Lessons From Bad Bertrich IX

MY round-nine game, in which I was White against international master Dieter Pirrot (2321), featured a classic example of how not to prepare for a game.

My opponent has 831 games in ChessBase's 2026 Mega database, of which 407 have him with the black pieces.

I fairly quickly decided I wanted to play 1.e4, which reduced the field to a still-healthy 156 games, of which 96 saw him continue 1...e5.

A Spanish would suit me, I felt, and it seemed he had 'only' two replies 3...a6 and 3...f5.

Against the former I would be happy with my normal Exchange Variation, but I had to do some work before deciding how I would meet the latter.

But my preparation could not end there, as in Mega26 he also has 21 games with 1...e6, 20 games with 1...d5, 10 games with 1...c5, and also games with 1...Nf6, 1...c6 and 1...Nc6.

I was not overly inclined to look at everything, especially as with some of those moves there was very little to go on, but on examining a list of my opponent's games by date, I noticed the French was very much his modern fancy, and I even found a game where he answered 1.d4 with 1...e6.

I have played many games with the Exchange Variation of the French, trying to emulate Morphy, but I decided this would be a too-insipid choice against an IM, and I ruled out the King's Indian Attack, a favourite of mine some years ago, on the ground of rustiness.

That left 2.d4 d5 3.Nc3, of which there were six games of DP's in Mega26 - four continued 3...Bb4, and two saw 3...Nf6.

I fairly quickly decided what I would play against the Winawer, and then looked at the two games with 3...Nf6.

I used to continue exclusively with 4.e5, but more recently I have mixed in 4.Bg5, and with this latter continuation the one game in Mega26 showed DP replying with the fashionable 4...dxe4 5.Nxe4 Be7.

Position after 5...Be7

*****

*****

*****

*****

*****

That game continued 6.Bd3?! Nxe4 7.Bxe7 Nxf2!? 8.Bxd8 Nxd1 9.Bxc7 Nxb2, which the engines reckon is level, although DP went on to win.

I knew the main move is 6.Bxf6, when 6...gxf6!? gives Black the sort of imbalanced game - pawn-structure v bishop-pair - that the superior player should be happy to get.

Accordingly I settled on 6.Nxf6+ Bxf6 7.Bxf6 Qxf6 8.Nf3, when I noticed the engines suggested meeting the main move, 8...0-0, with the little-played 9.Qd3, the idea being to follow up with queenside castling, the quicker the better.

And that was the extent of my preparation - it never occurred to me to look a few moves further ahead, to see how Black might react to White's queen move.

Over the board there came 9...c5 10.0-0-0 Rd8, and suddenly I realised White has difficulties.

Position after 10...Rd8

*****

*****

*****

*****

*****

Black has three men attacking d4, and is about to bring in a fourth attacker with ...Nc6.

White has three defenders, but adding a fourth defender with 11.c3? sees White's king exposed after 11...cxd4 12.cxd4, when the simple 12...Bd7, with ...Nc6 and Rac8 to come, gives Black a winning attack, according to Stockfish18 and Dragon1.

White therefore has to move the queen to unpin the d pawn, but my 11.Qe3?! showed my inexperience with the position (Qe4 is probably best, as I detail in my notes to the game).

After my move, Black played 11...Nc6, when 12.dxc5? Rxd1+ 13.Kxd1 Qxb2 is terrible.

Instead I played 12.d5!? (I had planned this when playing Qe3), but I was on the back-foot for almost the whole of the rest of the game - it felt as if I was playing with the black pieces, and my opponent with the white ones.

LESSON: half-baked preparation can be more dangerous than no preparation.

Lessons From Bad Bertrich VIII

MY round-eight game, in which I was White against Horst Prüsse (2006), began as an O'Kelly Variation in the Sicilian: 1.e4 c5 2.Nf3 a6.

I did not know my opponent played 2...a6, and I had nothing prepared

*****

*****

*****

*****

*****

*****

All I could recall was that 3.d4!? was considered a mistake, and that White should play a line in which ...a6 would not prove useful, eg going into an Alapin setup with 3.c3.

I cannot recall where this advice came from, but in the end I rather went against it by playing 3.c4, intending to play a Maróczy Bind, even though ...a6 in the Maróczy is a common move as, among other things, it helps prepare the break ...b5.

