Rapid Game Analysis: 10 April 2026 (Part 1)

We have a few new instructive rapid games to look at. Let's jump right in with the first one, where I'm white against another master.

Paciorkowski, Lev (2448) - Troutman, Charles (2216)
10-Apr-2026 Rapid, St. Louis, MO (Rd. 1)
G/15+3

1.Nf3 c5 2.g3 Nc6 3.Bg2 (D)

While we are here, I think it's useful to lay out exactly what systems black can choose from. There are a lot of transpositions which can make things a bit confusing, but the way I see it we have 8 different main line systems that would account for the vast majority of games from this point:

• Reversed Grunfeld if black plays 3. ... d5 4.d4 cxd4 5.Nxd4 e5
• The "e6-d5" system: 3. ... d5 4.d4 e6 (Just for my own convenience I sometimes call this the "Xiong system" because it's what GM Jeffrey Xiong once played against me)
• A Tarrasch, which would happen if black plays e6-d5 only after white has played c2-c4, at which point white can play c4xd5, then after ...e6xd5, to play d2-d4
• A variety of independent systems after 3. ... Nf6, where black plays for an early d7-d5 (in response to c2-c4) to try and get a reversed Maroczy bind structure
• The "e6-Nge7 system", where black develops g6-Bg7-e6-Nge7
• Botvinnik System, involving a e5-g6-Bg7-Nge7 setup
• The mainline Symmetrical English
• Symmetrical English with an early ...d6. There are some independent systems here; in fact, this was GM Bryan Smith's choice against me in Round 6 of the 2018 New York State Championship.

In this game, my opponent chose a mainline Symmetrical English:

3. ... g6 4.c4 Bg7 5.Nc3 Nf6 6.d4 (Starting with 6.0-0 would at least give black the extra option of transposing to some type of early d6 system with 6. ... d6; e.g. 7.d4 cxd4 8.Nxd4 Bd7, which is distinct from the mainline Symmetrical English tabiya we get in the game)

6. ... cxd4

Just for completeness, I want to mention how should white react to 6. ... d5!? (D)

It's a very rare move and doesn't have the best reputation for black. After some digging around, it appears there is not much to know. White should respond with 7.dxc5! now attacking the d5 pawn. Then:

• 7. ... dxc4 8.Qxd8+ Nxd8 9.Ne5 (9.0-0!?) leads to a typical Catalan-type position where white is enjoying a tiny plus in piece activity.
• 7. ... Qa5!? should be met with 8.Nd4! (8.cxd5!? Nxd5! 9.Qxd5 Bxc3+ is black's idea) when black has to come up with 8. ... Nxd4! (8. ... Qxc5? 9.Ndb5+/-) 9.Qxd4 Nh5! and there I think it's enough to know that white should play 10.Qd2! (10.Qd3? d4) 10. ... dxc4 11.c6! Rb8 12.0-0 0-0 13.Qg5! essentially forcing a very pleasant endgame after 13. ... Qxg5 14.Bxg5 bxc6 15.Bxe7 Re8 16.Ba3!+=.
• 7. ... d4? is just bad after 8.Nb5!, because 8. ... Ne4? can be strongly met with 9.Nfxd4!+- when the tactics are all working out for white.

Ok, back to the game.

7.Nxd4 0-0 8.0-0 (D)

This is the very well-known tabiya of the mainline Symmetrical English, with thousands of games played here. Since it is such an important position, perhaps it is good to talk a little history. This was a pretty obscure variation entirely until the early-mid 1960s when it started to appear in high-level games, involving such legends as (in no particular order) Averbakh, Polugaevsky, Portisch, Hort, Pirc, Gligoric, Szabo, Larsen, Spassky, Korchnoi, Geller, Reshevsky, Stein, Smyslov and Tal.

What is going on in this position? White is basically saying that we have some pressure on the Nc6 and arguing that makes it difficult for black to finish development, particularly to play ...d6 or ...b6 to get the Bc8 out. Meanwhile, white is thinking long-term about establishing a Maroczy Bind with e2-e4.

Originally, the main line (which I guess is still considered the "main line" in 2026) was 8. ... Nxd4 9.Qxd4 d6. This is the most obvious continuation to relieve the pressure on the Nc6 so that black can start to develop the queenside. A ton of theory developed from that point, with white trying all sorts of moves over the years: Qd3 (what emerged as the most popular), Qd2!? (with the idea b3-Bb2), Qh4, even Bg5!? just to name a few.

It was early on discovered that black could actually just offer a pawn with 8. ... d6, with the idea being that after 9.Nxc6 bxc6 10.Bxc6 Rb8 11.Bg2 Qa5, there appears to be quite a lot of compensation. With the open b/c files plus the a1-h8 diagonal, it isn't clear how white should finish development. As a matter of fact, you can find a game in the database where GM Alex Yermolinsky (a familiar name) played this pawn sac as black in the '80s. Over time, white found good ways to quell black's initiative, and the line is not that popular today, except in faster time controls where white can still quickly get into trouble if you don't know what to do.

Also in the 1980s, the curious-looking 8. ... Ng4!? started to become fashionable, although it never fully surpassed the old main line. The idea is to attack the Nd4 again, and if it retreats, then black could continue ...d6 without losing material. I think eventually the line garnered a suspicious reputation and wasn't fully solving black's problems after 9.e3, so it has fallen out of fashion since then, although you still see it occasionally.

For some reason, the move 8. ... a6?! has always lurked in the background, even being consistently more common than either of 8. ... Ng4!? or 8. ... d6, although almost never used by any strong players. In fact, this move is simply not very good because of 9.c5! which greatly cramps black's position and scores over 85% for white.

A more trendy move in the last decade or so has been 8. ... Qa5. It goes for a typical idea you see in these positions, where black wants to put the queen on h5, then later after ...d7-d6 there will be the option of ...Bh3, ...Ng4 with some kind of pressure on the kingside. This is where a lot of current theoretical discussions are, and strong players like Caruana, Svidler, Wang Hao, Dominguez Perez, Bacrot and Navara have all used this line in serious games.

Last but not least, there is the move my opponent chose in the game, 8. ... Qb6!? (D)

Somehow I never learned anything concrete about this natural move! It is actually the 2nd most common in the database, so better late then never to finally study it.

Of course, black is trying to disturb the knight on d4, but this also seems to be an awkward square for the queen, as she could get hit with Be3 and will not be able to swing over to h5. I didn't know anything about this move, and just quickly played 9.Nb3?! which believe it or not is slightly inaccurate and lets black completely equalize. I found out afterwards that Anand also once played Nb3 in a rapid game, so that's not the worst company to have!

