In all of the positions where the Tata Steel Masters resigned, I tried my hand against Stockfish


See my [previous](https://www.reddit.com/r/chess/comments/fn4ej7/in_positions_where_the_candidates_rds_15_resigned/) [posts](https://www.reddit.com/r/chess/comments/odshcu/in_positions_where_the_candidates_rds_614/) doing this for the Candidates.

I may or may not do this for the actual championship match; the problem is that I’ve already looked at those games a lot, and Game 8 and esp. Game 6 were resigned when they were way over and there isn’t much interesting going on anymore, so there’s a limited number of interesting games I could do this for.

I’m a little conflicted on whether to spoiler-tag all (or much more) of my commentary, to make it easier for people to do this themselves; if enough people ask me to I think I will.

[**Grandelius – Donchenko**](https://lichess.org/TQdmxQ7A) (Round 1)

Actual result: 1-0

My result: 1-0

It was a simple pawn ending that the engine made even simpler for me by just giving its e-pawn away on move 1.

[**Giri – Tari**](https://lichess.org/G4eaUEc9) (Round 1)

Actual result: 1-0

My result: 1-0

Another easy one. If you can win material and checkmate at the same time, that’s pretty nice.

[**Carlsen – Firouzja**](https://lichess.org/TsJcuvNS) (Round 1)

Actual result: 1-0

My result: 1/2-1/2

This is embarrassing because I actually thought Qg6 was mate in 1 for some reason. It’s actually a fairly tricky mate in 3 puzzle, for me at least. It took me a while in the analysis board to figure it out.

[**Donchenko – Caruana**](https://lichess.org/i4OfyaeQ) (Round 2)

Actual result: 0-1

My result: 0-1

The starting position was a confusing one; lots of stuff was hanging and there were lots of ideas on the board.

I forced myself to think about it for a while. I ended up coming with a pretty fool-proof winning idea: take the light-square bishop (preventing any shenanigans where white’s f7 pawn gets protected and white’s rook makes scary back-rank threats), and then bring my rook to b2 and b1 to force a queen (I didn’t realize Ra1 was possible but that wasn’t going to work in the long term).

The engine ended up saying that every reasonable move was winning for me, so it turns out none of these positions were too critical.

[**Duda – Grandelius**](https://lichess.org/3FR1rifM) (Round 2)

Actual result: 0-1

My result: 0-1

I’m very very proud of how I played, and won, this game. I’ll be spoiler-ing some text because this is one I think I’d recommend for other people to try too (particularly starting at move 4).

There’s a caveat that I set 20+20 time controls for myself in these games, and that might be more time than the actual players had. I used almost all my time for this game.

It was definitely not easy. I went into a rook+bishop+3 vs. rook+3 endgame, and I thought most likely I wouldn’t be able to pull it off (or that maybe the engine actually had an objective draw at one point or another). I had to dig deep and look for precise ideas to actually convert this, including the idea that finally ended it: >!35… h4, knowing that Rxh4 leads to Rg2+! leading to a forced rook trade and a win!<.

The engine agrees and said I had no innacuracies/mistakes/blunders, and only 7 average centipawn loss. The one asterisk is that the very first move the engine made was apparently a mistake, so I had a slightly easier starting position than Grandelius would have had.

[**Grandelius – Harikrishna**](https://lichess.org/VltX8wRP/black) (Round 3)

Actual result: 0-1

My result: 1-0

I ended up losing a tricky pawn ending where I forgot for a move too long about the split-pawns idea (where you hold on with two pawns against a king, because you exploit the fact that the king can’t take the back pawn because the forward pawn runs).

I gave myself a [consolation game](https://lichess.org/oOw8cJI2) from the position where I effed up, to allow myself to demonstrate the draw and to try to make myself feel better.

But in a plot twist, it turned out I resigned in a drawn position! Even though white was going to win my pawns, there was another draw I forgot about, namely that the g6 and h5 pawns don’t win against the h6 pawn when black’s king is in the corner, due to stalemate ideas.

The only actual blunder I made (other than resigning in the drawn position) was 52… Nb5; with the idea that if white takes on c2 I’m forcing a king and pawn ending. My logic was that it was my only practical chance of winning the game, but the engine says that d4 was the only winning move in that position.

[**Firouzja – Guijarro**](https://lichess.org/6nvug6XF) (Round 3)

Actual result: 1-0

My result: 1/2-1/2

I wasn’t as valiant with this one. Black was already so active that I didn’t have as much of a plan for how to try to win as I did with some other games (also I didn’t spend as much time).

The engine says Rd6 is when I blew the win and I should have played Ne6+ instead. I was thinking of that move but I didn’t think it did anything; I missed that when the knight takes on d5 it also hits a black pawn on c3, which would win it and force rook+knight+1 vs rook.