Imagine my surprise when I discovered, after the game, 3.d4!? is easily the commonest continuation in ChessBase's 2026 Mega database (I also found that I had played 3.d4!? last year, albeit with the special intention of switching into a Morra Gambit).

So, was the advice in my mind simply wrong, or had I misremembered altogether? Maybe not, on both counts.

For starters, the top two recommendations of Stockfish18 and Dragon1 are 3.c4 and 3.c3, although the former engine comes to like 3.g3!? the more time it is given to consider the position.

And then there are the stats to consider. 3.c4 scores 56% in Mega26, 3.c3 scores 54% and even 3.g3!? scores 52%, whereas 3.d4!? turns in a miserable 40%.

The main line in Mega26, after 3.d4!?, runs 3...cxd4 4.Nxd4 Nf6 5.Nc3, and now Black gets to play what is effectively an Accelerated Najdorf, ie 5...e5!?

Position after 5...e5!?

*****

*****

*****

*****

*****

The engines agree Black has already at least equalised, eg after the main move in Mega26, 6.Nf3, they reckon 6...Bb4!? starts to tip the balance in Black's favour.

So it seems the half-remembered advice I had at the back of my mind was sound after all. But it would have been nice if I had originally spent some time learning the reasons behind the advice, rather than just noting it.

LESSON: as usual in chess, knowing something is good, but knowing why it is good is much better.

Lessons From Bad Bertrich VII

THE progress of players is in some ways mirrored by their relationship with the chessboard's back rank, and in particular with the danger of falling victim to a back-rank mate.

Novices often succumb to such a mate, until in time they learn it can be worthwhile playing h3/...h6, or similar, to give the castled king an escape square.

But as players get stronger, they realise playing h3/...h6 - a "bookend move," as it has been called - can be a serious waste of time, and even make the king's position less safe.

Eventually players develop a feel for when a back-rank mate is on, or, more precisely, for when it is useful to take precautions against the possibility of such a mate.

My round-seven game at the Bad Bertrich 50+ tournament, in which I was Black against Michael Hillenbrand (2113), featured an interesting back-rank scenario, arising after I played 22...Rfd8.

All four black pieces (five if you include the king) are either on the back rank or are defending it - even the bishop covers f8

*****

*****

*****

*****

*****

In the game, White saved his queen by playing 23.Qf3?, but after  23...Rxd1 24.Qxd1 Qe4+ 25.Kg1 h6 White was lost, thanks to the pin on the c file.

The move 25...h6 was necessary because, although the peril is not immediately obvious, Black's back-rank is weak.

So much so, that in the diagram position White has three moves that maintain at least equality, Qe5, Ba1 and Be1, and it is worth looking at each in turn.

*****

23.Qe5 threatens mate on g7, so Black's best is 23...Qxe5, and after 24.Bxe5 Black cannot capture either white rook, eg 24...Rxc2?? loses to 25.Rxd8+ Bf8 26.Bd6.

Black should probably play 24...f6, when Dragon1 gives White a slight edge, but Stockfish18 calls the position equal.

*****

23.Ba1!? gives an even stranger impression at first sight, but 23...Rxd5?? loses to 24.Rxc8 Qf8 25.Rxf8+ and 26.Rxd5.

Not so bad is 23...Rxc2?!, although 24.Qxd8+ Qf8 25.Rd7!? is promising for White, eg 25...Rxa2 26.Be5 Rc2 (preventing Qc7) 27.Qg5 h6! (only move) 28.Qxf5 Rc5 29.Qe4, after which White's greater activity gives the upper hand, according to the engines.

The engines reckon Black's best is 23...Bc5!?, after which White should play 24.Qe5 or 24.Qf3, both with equality.

*****

23.Be1?! also looks odd, but the engines agree it leaves the game completely equal.

However, unlike Qe5 and Ba1, it sets Black no problems, and several moves maintain the balance, including 23...Bc5, 23...g6 and 23...Rb8.

LESSON: the threat of back-rank mate creates possibilities that otherwise would not exist, and this is especially so when most minor pieces (bishops and knights) have been exchanged, but heavy pieces (queens and rooks) remain.

Lessons From Bad Bertrich VI

MY round-six game, in which I was White against Martin Schopf (1909), was decided in a rook-and-pawn ending.

A critical position arose after I played 48.Rg3-g5

*****

*****

*****

*****

*****

*****

*****

My opponent replied 48...Rxg5?, which, as I commented in my post, "makes White's task easier."