We'll see what happened in my game, but I'll just say before continuing that I think white has much more dangerous options to try which I will know about for next time.

In the game we got 9. ... d6 10.Be3 Qd8 This looks like a very strange way to play for black. After all, can you really waste two full moves like this bringing the queen out and right back again, while white seemingly is close to finishing development? However, the problem is that the Nb3 is really poorly placed (which I understood, but didn't think it was a big deal). Also, the Be3 is constantly vulnerable to ...Ng4 shots. This position is simply equal!

I tried 11.c5 it's as good an attempt as any, and worked out, because it's not super obvious how black should react to this. My opponent played 11. ... Be6?! (Actually 11. ... d5! is the only move that keeps equality, because of 12.Nxd5 Nxd5 13.Qxd5 Qxd5 14.Bxd5 and black can even play 14. ... Bf5!= as there is no way to comfortably defend b2. Not easy to figure out in a rapid game though.) 12.cxd6 exd6+= (D)

So after our mutual imprecisions, we end up reaching an IQP middlegame where I understood white should have a stable (but not too serious) advantage.

13.Nd4 I think this is a good strategic decision, wasting no time in re-improving the poorly placed Nb3. I had to calculate something in case of 13. ... Ng4 (D), which my opponent indeed played.

Without this move I think black would just be really suffering.

14.Nxc6?! I like bishops, but I somehow talked myself out of the stronger 14.Nxe6! on general grounds, not wanting to activate black's rook on f8. I think I even saw during the game that after 14.Nxe6! fxe6 15.Bd2+= Qb6!? 16.e3!+/- there is just nothing impressive for black since I am gaining a tempo on the Ng4, so there is no time to take on b2. (Even without the hanging knight, Rb1-xb7 would still be plenty good for white.) Also not possible would be 15. ... Bd4? 16.e3+-. So black really has no counterplay and white just enjoys two bishops. Not working is 14.Nxe6 Nxe3? 15.Nxd8 Nxd1 16.Nxc6+- when white ends up winning a piece.

14. ... bxc6 The thing is, black now gets an open b-file, so it's not like this doesn't activate one of his rooks too. 15.Bd4 Ne5 (D)

The engine still thinks white is slightly better, but I am not sure how true that is. During the game I wanted to play 16.b3, which initially is also one of the engine's top moves. However, I think I was correct not to like 16. ... c5! 17.Be3 (I don't want 17.Bxe5 Bxe5 18.Bxa8 Qxa8= in my experience these positions are way easier to play for the side with the bishops, even if you can keep the exchange) 17. ... Ng4! again offering the exchange 18.Bd2 d5 and black is the one fighting for an advantage in my opinion.

Maybe 16.Rc1 is a healthier move instead. But I think the real inaccuracy was not taking on e6 earlier. I ended up playing 16.Qa4!? which turned out not to be very impressive. 16. ... Rb8 17.b3 a5! 18.Rad1 Rb4 19.Qa3= This position is just balanced. Black's piece activity compensates for the long-term strategic risk of the hanging pawns. 19.  ... Qc7 20.Ne4 Bd5 21.e3 Re8 22.Nd2 Trying to exchange some of black's active pieces. 22. ... Bxg2 23.Kxg2 Qb7 24.Bxe5 Bxe5 25.Nc4 a4 By this point we are both down to under a minute, so it basically becomes a blitz game. It's all still level. 26.Nxe5 dxe5 27.bxa4 Ra8 28.a5 Qb5 29.Rd6 Rxa5 30.Qc3 Rxa2 31.Qxc6 Qxc6+ 32.Rxc6 (D)

Another rook endgame! You would be forgiven for thinking that nothing interesting could possibly happen from this position. However, we both were down to about 15 seconds (still with the +3 second increment), and the funny thing is I think both of us were still playing for a win. Black's rooks are more active, but there isn't a way to really crack white's position, unless I blunder...

32. ... Rbb2 33.Rc3 Kind of with the idea of e3-e4, although I am not sure I really want to play that now because of ...Re2. 33. ... e4!? Either way, my opponent believed me. Now I actually get a target to attack. 34.g4 Trying to discourage black from uniting his pawns with f7-f5. Now we shuffle around a lot. 34. ... Ra4 35.Kg3 Kg7 36.Rfc1 Raa2 37.Rf1 Ra4 38.Rc5 Rd2 39.h3 h6 40.Rfc1 Raa2 41.Rf1 Ra4 42.h4 Rb2 43.Rfc1 Raa2 44.Rf1 Ra6 Black now tries to reposition his rook. 45.Rc4 Re6 46.Rd1!? Rf6! (D)

I have to say I missed that one. We are basically just playing an over-the-board bullet game at this point, with only the 3 second increment each move. In the few seconds I had I didn't notice that black actually has two threats to win a pawn!

47.Rf1?! (47.Rxe4 Rfxf2 48.Rf4 is arguably easier for white to defend) 47. ... Rf3+ 48.Kg2 Rxe3 (D)

Black has stolen a pawn. Here I made a split-second decision to force a rook trade, which I thought would make my defensive task easier. That was a good idea on my part, but I didn't execute it in the best way.

49.Rb4!? I didn't consider the other option of 49.Rb1! which is a much cleaner solution, getting rid of my passive rook. Black even would have a chance to lose there after 49. ... Rbe2?? 50.Kf1!+-.

49. ... Rxb4 50.fxe3 This is still a draw, but now it gets tougher, especially with no time to think. 50. ... Rb3 51.Kf2 f5 52.gxf5 gxf5 53.Rg1+ Kf6 (D)

The best technique here would be to right away get the rook active to hit one of black's weaknesses, the pawn on h6. This can be done with 54.Ra1!, intending 54. ... Kg6 55.Ra6+ Kh5 56.Rf6 Kg4 57.Rg6+!= (only move). I had the right idea, but unfortunately the way I chose to carry it out blunders my e3 pawn.

54.h5? I simply missed or forgot about the possibility now of 54. ... Rb2+! 55.Kg3 Re2 56.Kf4 Rf2+ 57.Kg3 Rf3+ which is winning for black.

54. ... Ke5? (D) Returning the favor.

55.Rg8? The comedy of errors continues. This is a completely ridiculous move, explainable only by the fact that I have 10 seconds on my clock. 55.Rg6 should draw easily, always meeting Rb2+ with Kf1/e1.