[**Grandelius – Vachier-Lagrave**](https://lichess.org/T1mpo3O6) (Round 5)

Actual result: 1-0

My result: 1-0

In the starting position I couldn’t find anything that completely puts the game away. I settled on just going with the obvious move of Qxh7.

Ke8 then posed me with a decision between taking a third pawn with check, or playing a back-rank check that would force an eventual queen trade. In the end, I took the third pawn and forced the queen trade anyway.

Any endgame will be a little stressful against an engine with a knight, but this one was pretty straight forward. Black had no real counterplay and I had way too many pawns for black to deal with.

[**van Foreest – Guijarro**](https://lichess.org/EBZcM5Cq) (Round 6)

Actual result: 1-0

My result: 1-0

This was a laugher. Not only was I up a piece, but Stockfish played an apparent blunder that let me win its queen.

[**Donchenko – Firouzja**](https://lichess.org/e2HL4g9U) (Round 6)

Actual result: 0-1

My result: 0-1

Another easy one. I had to find Rd1+ for this to be easy, but I did and the rest of the moves sort of played themselves. It was also easier to find Rd1 in a context where I knew white resigned in this position.

[**Caruana – Vachier-Lagrave**](https://lichess.org/UTbjLQ4A) (Round 7)

Actual result: 1-0

My result: 1-0

Another easy-to-medium one. I did have to figure out how to stop my king from getting checked, but I managed to figure it out relatively easily.

[**Giri – Grandelius**](https://lichess.org/Py4jCTsV) (Round 7)

Actual result: 1-0

My result: 1-0

There was a bit of a blooper here. I initially premoved g8=Q, but then realized in time it would be stalemate on Ka1.

[**Firouzja – Duda**](https://lichess.org/nL8gTAmH) (Round 7)

Actual result: 1-0

My result: 1-0

Seeing this position where I only had a rook, bishop and pawn had me thinking “oh brother”. Fortunately, I did realize after a little bit that black’s king was in a mating net and had to give up its rook to prevent checkmate.

Even then, though, the endgame wasn’t so trivial because black had several pawns that could get dangerous. My main way of avoiding shenanigans was to make sure black’s king could never get to the third rank to start harassing my g-pawn.

[**Tari – van Foreest**](https://lichess.org/xR5EfCwK) (Round 7)

Actual result: 0-1

My result: 0-1

Another laugher.

[**Harikrishna – Firouzja**](https://lichess.org/cluugo0I) (Round 8)

Actual result: 0-1

My result: 0-1

Required a *little* bit of thought but it was mostly just boring and drawn-out.

[**Esipenko – Carlsen**](https://lichess.org/uOcMB8JE) (Round 8)

Actual result: 1-0

My result: 1-0

The only challenge was preventing perpetuals. Black had one last trick in Qb6, threatening Qxb3 with a perpetual (so not allowing me to get a second queen immediately). I found Bd5, which the engine said was the most precise.

[**Giri-Vachier – Lagrave**](https://lichess.org/ERzlFzsC) (Round 9)

Actual result: 1-0

My result: 1-0

This is a somewhat interesting endgame to study- although white has a little too many winning options to say this is a proper endgame study or something. I’ll spoiler-tag the rest in case someone wants to look at this themselves.

>!The engine didn’t play Kf5 after Re4, after which I calculated e6! would win (if black takes the rook, the e-pawn is queening with check and in time to stop black’s pawns). Turns out the engine recommends Kd5 as even better than e6.!<

>!The engine also didn’t play Kf5 after Rxh4, which I had to calculate after the game. I settled on Rh1 as the solution there, but that turns out to be a draw in a position where white has a million winning moves. The flashiest way to win would again, be Kd5!, because after g2, white can’t stop the queening directly, but Rh8! prepares to meet g1 with Rf8+, forcing black’s king to the g-file where Rg8+ gets the skewer.!<

[**Wojtaszek – Caruana**](https://lichess.org/O8LOiI8C) (Round 9)

Actual result: 0-1

My result: 1/2-1/2

Playing this endgame against the two bishops was an experience. It’s absolutely crazy how many moves I wanted to make that were either directly covered or covered with tactics.

The engine says my biggest mistake (from -2.1 to -0.6) was 54…Ke7. I should have just swallowed my pride with Kg6 and allowed the repetition, after which I had a win with either b5 or Rc3. I also missed 43…Rd1, where Kg2 is met with the nice follow-up f3+!, forcing the king to either stop defending the f1 bishop or to allow a fork winning the b2 bishop.

Still, I don’t really feel bad for not converting this game overall.

[**Carlsen – Grandelius**](https://lichess.org/ZHuw2kVAjQ5V) (Round 9)

Actual result: 1-0

My result: 1-0

This is another one I’m somewhat proud of.

