be able to rattle off the next bunch of moves and draw relatively quickly. White must have erred later on.
That’s enough schematic thinking for now, let’s bring on the diagrams and variations!
Critical Position #1: The Simple Case
How did we get here? White executed the Kf3-g3 plan and black didn’t really test it. From the diagram at the top, to get here the game could have continued: 27 Kf2 Ke7 28 Kf3 Ke6 29 g3 b6 30 a6 b5 31 c5 f5 32 exf5+ Kxf5 33 e4+ Kf6.
What’s the assessment? The deadest of draws. It’s hard to imagine anything that would even threaten to make progress.
Critical Position #2: Too Close to the Sun
How did we get here? White executed the Kf3-g3 plan and black went a little crazy trying to contest it. The simplest path to this position is 27 27 Kf2 Ke7 28 Kf3 Ke6 29 g3 b6 30 a6 b5 31 c5 h5 32 Kf2 g6 33 Kf3 f5 34 exf5+ gxf5.
What’s the assessment? It turns out that black recapturing on f5 with a pawn is a blunder. Now white is winning thanks to the outside passer: 35 e4 fxe4+ 36 Kxe4 Kf6 37 g4 h4 38 g5+! Kxg5 39 Kxe5 with Kd6 to follow.
Critical Position #3: The Slightly Less Simple Case, but Only Slightly
How did we get here? Black used his h-pawn to try to soften up white’s defensive set-up. For example: 27 Kf2 Ke7 28 Kf3 Kf7 29 g3 b6 30 a6 b5 31 c5 h5 32 Kf2 Kg6 33 Kf3 h4 34 g4 Kg5 35 Kf2 g6 36 Kf3 f5 37 gxf5 gxf5 38 exf5 Kxf5 39 e4+ Kg5.
What’s the assessment? White king shuffles between e3 and f3, black king between g5 and h5. Handshake.
If at this point you’re not convinced that Kf3 and g3 is all it takes to draw, feel free to play around with the position: there’s really not much there. This leads us to part two of our investigation, how did a strong GM lose this position? Which move was the culprit? Here’s the starting position again:
The game continued 27 Kf2 Ke7 28 g4 Kf7 29 Kg3 b6 30 a6 b5 31 c5 Kg6 32 h4 h5 33 Kh3 Kh6 34 Kg3 hxg4 35 Kxg4 g6. Let’s call this …
Critical Position #4: Calling Mark Dvoretsky, We’ve got a Tragicomedy
How did we get here? Look up, this is what actually happened.
What’s the assessment? Losing for white, who played the hopeless 36 h5 rather than get ground down after 36 Kg3 Kh5 37 Kh3 g5 38 hxg5 fxg5 39 Kg3 Kg5 40 Kg2 Kh4, etc. When did things go so wrong? My initial hypothesis was that 28 g4 was the lemon.
Critical Position #5: Tragicomedy? Hold My Beer.
How did we get here? I tried to show that after 28 g4 black is winning, but then the computer laughed at me and then kicked my butt. Here’s the variation: 27 Kf2 Ke7 28 g4 Kf7 29 Kg3 b6 30 a6 b5 31 c5 Kg6 (as in the game) 32 Kh4! h6?! 33 Kg3 h5 34 Kh4! hxg4 35 Kxg4!
What’s the evaluation? I was hoping for 35… Kf7 36 Kf5? g6+, slowly forcing back the white king (and getting to Critical Position #6), but that’s not even close to what happens. Instead the machine gives this outlandish outflanking idea: 35… Kf7 36 Kh5! g6+ 37 Kh6 g5 38 Kh7! f5 39 exf5 e4 40 f6 Kxf6 41 Kg8 g4! 42 hxg4 Kg5 43 Kf7 Kxg4 44 Ke6 Kf3 45 Kd6 Kxe3 46 Kxc6 Kd3 47 Kb7! e3 48 c6 e2 49 c7 e1=Q 50 c8=Q Qxb4 51 Kxa7! with the following position, judged a white win by the all-knowing tablebase:
Don’t ask me exactly how to win this, but I’m guessing that the white king gets some shelter behind the black b-pawn and the a-pawn is too hard to stop.
Critical Position #6: Good Practical Chances but no Forced Win
How did we get here? Having seen the error of my previous winning attempt, I aimed something a little more sound. There are a variety of ways that we might trade g- and h-pawns, including: 27 Kf2 Ke7 28 g4 Kf7 29 Kg3 b6 30 a6 b5 31 c5 Kg6 32 Kh4! Kh6!? 33 Kg3 Kg5 34 h4+ Kg6 35 h5+! Kg5 36 Kh3 g6 37 hxg6 Kxg6 38 Kh4 h6 39 Kh3 Kg5 40 Kg3 h5 41 gxh5 Kxh5 42 Kh3 Kg5 43 Kg3.
What’s the evaluation? White can theoretically hold on, although it will be an uphill battle. This holds true even if it’s white’s move in the diagrammed position; shuffling the king between g2-f2-f3 keeps black from making further progress. Here’s the defensive idea that I initially missed: 43… f5 44 exf5 Kxf5 45 Kf3 e4+ 46 Kg3 Ke5 47 Kg4 Kd5 48 Kf4 Kc4 49 Ke5! (the only move, white has to go after the c-pawn) 49… Kxb4 50 Kd6 Ka3 51 Kxc6 b4 52 Kd7 b3 53 c6 b2 54 c7 b1=Q 55 c8=Q Qd3+ 56 Ke6 Qxe3:
Apparently this queen ending is drawn. Far be it from me to contradict the tablebase, but I would be quite disappointed to have missed the simple draw via the Kf3-g3 plan only to arrive at this position instead.
Too Long, Didn’t Read: What Was the Losing Move?
It’s fair to say that 28 g4?! was an unforced error that could have lost the game even without a second clear blunder, but that blunder came in the form of 32 h4? The reason is clear when we compare critical positions 4, 5, and 6. To get to position 6, the white h-pawn made it to h5. In position 5 (even if white doesn’t play for a win) the h-pawn hangs back on h3. Only in position 4 did the h-pawn get stuck on h4. This meant that there would be an exchange on g5 rather than on g6 or g4, allowing black to take back with the f-pawn, creating the extra, outside passed pawn that makes all the difference.
This game turned out to be critical for Sam’s first place ambitions, as he got next to nothing in his final two games and had to settle for draws. For Hans, it must have been a major disappoint, and games like one this tend to muddy the already cloudy waters of the “Hans Niemann: Cheater or Genius” narrative. I think that he simply got careless and assumed that he could draw any way he wanted; he was certainly playing well in the tournament to this point. For Hans’ best game from the event, check out Dennis Monokroussos’ annotations on his substack.
Norway Chess is already upon us. The first round certainly had its ups and downs. I’ll post about it at some point, but for the moment, who are you rooting for?