55. ... Rb2+ 56.Kg1 I realized now that otherwise black will simply play ...Rh2, picking up the h5 pawn. 56. ... Re2 57.Rg3 (D) Now it is black's turn to make a mistake with under 10 seconds!

Black can win just by marching the king to d3. After 57. ... Kd5! there is literally nothing white can do. 58.Kf1 Rh2-+, and otherwise the best I can do is trade h6 for e3, when the resulting position is winning for black with two connected passers vs. my one passer.

57. ... f4? 58.exf4+ Kxf4 59.Rg6 Now it should be a draw again. 59. ... Re1+ 60.Kf2 Rh1 61.Rxh6 Rh2+ 62.Kg1 Rh4 63.Kg2 Ke3 (D)

White has numerous ways to make a draw. By far the easiest is just to run up with the king to support the h-pawn, preparing to sac the rook for black's e-pawn at the last moment. Somehow I was still entrenched in the pure "defense" mindset and the thought of playing actively with my king didn't occur to me.

64.Kg3 Rh1 65.Kg2!? (65.Kg4 Ke2 66.Kg5 e3 67.Re6 is much simpler) 65. ... Rh4 66.Kg3 Rh1 67.Kg2 Ra1 68.Re6!? Another strange decision. Much easier is 68.Rh8! Kd2 69.h6 when now black even has to be careful. 68. ... Kd3 69.h6 Ra7 70.Kf2 Rf7+ 71.Ke1 (71.Kg3! e3 72.Kg4 e2 73.Kg5) 71. ... Ke3 72.Ra6 Kf3 73.Ra3+ e3 74.Rb3 Rf6 (D)

If you calmly look at the position for a few seconds, it will quickly become clear that white will still draw this trivially after a move like 75.Rb8! Rxh6 76.Rf8+ (the Philidor technique). This is of course something I am familiar with, but somehow my brain didn't recognize that pattern quickly enough in this specific case with the h6-pawn. Both of us had been about to flag for the last 50 or so moves, and here I make the last comical blunder: 75.h7?? Rh6-+ (D)

Whoops! Now it is just game over. 76.Kd1 Rxh7 77.Rb8 e2+ 78.Kd2 Rd7+ 0-1

An very well-played game by my opponent. Except for his one opening imprecision with 11. ... Be6?! and our back-and-forth mistakes in the time scramble, I don't think he made a single inaccuracy giving me a chance for an advantage. In the future, I will have a better idea of how to meet 8. ... Qb6.

I think it is also interesting for me to note how my split-second intuition never led me to consider ditching passive defense with my king on moves 65 and 71, and to instead run up the support my own h-pawn. That would have been far easier than what I did in the game.

Game Analysis: 2026 Mid-America Open, Round 1

Continuing our series of looking at my games from the 2026 Mid-America Open, we have a quick one today. In the first round of the event I managed to win a miniature against an expert.

Previously we've covered rounds 3, 4 and 5 from this event.

Paciorkowski, Lev (2446) - Boaz, Ben (2051)
2026 Mid-America Open, St. Louis, MO (Rd. 1)
G/60 d10

To fit all the games in one weekend, the tournament used a faster time control in the first two rounds of one hour with a ten-second delay.

1.Nf3 e6 2.g3 f5 3.Bg2 Nf6 4.0-0 d5!? (D)

A Stonewall Dutch! Actually, one advantage of starting with Nf3-g3-Bg2 is that in particular against the Stonewall, white does not need to play d2-d4, but instead has a stronger plan of playing for a quick e2-e4 break which is what I tried to go for in the game.

Although I did win quickly with 5.d3!? here, I will be nitpicky and point out an even more precise way for white to play that begins with 5.c4! c6 and only now 6.d3 (D)

This position scores over 70% for white according to my database. The point of including c2-c4 is that black's central formation will come under even more pressure after later including e2-e4. I will not delve into all the details here, only to say that after the most commonly played 6. ... Bd6!? (which I think is already slightly imprecise) 7.Nc3 0-0 8.e4+/- black just doesn't have a good way of keeping the tension in the position (e.g., 8. ... fxe4 9.dxe4 dxe4 10.Ng5+/- already scores in the 80-90% range for white). In light of this, I honestly think the Stonewall just is not a viable setup against Nf3-g3-Bg2, if white knows about this most critical line (which I definitely will know for next time).

In the game, after my 5.d3!? we continued 5. ... Bd6 6.Nbd2 (6.c4!) 6. ... 0-0 7.Re1 (I had come to the realization that 7.e4!? fxe4 8.dxe4 dxe4 9.Ng5 e3! 10.fxe3+= really isn't that impressive for white, which is one reason why the knight is much better on c3 than on d2, as we could then take with the bishop on e3 to avoid ruining our pawn structure.) 7. ... e5 (7. ... f4!? deserved attention) 8.e4 fxe4 (don't blink; this game is already halfway over...) 9.dxe4 d4 (D)

This position we got out of the opening is unusual and perfectly playable for black, in my opinion.

10.Nc4 (I actually did not even consider 10.Nxd4!? dxe4 11.e5 here, although I later thought about this idea.) 10. ... Qe7!? I think this is the start of black's problems. 10. ... Nc6 is a much healthier developing move. I thought for about 13 minutes before playing 11.c3 (D)

My intention was 11. ... dxc3 12.Qb3 Kh8 13.bxc3 which I thought would be comfortable for white. That perhaps may be true, but objectively black would be doing fine there after 13. ... Nc6 or 13. ... Nbd7.

However, my opponent played a much more ambitious move that, perhaps surprisingly, turns out to be the decisive mistake.

11. ... c5? Although it would be nice to keep the cramping pawn on d4, unfortunately black's position is just not well developed enough to support such an extension. As invariably happens in the Dutch, the open diagonal to the king on g8 will turn out to play an important role.

12.b4! (D) I played this quickly. During the game I only thought white would be much better, but black's position is in fact just collapsing.

12. ... Nc6?! not the most resilient, but the position is very difficult. During the game I had seen after 12. ... b6 I have at the very least 13.Nxd6 Qxd6 14.bxc5 bxc5 15.cxd4 cxd4 (15. ... exd4 16.e5+-) 16.Qb3+ Be6 17.Ba3 Qa6 18.Qb2 Rc8 19.Nxe5+- when white wins a pawn, gets the bishop pair, and will likely pick up d4 later as well. The engine confirms that is indeed winning for white.