It was a headache, and at times I thought I wasn’t going to be able to win. But instead of just accepting a repetition I was able to win a second black pawn, and eventually black allowed 91. Qd5+, trading queens and ending the game. A human wouldn’t have done that and would have made me keep going with the queens on the board, but oh well.

[**Vachier-Lagrave – Donchenko**](https://lichess.org/w24MAwtC) (Round 10)

Actual result: 1-0

My result: 1-0

Laugher. I saw queens on the board and was bracing myself for a repeat of the last game, but the queen trade was already forced in the initial position.

[**Esipenko – Guijarro**](https://lichess.org/xMKNJDNq) (Round 10)

Actual result: 1-0

My result: 1/2-1/2

I wasn’t really sure what made this such a resignable position at first; it took a while to find Qxg6, a tactic I often miss (where you’re exploiting the pin the bishop has on the f7 pawn).

Still, I screwed up the win with a bunch of mistakes in a row.

The engine says I made two blunders: 66. Ke2 in lieu of Kg4 (I got too scared of black’s passer I guess), and 67. f4, which just blundered Ke7, turning the game into a dead draw.

[**Giri – Wojtaszek**](https://lichess.org/k12BJXDe) (Round 10)

Actual result: 1-0

My result: 1/2-1/2

Two straight games that I didn’t win, with this one culminating in a *drawn* rook+2 vs rook position.

The only mistake I made per the engine was the f5 move that allowed said draw. It’s not immediately obvious to me how to actually win knowing you can’t allow that, and the engine isn’t very helpful when it just says every move wins eventually.

But it looks like a key feature of the drawing position after f5 is that black is allowed to put its king on g6, forcing white’s rook to perpetually guard the g-pawn laterally, preventing any progress. But in, say, [this position](https://imgur.com/Px2S4dT), black’s king is stuck on the back rank so Kg6 is impossible. Black also can’t just take the pawn and wait with its rook on the 4th rank, because g6! threatens mate and forces black to passively defend, which is obviously a losing strategy against two pawns.

[**Tari – Esipenko**](https://lichess.org/11Dh8Gft) (Round 11)

Actual result: 1-0

My result: 1-0

The piece-up endgame with minor pieces is pretty easy, especially when I have a passed pawn too. The lack of heavy pieces reduces complications, and trades are always good because the 1 piece vs. 0 with just the kingside pawns on the board is an easy win.

[**van Foreest – Harikrishna**](https://lichess.org/B1LiBYxI) (Round 11)

Actual result: 1-0

My result: 1-0

I got some slight flashbacks to the rook+2 vs rook draw from before, and black did get one stalemate trick in, but I was able to figure this one out fairly easily.

[**Harikrishna – Donchenko**](https://lichess.org/zHT208Ql) (Round 12)

Actual result: 1-0

My result: 1-0

It took an embarrassing amount of time for me to realize black’s rook was hanging.

[**Carlsen – Vachier-Lagrave**](https://lichess.org/mXmkZaN5) (Round 13)

Actual result: 1-0

My result: 1/2-1/2

I saw Bd2+, but then I saw I had an escape square and thought that was it. I missed the perpetual afterwards, and I think I should have suspected something was up when the engine just straight-up let me promote like that.

The [consolation game](https://lichess.org/CqBGWE7x) where I gave myself e8 back, though, was pretty easy.

[**van Foreest – Grandelius**](https://lichess.org/nYaznp2I) (Round 13)

Actual result: 1-0

My result: 1-0

Again, it look embarrassingly long for me to realize black’s queen was just hanging, and that giving the queen away was just the best/only way to avoid mate.

It seems like a queen being able to take the other queen for free is a hard thing for me to spot for some reason. Maybe I’m just too focused on the tension between the queens, so I don’t even realize that one of them is just hanging.

[**Donchenko – Esipenko**](https://lichess.org/T08pZuUc) (Round 13)

Actual result: 0-1

My result: 0-1

This was a little tricky, because it took me a little while to realize Qxf2+ was the best move, and there were also some stalemate traps in the knight ending. But I still managed somewhat easily.

[**Giri – van Foreest**](https://lichess.org/wuJXk1a8) (Armageddon)

Actual result: 0-1

My result: 0-1

I was going on a bit of a goose chase in the starting position before I realized I could just queen, and if white takes my rook I take white’s rook.

# Final Results

Out of 29 games, I got 22 wins, 6 draws, and 1 loss, for a score of 25/29 (86%). You might not think it’s that good because these are all winning positions of course, but it’s markedly better than the 14.5/25 (58%) I got for the Candidates installment(s).

I’m guessing the improvement is due to a combination of some easier positions (I found myself with a lot of “laughers” this time; maybe the Candidates were more prone to resigning early because the tournament itself was more important, or something), plus the fact that I’m pretty sure I bore down and paid more attention to the games this time.



View Reddit by bluecjjView Source

Recommended For You

About the Author: chess99

Leave a Reply