The evaluations of my main engines jump from around +3.00 for White to more than +5.00 (Stockfish18) or almost +9.00 (Dragon1).

But it occurred to me afterwards that it would be instructive to see how White wins after a more sensible move, such as 48...Ke7, assuming the engines are right, and there is a way for White to break through.

Position after 48...Ke7

*****

*****

*****

*****

*****

Material is level, but White's advantage is based on the fact that Black's 4-3 queenside majority is blockaded, while White's 4-3 kingside majority still has some flexibility.

The moves that follow are the engines' consensus on how to best play the position. Where they disagree, and it is not obvious which is correct, I will choose what seems to me the more-human continuation.

49.Rf1!? Rg7 50,g4 hxg4 51.Rxg4 Rh5 52.Kf3 Rh6 53.Rc1 Kd7 54.Rh1 Rh5 55.Rg5 Ke8!? 56.Kg4 Rh8 57.h5 Rgg8 58.h6 Kf7.

White has created a passed pawn, but how does White make progress?

*****

*****

*****

*****

*****

59.Kf3 Rh7 60.Rg2 Rgh8 61.Rc2!? Rxh6 (61...Rc8?! 62.Kg4) 62.Rxh6 Rxh6 63.Rxc6 Rh3+ 64.Kg4 Rb3 65.Rxa6 Rxb4 66.Ra7+ Ke8 67.a6 Kd8 68.Kg5 Rxd4

Black is a pawn up, but White's more-active king and advanced passer add up to an advantage worth well more than a rook, according to the engines

*****

*****

*****

*****

*****

69.Rh7 Ra4 70.a7 Kc8 71.Kxg6 Ra6 72.f5!? exf5+ 73.Kxf5 b4 74.Rh8+ Kc7!? (74...Kb7 75.e6) 75.a8=Q Rxa8 76.Rxa8, and even the most-stubborn Black can resign.

LESSON: engines can make endgame technique seem simple, but watching such technique is instructive.

Lessons From Bad Bertrich V

MY round-five game, in which I was Black against Fide master Norbert Heck (2235), featured an instructive middlegame moment.

White has just played 22.Rc1-c7, defending the bishop on b7, which has helped snaffle a pawn

*****

*****

*****

*****

*****

I replied with 22...Rfd8?, the idea being that White is forced to move the bishop from b7 as Black threatens to kick the queen's rook with 23...Be5.

However, after 23.Ba6 Rxb2 (there is nothing better) White played 24.Bc4, when Black must lose either the f or a pawn.

Black is lost, according to Stockfish18 and Dragon1, as I show in my notes to the game.

What is more important is that in the diagram Black has an instructive way to equalise (Stockfish18) or at least be only slightly worse (Dragon1).

*****

*****

*****

*****

*****

The way to do this is to give up a second pawn by 22...d3!?, the point of the move being to activate Black's bishop.

After 23.exd3, Black does not take on b2 straightaway (if at all!), but instead hits the white queen's rook with 23...Be5.

That effectively forces 24.Rd7, and now 24...Rfd8 25.Rxd8 Rxd8 results in two White pawns being en prise.

White cannot defend both of them, so, after Stockfish18's line, 26.b3 Rxd3 27.Re1 f6, Black has regained one pawn, and has the better coordinated pieces.

Dragon1 suggests 26.Re1 Bxb2 27.Ba6, after which White has a passed pawn, but it is difficult to see how White makes progress following, for example, 27...Bd4.

LESSON: piece activity is hugely important, and can be well worth achieving at the price of a pawn, especially as the pawn may well be regained in short order.

Lessons From Bad Bertrich IV

MY round-four game, in which I was White against Joachim Hiller (1879), featured an instructive moment in the opening, an Exchange Variation of the Spanish.

I have just played 10.Bc1-f4, and it is up to Black to decide what to do about the attacked c7 pawn

*****

*****

*****

*****

*****

As is well known, White in the Spanish Exchange gives up the bishop-pair in order to obtain the superior pawn-structure.

White gets a healthy 4-3 kingside majority, from which a passed pawn can be made with normal play, while Black gets a damaged 4-3 queenside majority, which will not produce a passed pawn with normal play, unless White gets careless.