I also had considered 12. ... Bc7 for black, which is what the engine wants to play. There I saw 13.bxc5 Qxc5 (otherwise black just lost a pawn for nothing) 14.Bf1!?+- threatening Ba3 and still triple attacking the d4 pawn. White wins material there.

After 12. ... Nc6 the game ended quickly: 13.Nxd6 Qxd6 14.bxc5 Qxc5 15.Qb3+ Kh8 16.Ba3 1-0 Black resigned as he is losing a full rook.

Actually, my opponent went on to have a really good tournament after this game, ultimately finishing with 3.5/5 and a performance over 2300, gaining 40+ rating points.

Next time though, I will know to play 5.c4! c6 6.d3 against the Stonewall.

Game Analysis: 2026 Mid-America Open, Round 4

We have quite a long game to look at today, so let's jump right in. This is a continuation of a series where I have been analyzing in detail some of the games I played in the recent Mid-America Open. Previously we have looked at rounds 5 and 3.

My opponent this round is FM Gabriel Petesch. I was black this game.

Petesch, Gabriel (2454) - Paciorkowski, Lev (2446)
2026 Mid-America Open, St. Louis, MO (Rd. 4)
40/80 SD30; d30

In my preparation, I was able to find some recent games Gabriel had played online and saw that with white he tends to use a non-specific Nf3-e3-b3 system, which in fact I play sometimes myself. So I did some opening work and figured out exactly how I would want to play as black against this system. As it turned out, my preparation was spot on and I correctly predicted what he would play.

1.d4 This was a slight surprise, as I had seen he more typically starts with 1.Nf3, but we quickly transposed back into what I expected anyhow. 1. ... Nf6 2.Nf3 g6 3.b3 (D)

3. ... c5 (D) This is the system I had prepared for black, immediately challenging white's position.

4.e3!? In my humble opinion, I think this is slightly imprecise for white and plays exactly into black's hands. I believe white should prefer 4.dxc5 Qa5+ 5.Nbd2 Qxc5 6.Bb2= which is about equal.

Our game continued 4. ... cxd4 5.exd4 Bg7 6.Bb2 0-0 (First playing 6. ... d5!? 7.Bb5+ would at least give white other options) 7.Bd3 d5 8.0-0 Nc6 9.a3!? (D)

I am still in my preparation here, this being one of the possible lines I expected. White's last move was directed against ...Nb4, which would annoy the bishop on d3. At first glance, it does not seem like there would be any danger for white in this structure, but during my pre-game exploration of this middlegame, I learned about a strong maneuver for black that poses white some serious practical problems. In fact, because of this maneuver, I believe black is already objectively slightly better.

9. ... Nh5!=- (D)

There are 55 games in the database here, and of those black has won 36 and lost 9 (the rest draws) for an impressive score of 74.5%. This statistic alone indicates the extent of white's difficulties here.

What is the purpose of this strange-looking knight sortie? The obvious positional threat is ...Nf4, which will win the bishop pair. However, it also opens up the Bg7 to attack d4. This means that black also has a much more serious threat to play ...Bg4 next move, which will make it quite difficult to defend the d4 pawn. A move like c2-c3 would be extremely undesirable for white.

To illustrate how careful white already has to be, the most commonly played move here, 10.Re1?, is already a serious mistake since black just plays 10. ... Bg4! (10. ... Nf4? 11.Bf1 is not as effective for black) 11.Be2 and now 11. ... Nf4-/+ makes white's position very unpleasant [0-1 Yewdokimov (2375) - Dorfman (2575), 1991].

My opponent thought for a little bit and played 10.g3?! (D)

This was one of the moves I had looked at during my pre-game preparation. I confidently played my response, only finding out after the game that I had actually mis-remembered what I recommended for black in this exact position. (By the way, according to the engine, white's best move to limit the damage is 10.Qd2, but still after 10. ... Qc7 11.Re1 Nf4 12.Bf1 Bf5=- black's position is very comfortable and white's queen is awkwardly occupying the Nb1's natural square.)

10. ... Bg4!? Black is still slightly better, but this is a little imprecise. Even stronger, and what I had actually analyzed before the game, is 10. ... Qb6!-/+ (D)

Simply attacking the d4 pawn. The obvious move for white is then 11.Be2 (11.c3!? f5!-/+ quickly generates a strong attack for black after ...f5-f4), after which now black continues 11. ... Nf6! (The knight has done its job on h5 and should be immediately redeployed.) 12.Nbd2 Ne4!-/+ white has no comfortable way to keep d4 defended, and black can continue building the pressure with ...Bh3 and ...Rd8.

11.Be2 Nf6 12.Ne5!? White really should prefer 12.Nbd2=-, continuing development 12. ... Bh3-/+ 13.Nxc6 bxc6 14.Re1 Ne4 15.Nc3 (D)

The opening has clearly been a success for black, and there are multiple good-looking options here to continue. What I chose here was tempting, but in hindsight a bit too impatient.

15. ... c5?! I somewhat overestimated my chances in the position this led to. In actual fact, it would have been stronger to first play 15. ... f5!-/+ (which I didn't consider), allowing black to recapture on e4 with the f-pawn and open up the Rf8. The break c6-c5 is not going anywhere and can always be played later. In the game, we got 16.Nxe4 dxe4 17.c3 White doesn't have any other reasonable options. During the game I saw 17.Bf1? Bxf1 18.Kxf1 cxd4 19.Rxe4 Qd5-+ and rightfully concluded that despite the equal material, this should be strategically winning for black, with ...e5-f5 to follow. 17. ... cxd4 18.cxd4= (D)

I thought this kind of position would be very attractive for black, and indeed white still has to tread carefully here. But the advantage is only minimal (actually, at higher depth, my engine even assesses this as just equality. Definitely could have fooled me!). With careful play, white can keep the d4 pawn adequately protected and maintain the balance. There followed 18. ... Qb6 (threatening to continue with ...Rd8, winning the d4 pawn) 19.Bc4! The only good move, opening up a counter-attack on e4. 19. ... Rac8 (D)

With the idea 20.Rxe4? Rxc4 21.bxc4 Qxb2-+. I was now expecting white to play 20.Qd2, defending b2 and reinstating the threat on e4. In that position, I was considering to sacrifice a pawn with 20. ... Rfd8!? (I also thought I could just play 20. ... Qb7=) 21.Rxe4 Qb7 22.f3 (forced) 22. ... Bf5 23.Re3 (D)

I was trying to evaluate this position while my opponent was thinking. Even with an extra pawn, I didn't think white should be better since my pieces are so active and the Kg1 is always a little weak (this intuitive assessment is correct). However, I didn't see a clear way to break through for black (indeed, there is none). For example, I saw that the attempt 23. ... e5? (23. ... h5=) 24.d5 Bh6 25.f4!+- is suddenly good for white, since I cannot continue 25. ... exf4 26.Qc3!+-.

All of this turned out to be irrelevant though because my opponent instead played his first significant mistake with 20.Qe2? (D)

The problem here is that white is not actually threatening to take on e4 because then black would again have ...Rxc4 followed by ...Qxb2. So I can actually just pick up the d4 pawn right away with 20. ... Rfd8-/+ The game then continued 21.Rac1!? (somewhat more resilient would have been 21.Rad1, when during the game I had the right idea by trying to get my bishop to g4, but hadn't quite figured out the proper execution. The solution is to play 21. ... h5!, followed by ...Bg4, which picks up either the exchange or the d4 pawn) 21. ... Bxd4 22.Bxd4 Rxd4 23.Qe3 After the game my opponent confessed that his intention was 23.Bxf7+? but he forgot that my bishop on h3 still defends the rook on c8! A classic one that I have definitely overlooked before too...bishops move backwards. 23. ... Rcd8 I remember spending a lot of time on this perfectly natural move. 24.Be2 (D)

White threatens g3-g4 to trap my bishop. This is one of the more important critical moments of the game. I spent a significant amount of my remaining time here, as I could not figure out a clear way to make progress. Besides g3-g4, I also saw that white has some real intentions to get counterplay with b3-b4, Rc5, Rec1, then even b4-b5 and Rc6. The Qe3 and Be2 is such a sturdy construction, defending all of white's weak squares on the kingside.

Intuitively, I felt like black should have a near-decisive advantage here. After all, white's king should be weak, and we still have the extra pawn, right? This assessment is correct, but black has to proceed accurately here, otherwise a large portion of the advantage will disappear.

I did at least partially hit upon the correct strategy: we need to exchange off white's Be2. It is such a strong defensive piece, not only guarding entry squares for my rooks like d3 and d1, but also physically shielding the 2nd rank, making Rd2 ineffective. So I thought about 24. ... Qe6?! (to play ...Bg4), but rightfully did not like 25.Rc7, when white gets counterplay of his own.

Unfortunately, I did not really consider the alternative execution of 24. ... h5!-+. This prepares ...Bg4 without readily allowing counterplay on the 7th rank. In fact, black can even throw in a few housekeeping moves like ...Kg7 and ...Qf6 before proceeding with ...Bg4. White is helpless, e.g. 25.b4 Kg7 26.Rc5 Qf6 27.Rec1 Bg4!-+ after the Be2 leaves, black will be able to invade with the rooks.

Why did I miss this? I think I did not quite sense the urgency in this position. I did have the correct intuitive assessment that this position should be at least nearly winning, but as we have discussed before, winning position does not equal easy position! This would be considered as a hybrid of a Type 2 and Type 3 winning position. We have a small material advantage plus a large positional advantage (white's weak king).

In this case, where part of our advantage includes our opponent's weak king, the correct approach is to look for an energetic way to "go for the throat", i.e., play for the attack. We need to do this quickly to make the most of our position before white is able to organize sufficient counterplay of his own along the c-file. We can combine this thinking with asking a useful question, "what is our opponent's most important defensive piece?" (I feel like I did at least subconsciously notice that it is the Be2 during the game) and figure out that we need to play ...h5 and ...Bg4, then look to invade decisively with the rooks. This broader strategy, and particularly the urgency of carrying it out, is what was escaping me in this position.

Watch what happened instead: 24. ... Be6?! 25.b4 Kg7?! (25. ... h5! is still strong) 26.Rc5-/+ already the trend is going in white's direction. 26. ... R4d5 27.Rec1 and here, starting to get into time pressure, I made the practical decision to return the pawn to simplify the position. 27. ... Bf5!? (27. ... Rxc5 28.Rxc5 Bd5=- was at least objectively keeping a small advantage) 28.g4 (now I thought white's king might be a little more exposed) 28. ... Be6 29.Qxe4 Qd6= (D)

Quite the turnaround. At this point both of us have about 4 minutes left to make it to move 40, so we both just started to make some cautious moves. 30.Qe3 Rd4 31.h3 Qf4!? (It was OK to keep the queens on with something like 31. ... h6= but with little time I wanted to lower the stakes) 32.Qxf4 Rxf4 33.R1c2 Rfd4 34.f3!? (D)

The position is about equal, but I thought now I could start to ask white some questions. Instead, 34.Rc7 was perhaps a better practical decision, keeping me honest with my a7 pawn.

34. ... Bb3! This disrupts white's piece coordination. 35.R2c3 Rd2! 36.Re5 (36.Rxb3? Rxe2-+ cannot be good for white) 36. ... Be6 37.Rce3!? (Again, 37.Rc7 probably was to be preferred) 37. ... Ra2 Now black is able to double rooks on the 2nd rank. The position is still equal, but it starts to get tricky for white. 38.f4 Rdd2 39.f5 gxf5 40.gxf5 Bc8!? (D)

Finally we make it move 40, and without any accidents. I did not like 40. ... Bxf5 41.Rxf5 Rxe2 42.Rxe2 Rxe2 43.Ra5 when that endgame gets unclear with both of us having connected passers. If white now plays 41.Rxe7 then I wanted 41. ... Kf6=, and white is in kind of a weird bind. That position would be equal.

My opponent instead played 41.Kf2 to which I responded 41. ... Kf6 (D)

Now if white tries 42.Rxe7, I had prepared 42. ... Ba6 when 43.Kf3 (white has nothing better) 43. ... Bxe2+ 44.Rxe2 Rxe2 45.Rxe2 Rxa3+ 46.Kg4 Ra4 will lead to a rook endgame with an extra pawn for black. That should still be a draw, but white is the one who would have to prove it.

Perhaps wanting to avoid that, white instead played 42.Kg3? but it turns out that now after 42. ... Bxf5 (D)

White can no longer take on e7, because 43.Rxe7? Be6!-+ suddenly wins a piece for black. This is now a very difficult position for white, as I have basically just stolen the pawn on f5 for free.