In the diagram position, Black has three plausible ways of saving c7: (i) by castling, (ii) by defending the pawn with the dark-square bishop, (iii) by advancing the pawn.

I suspect most players would pretty quickly dismiss 10...c6 as not helping development, as well as weakening Black on the b8-h2 diagonal, including the square d6.

The overwhelmingly most popular choice in ChessBase's 2026 Mega database is 10...0-0-0, which not only defends the pawn, but gets the king to safety, and develops the rook - three good things for the price of one!

However, I was surprised 10...Bd6 appears only once in Mega26.

After 11.Bxd6 cxd6, Black no longer has the bishop-pair, but also no longer has doubled pawns. However, Black has also contracted a backward pawn on a half-open file.

The engines very much disapprove of 10...Bd6.

Part of the reason, no doubt, is the backward pawn, but there is also the fact that engines rate the bishop-pair as more than cancelling out the disadvantage of having doubled pawns - indeed, engines are not keen on White playing the Spanish Exchange from the get-go.

The move 4.Bxc6 sees Stockfish18 and Dragon1 drop their evaluation of the opening from slightly better for White to a game with equal chances.

Garry Kasparov might well agree, as I recall a John Watson review for The Week In Chess in which he quotes Kasparov as joking that "the bishop is always better than the knight!"

LESSON: the bishop-pair is a major factor in chess, and should not be given up without very good cause.

Lessons From Bad Bertrich III

THERE is a fairly well-established hierarchy when it comes to the drawishness of endings.

At the top, and by a considerable distance, are endings with opposite-colour bishops.

Next come rook-and-pawn endings, their drawing tendencies illustrated by the aphorism, usually attributed to Tarrasch or Tartakower, that "all rook endgames are drawn."

Third in the drawishness stakes are queen endings, thanks to the danger of the weaker side escaping defeat with perpetual check.

After that the hierarchy becomes somewhat murky, but I would suggest the list continues with same-colour bishop endings, followed by knight endings, and finally pawn endings (note I am only listing endings where both sides have the same pieces, ie leaving out, for example, rook versus bishop and knight).

All of this is a preamble to the strange finish in my round-three game, where I was Black against Børge Svanholm (1871).

The following rook-and-pawn ending was reached after White's 27th move.

White has just captured on d8, leaving a commonly occurring rook-and-pawn ending in which material is level, but one player has a farside passed pawn

Such positions are normally quickly agreed drawn at high levels, and often at club level too.

However, if the game continues, the normal procedure is that the player with the passed pawn - in this case White - protects it from the front, and advances the pawn to the seventh rank, while the black rook keeps an eye on the pawn from behind.

There are two dangers for Black to avoid, once the pawn has reached White's seventh.

One is to carelessly put the black king in such a position on Black's third rank, or higher, that it can be checked by the white rook, allowing the a pawn to be advanced on the next move.

The other is to let the white king get over to the queenside to safely help promote the pawn, although this is a risky business if it means Black mops up White's kingside pawns.

But if White tries nothing, and Black does not get careless, the game is inevitably drawn.

Here is the position seven moves after the previous diagram

Material is still level, and the ending is still completely drawn. Indeed, the ending would be a tablebase draw even if the h pawn were removed from the board.

Six moves later, the h pawn was removed from the board after Black in this position played 40...Rxh5

The Syzygy endgame tablebase confirms that the position is still drawn, and would remain so after any move by the white rook along the back rank.

Instead, my opponent committed the cardinal sin of leaving a drawn rook-and-pawn ending for a lost pawn ending by playing 41.Rxh5+, resigning 10 moves later.

LESSON: unless you have no other option, never swop off into a pawn ending unless you are certain of the outcome.

Lessons From Bad Bertrich II

MY round-two game, in which I was White against Wolf Ebert (1725), was a French Defence that began 1.e4 e6 2.d4 d5 3.Nc3.

A position familiar to every tournament player, even if it has not appeared in one of their games

*****

*****

*****

*****

*****

*****

*****

The position occurs 239,392 times in ChessBase's 2026 Mega database, with the most popular continuations being 3...dxe4 (27,031 games), 3...Nf6 (94,362) and 3...Bb4 (106,481).

Geoffrey Stern (1931, Luxembourg) told me later that, although he had played the French for many years, my opponent's continuation was one that GS had never even considered playing.

The move my opponent played was 3...Ne7!?, which frankly looks rather silly at first, but appears 251 times in Mega26.