43.Bf3? (D)

It was not obvious, but the only move to stay in the game for white was 43.Bg4! Bxg4 44.hxg4 e6 45.Ra5 when black has nothing better than 45. ... Rd7-/+ and still has good winning chances, but the game is not over.

Now something strange starts to happen. I felt like something was fishy about this position, and that white's king is kind of in a mating net. So I started to look for anything that seemed to be a forced win, thinking that I might have to find a precise move here or my chance might evaporate.

Turns out my intuition was kind of correct. Black actually is completely winning here (engine gives -4!), but there is not exactly any kind of "forced win". There are even multiple moves that maintain a decisive advantage. This is just a winning position because of the extra pawn which white will not have time to regain, plus the long-term vulnerability of the king on g3.

The best move is the very simple 43. ... e6!, then if white tries 44.Ra5 (the only active move) black has multiple strong continuations. The easiest is 44. ... Rh2, which is similar to what happened in the game.

43. ... Ra1 is also strong, when white's king is just not going to survive long-term after 44.Bg4 Rg1+ 45.Kf3 Rf1+ 46.Kg3 e6. I considered that, but during the game it wasn't necessarily obvious to me that the king is fatally weak there, and white still has some potential counterplay with Ra5.

The move I played is 43. ... Rh2 (D). It's technically not as strong as the other options, but it's still good enough to get the job done. However, the problem is not the move itself, but how much time I spent on it! I used 19 of my remaining 25 minutes in this position, which really was a poor choice on my part. We will see why later.

The threat to take on h3 will be absolutely crushing, so white's next move is apparently forced. 44.Bg4 One reason why my move was not as precise is white actually can (and perhaps should) try 44.Rxf5+!? Kxf5 45.Rxe7 Kf6 46.Rxa7. I kind of assumed that should just be winning for black, but actually one further precise move 46. ... Rhd2!-+ is required there to stop Bd5 (after the game, Gabriel told me that was why he rejected that line, which is fair enough).

44. ... Rag2+ 45.Kf3 (I had correctly seen 45.Kf4 Rf2+ 46.Bf3 (46.Rf3 Bxg4-+) 46. ... e6 47.Kg3 (47.Ra5 Rxh3 48.Rxa7 e5+ 49.Rxe5 Rfxf3#) 47. ... Ra2 48.Bg4 when I at least get the same exact position with the additional free move ...e6 for myself). 45. ... Rf2+ 46.Kg3 Rhg2+ 47.Kh4 e6 48.Ra5 Rf4! (D)

This is the position I had seen back when I was calculating on move 43, and I had further noticed that 49.Rxa7? h5! (there are other moves too) would win immediately, with the intention 50.Kxh5 Bxg4+ 51.hxg4 Rh2+ with mate next move. At this point, my opponent found some very strong defensive resources with little time on his clock.

49.Rf3! This came as a surprise to me. I had seen 49.Rg3 Rh2! when 50.Rxa7 h5! still wins: 51.Kxh5 Bxg4+ with mate to follow. 49. ... Rd4! black is still in control, but now my opponent blitzed out another very strong move: 50.Ra6! (D)

This is such a weird position. White's bishop on g4 is pinned and I am attacking it three times, but my own bishop on f5 is also pinned! To add to the confusion, my e6 pawn is now pinned to my king, so white is actually simply threatening Rxf5+ here! There is nothing for black after 50. ... Rdxg4+? 51.hxg4 Rxg4+ 52.Kh3.

50. ... Kg7! Still in control, because h7-h5 is a renewed threat. 51.Rxa7? h5 52.Kxh5 Rgxg4! (not 52. ... Rdxg4? 53.Rxf5!=) 53.hxg4 (now 53.Rxf5 Rh4+ 54.Kg5 exf5-+ wins for black) 53. ... Bxg4+ 54.Kh4 Bxf3+ with a trivially winning position for black. At minimum we can give the bishop for one of white's queenside pawns, and then two connected passers always beat one pawn in a rook endgame.

51.Rg3! (D)

White keeps finding the most resilient defensive moves. Black is still winning here with pretty much any move that avoids trading rooks (there are no discovered checks because white's bishop is still pinned). But it is still not completely trivial. The strongest move is 51. ... Rgd2!, when after 52.Rxa7 Bg6! white is utterly defenseless against the threat of h7-h5, e.g. 53.Kg5 h6+! 54.Kh4 h5-+.

Unfortunately, I only had 1 minute in this position and didn't see that. Instead I started to get worried about still not finding any forced win and white having two connected passed pawns after taking on a7, and decided to bail out.

51. ... Rxg3? 52.Kxg3 Bxg4 53.hxg4 Rd7 (D)

I knew that playing passively like this is not likely to win me the game (in fact, black is not even better anymore; it's just equal), but at least the position is simplified and I knew I shouldn't lose either. The game ended in a draw after a few more moves. 54.Kf4 Kg6 55.b5 h5 56.gxh5+ Kxh5 57.a4 Kg6 58.Ke5 Kg5 Black just cannot make any progress because my rook is tied to the defense of a7. 59.Rd6 Re7 60.Rd8 Kg4 61.Kd6 Rb7 62.Kc6 Re7 63.Kd6 Rb7 64.Kc6 Re7 1/2-1/2

It was painful not to win that game, but I think it also revealed some weaknesses I still have to work on. To recap what happened:

• My opening prep was spot on, and by move 9 I was already a little better with black, even though I did slightly mis-remember some of the details of my analysis.
• It's a minor note, but my 15. ... c5?! was a little impatient and objectively threw away a large chunk of my advantage, although the position remained hard for white to play. I essentially just over-estimated the objective evaluation of the IQP position we later got.
• After I had won a pawn, around move 23 I correctly felt that I probably should have a near-decisive advantage, and at least intuitively understood that white's bishop was my opponent's strongest defensive piece and should be exchanged. But I didn't quite appreciate the urgency with which I had to carry out that plan. I was correct to reject 24. ... Qe6?! 25.Rc7, but should have been more persistent and asked myself how else I can try to trade the bishops without allowing as much counterplay. This would have led me to appreciate the strength of 24. ... h5!-+.
• After giving the pawn back to "bail out" into an equal endgame, I felt like I was still able to pose practical problems for my opponent from moves 34-42 to eventually reach an objectively winning position again. But my time management was not great, spending 19 of my 25 minutes on move 43. It did not take me that long to see the position where I can effectively force white's king to h4, and a more pragmatic decision would have just been to play that as soon as I saw it, having the foresight to save my time for subsequent critical moments. I should have just trusted my intuition that white's king is not supposed to be able to survive there, even if I didn't see a forced win yet.
• My opponent was extremely resourceful and found all of the most resilient defensive tries from moves 49-51, with little time on his clock. In the final critical moment on move 51, I could not make the best decision because my poor time management had left myself with only 1 minute on the clock. So I bailed out again and the game ended in a draw.

I think one moral of the story is that I should trust my intuition more! For the most part, my intuitive evaluations of several key positions throughout this game were correct. I would criticize my evaluation of the IQP position after 15. ... c5?! but other than that I was spot on.

However, because I didn't trust my intuition, I was very indecisive in a couple key moments, burning a lot of my time senselessly. In particular, I spent 19 minutes on 23. ... Rcd8, which was a completely natural (and the best) move that I could have and should have made in maybe 30-60 seconds, or at most a few minutes. Then I burned another 19 minutes later on 43. ... Rh2. That time adds up and if I even had left myself with 10 minutes on move 51, I am sure I would have been able to keep my advantage and probably would have won the game. This indecisiveness led me to not fully trust my position assessments while I was calculating on move 43, and I ended up wasting a lot of time re-checking variations. Perhaps we can also talk about more organized calculation.

Definitely some food for thought, but that's all for now. Later we may look at two other games from this event.

Rapid Game Analysis: 03 April 2026 (Part 3)

This is the final rapid game from a small 3-round event. We have so far looked at rounds 1 and 2 where I was up against two expert-strength players. This game, I am playing none other than GM Alex Yermolinsky, a two-time US champion who used to be in the top 25 in the world back in the 1990s. Although that was a while ago, he is still obviously a very strong and experienced player, and I was curious to see how he would approach the opening against me. I'm playing the black pieces this game.

Yermolinsky, Alex (2442) - Paciorkowski, Lev (2448)
03-Apr-2026 Rapid, St. Louis, MO (Rd. 3)
G/15+3

1.Nf3 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.d4 0-0 6.Be3 (D)

I did not shy away from a fight this game, entering the King's Indian which is one of the sharpest ways black can play against d4/c4/Nf3 systems. In response, Alex has chosen quite a rare move for white which I suspect was to try and quickly get me out of my own opening knowledge (mission successful). The way I see it, there are the following main systems white can choose from in the position after 5. ... 0-0:

• Classical Main Line with 6.Be2 e5 7.0-0 - this is where most KID theory is
• Petrosian System: 6.Be2 e5 7.d5 - this is kind of rarely seen nowadays, and I believe not considered to be too dangerous for black
• Gligoric-Taimanov System: 6.Be2 e5 7.Be3 - this has a little more poison to it and certainly could be considered a more serious weapon for white
• Exchange Variation: 6.Be2 e5 7.dxe5 - this is very rare nowadays, but it used to be a penchant of GMs like Ulf Andersson, who would try to use their technical prowess to win the resulting drawish endgames
• A Makogonov-type system with 6.h3. This is also a serious move. In the traditional move order, white plays h2-h3 before necessarily committing the knight to f3.

After a small amount of thought, I played 6. ... Ng4 I was a little surprised to find out after the game that this is a fairly uncommon move for black, even though it seems like the natural way to test white's early placement of the dark squared bishop without including h2-h3. The main alternative would be 6. ... e5 when white would have the option of transposing back into a Gligoric system with 7.Be2 (which perhaps black would prefer not to allow) or we could go into more uncharted territory with a line like 7.d5!? Ng4 8.Bg5 f6 9.Bh4.

The game continued 7.Bg5 This is the only move that could ask any serious questions to black; otherwise I would just play e7-e5 right away. 7. ... h6 8.Bc1!? (D)

This is pretty unusual though. In my database I could only find one high-level blitz game where this was played. Over the board, the only move I was still a little concerned about was 8.Bh4, keeping the pin on the e7 pawn. It turns out this move has been used with some success by white, mostly in faster time controls. Black will have to decide if it is worth it to commit to g6-g5 so that we can achieve our desired ...e5 break in the center, a decision which is not obvious if you've never looked at this before.

8. ... e5 There is nothing wrong with this natural move, and I think black has equalized. 9.dxe5 This was also white's choice in a 2025 blitz game Paravyan-Lazavik. The try 9.h3!? exd4! (as was my intention during the game) 10.Nxd4 Ne5 is nothing special for white, while 9.d5!? f5 only enters a traditional KID structure where it seems like black has gained several useful tempi.

9. ... Nxe5 My choice here deviates from any established previous games. I did consider 9. ... dxe5, but felt concerned about my knight on g4 being out for lunch in those positions, and I have to constantly think about if white can take on e5 after ...Nf6. The aforementioned blitz game went 9. ... dxe5 10.Qxd8 Rxd8 11.Nd5 eventually ending in black's favor [0-1 (42) Paravyan (2582) - Lazavik (2605), Blitz 2025]. For what it's worth, the engine does actually slightly prefer my move, so perhaps there was something to my intuition after all.

10.Nxe5 Otherwise black will quickly be able to organize some pressure on the c4 pawn with ...Be6. 10. ... dxe5 11.Qxd8 Rxd8 12.Nd5 Na6 (D)

This kind of endgame is always strategically a little risky for white, since the d4 square is a proper outpost for black, while the d5 square can still be covered by my c-pawn after ...c6. In this particular position, black also has a small lead in development. White could try here to liquidate the position with 13.c5!? when I was planning 13. ... Nxc5 14.Nxc7 Rb8=, but actually, because black has such a lead in development, the engine even suggests the insane-looking 13. ... Rxd5!(!?) 14.exd5 Nb4! 15.Kd1 Nxd5 when black apparently can fight for an advantage in that position despite being an exchange down. I don't think I ever would have seriously considered that during the game but it's a neat concept to see!

13.Be3 Be6 14.0-0-0 c6 15.Nc3 b6!? Not strictly necessary, but I wasn't convinced by 15. ... Bf8?! 16.Rxd8 Rxd8 17.Bxa7+= when there is not quite enough compensation for black. Nothing much happens for a while 16.Be2 Bf8 17.Rxd8 Rxd8 18.Rd1 Rxd1+ 19.Kxd1 Bc5 It is almost always desirable for black to exchange the dark squared bishops in this structure, leaving white with a more passive light squared bishop. 20.Bxc5 Nxc5 21.Kd2 Kf8 22.Ke3 Ke7 23.h4 g5 24.hxg5 hxg5 25.b3 (25.b4!? would always have been risky because of 25. ... Nb7! when white's c-pawn can become a target later after ...Nd6) 25. ... f6 26.Bd1 Bf7 27.Bg4 Kd6 28.g3 a5 29.Be2 Ne6 30.Na4 Kc7 31.Nc3 Nd4 32.Bd1 b5 33.cxb5 cxb5 34.Ne2 Nxe2 35.Bxe2 b4 (D)

The start of an interesting bishop ending. During the game I was correct in thinking that this should objectively be a draw, but white is the one who has to be careful, since all of my pawns are on the opposite color square from my bishop. In general, my immediate plan is to use the pressure from my Bf7 to achieve ...a5-a4-a3, then look for some way to make progress after that.

36.f4 gxf4+ 37.gxf4 Kb6 38.fxe5 fxe5 39.Bd1 Kc5 40.Bc2 Be8 41.Bd1 Bb5 Here I am just moving around to gain some time from the increment. At this point I had about 30 seconds and Alex had about a minute and a half on the clock. 42.Bc2 Bf1 43.Bd1 Bg2 44.Bc2 Bh3 45.Bd1 Be6 46.Kd3 a4 (D)

The first real decision for white to make. Should this pawn be taken, or is it OK to allow black to achieve ...a4-a3?

It turns out that both options still draw, although it would start to get very tricky for white after 47.bxa4!? Bxa2. So understandably, Alex allows my pawn to go forward, but at the same time that also feels like a big accomplishment for black. 47.Ke3 a3 48.Kd3 (of course white can never allow 48.Be2? Bxb3!-+) 48. ... Bc8 49.Ke3 Ba6 50.Bc2 Bf1 51.Bd3 Bh3 (51. ... Bxd3 52.Kxd3 is a dead draw) 52.Bc2 Bg4 53.Kd3 Bf3 54.Ke3 Bh5 55.Kd3 Bf7 56.Ke3 Kd6 (D)

After some more shuffling to gain time on the clock, here I start to march my king over to the kingside to try and invade somehow. My bishop on f7 always keeps c4 under control, so white will never be able to invade with his own king on the queenside. White's bishop also always has to stay on c2 or d1 to prevent ...Bxb3.

57.Bd1 Ke7 58.Bc2 Kf6 (D)

Again white has a decision to make. Should anything be done now to stop my king from invading too far? Had I been playing white, I think my intuition would have been (of course, in a classical game these things should be decided by pure calculation, not intuition. This doesn't apply in fast time controls though.) to try 59.Kf3 Kg5 60.Kg3, but that is not so easy because black can try to set up a zugzwang position - e.g. 60. ... Bh5 61.Bd3! (61.Bb1? Bd1-+ is zugzwang) 61. ... Bg4 62.Bc2! Be2 (D)

Here white is only surviving because of 63.Kf2! momentarily attacking my bishop. Any other move loses, e.g. 63.Bb1? Bd1!-+ Here in the post-game analysis perhaps that doesn't seem so hard, but over the board with little time that isn't obvious.

In the game Alex decided on 59.Bd1 which also draws, but at least my king can invade more now. 59. ... Kg5 60.Bc2 Kg4 61.Bd1+ Kg3 62.Bc2 Kg2 63.Ke2! (white cannot allow my king to f1) 63. ... Bh5+ (D)

Quick - where should white go with the king? Only one move holds the draw.

64.Kd3! (D) Threatening to invade through c4 now that my bishop has momentarily left that diagonal.

I actually can win white's bishop in the variation (I didn't have time to calculate this during the game) 64. ... Kf2 65.Kc4 Ke3 66.Kxb4 Kd2 67.Bb1 Kc1 68.Kxa3! (68.Bd3? Kb2!-+) 68. ... Kxb1 but white is just in time to run back with his king to eliminate my e-pawn: 69.Kb4! Bf3 70.Kc5 Bxe4 71.Kd6=

So I go back with my bishop. 64. ... Bf7 65.Ke2! Be8 and try to dislodge white's king from another diagonal, where I keep c4 defended at the same time. 66.Kd3 (just to show how careful white has to be, 66.Bd3? loses after 66. ... Bh5+! when white's king can no longer go to the needed d3 square. The last finesse is 67.Ke3 Bf7! threatening to take on b3 68.Bc2 Kf1-+) 66. ... Bb5+ 67.Ke3 (D)

White is still holding on by the slimmest of margins. 67. ... Kf1 68.Bd3+! would lead to a drawn K+P endgame: 68. ... Bxd3 69.Kxd3 Kf2 70.Kc4 Ke3 71.Kxb4 Kxe4 72.Kxa3 Kd3 73.b4 e4 74.b5 e3 75.b6 e2 76.b7 e1=Q 77.b8=Q and this is a draw. By the way, white also would have to be sure that 67. ... Kf1 68.Bd3+ Ke1? 69.Bxb5 Kd1 doesn't work for black - in fact, white's extra bishop is winning there, but only with 70.Kd3 Kc1 71.Kc4 Kb2 72.Kxb4 Kxa2 73.Bd3! Kb2 74.Bb1!!+- and white wins. Without this resource, or if the bishop couldn't get back there in time, black would actually be winning.

67. ... Ba6 (D)

68.Bd1? After a lengthy defense, the difficulty of chess finally makes itself felt and my esteemed opponent goes astray. The only move was 68.Bb1! to still keep the Bd3+ resource available in case of ...Kf1.

68. ... Kf1-+ White cannot effectively deal with this king invasion, although my opponent kept kicking and fighting, still finding an ingenious defensive idea in the next few moves. 69.Bg4 Ke1 70.Be2!? (D)

I remember being quite startled by this because it just comes out of nowhere. 70. ... Bxe2? is stalemate! Fortunately, I have better options.

70. ... Bc8 White now cannot deal with both ...Be6 and ...Kd1. 71.Bc4 Kd1 72.Kd3 Kc1 73.Bd5 Ba6+ 74.Ke3 Kb2 75.Kd2 Bc8 (75. ... Kxa2 76.Kc2 Be2! is also winning with the idea Bd1+!) 76.Kd3 Ba6+ 77.Kd2 Bf1 78.Kd1 Bd3 0-1 Now my bishop goes to b1 and cleans up white's queenside.

It is really incredible to see how complex a bishop endgame can be with such a limited number of pawns left on the board. This game also goes to show how difficult defending such an endgame can be, even if you know or feel it should be a draw.