One of my first thoughts was to reply 4.Be3, to make it difficult for Black to get in ...c5. But then one of the points of my opponent's move becomes apparent, as he can play 4...dxe4 5.Nxe4 Nf5, with ...Nxe3 to come, when Black at least has the bishop-pair for his trouble.

I finally settled on 4.Bd3, which is reasonable enough, but the commonest continuation in Mega26, and the top choice of Stockfish18 and Dragon1, is 4.Nf3.

(Note that a move such as 3...Ne7!?, no matter how awful it may look, cannot be refuted out of hand - there is no tactical blow to punish Black.)

After 4.Nf3, the engines for quite some time completely disagree as to how Black should continue, but eventually agree that best is 4...Nbc6!?

This may look odd, but a knight on c6 is well-placed to harass a white bishop on d3 (by continuing ...Nb4) and may support a later ...e5.

After 4...Nbc6!?, the continuation 5.Bd3 scores a disappointing 47% in Mega26, but 5.e5, ensuring Black cannot play ...e5 in the foreseeable future, rings up a much more satisfactory 61%.

Black can attack the head of the pawn chain with 5...f6 - Nimzowitsch might have disapproved - when the relatively high-level game, Yunguo Wan (2471) - Ngoc Truong Son Nguyen (2579), PGMA Cup (Manilia, Philippines) 2008, continued 6.exf6 gxf6 7.Nh4!? Ng6?! (the engines reckon 7...Bg7 improves) 8.Qh5 Ne7 9.Nxg6 Nxg6 10.Bd3 f5 11.Bg5 Be7 12.h4 Qd6 13.0-0-0 Bd714.Bxe7 Kxe7! (14...Qxe7 15.Nxd5!) 15.g3 Rag8, when the engines agree White has at least the upper hand, but the game was drawn after 41 moves.

A later game, but also at quite a high level, Axel Delorme (2495) - Aleks Sahakyan (2243), Chess.com Blitz 2023, varied with 7.Bd3. There followed 7...a6!? 8.Bf4 Ng6 9.Bg3 Bd6 10.Qe2 Nb4!? 11.0-0-0 Nxd3 12.Rxd3 Nf4!? 13.Bxf4 Bxf4+14.Kb1, with a slight edge for White, according to the engines (but 0-1, 49 moves).

LESSON: following established theory is often be safe and reliable, if a little predictable, but striking out on one's own, even with a move that may cause 'superior' players to smile or wince, can be invigorating, and may bring surprising success if the move is played with specific worked-out ideas in mind.

Lessons From Bad Bertrich

I CAN recall reading an article by Cecil Purdy, the first chess world correspondence champion, in which he stated that, though it is hard to believe, novices can struggle with mate-in-one problems.

It is easy to laugh at such difficulties, but I suppose the equivalent among experienced players would be to miss a mate-in-two.

Unfortunately, that happened to me in the 50+ at Bad Bertrich in game one, in which I was Black against Jürgen Dieter (1769).

The critical position arose after White's 40th move

*****

*****

*****

*****

*****

Knowing something is on, it is not difficult to find 40...Rb3+! 41.axb3 Rxb3#.

The situation of the white king, and specifically its inability to retreat, is a little unusual, but hardly unprecedented.

I have preached in previous blogs about how I try to follow a mantra, Captures-Checks-Hangers-Threats, at every move.

This entails firstly looking at every possible capture, not matter how apparently absurd, followed by looking at every possible check, again no matter how foolish-looking.

I am then supposed to look at hangers, ie pieces and pawns that are not protected, and so are 'hanging' or 'loose'. If a player has two or more, a forking combination may be in the air.

And finally I look at threats, eg a minor piece that can 'hit' the opponent's queen, or a pawn advancing against a piece short of squares, or a move lining up a checkmate.

The mantra takes longer to explain that it usually takes to carry out - most possible captures, for example, can be dismissed at once as of poor value, and lots of possible checks are either pointless or downright harmful to the checking side.

But if, in the diagrammed position, I had seriously considered, just for a moment, ...Rb3+, I would surely have seen that it was by no means as ridiculous as it might appear at first sight.

LESSON: there is no point in having a mantra if you do not use it.

Summing Up Bad Bertrich 50+

Carving - Falkenlay

My score of +5=3-1 gained 35.4 Fide elo. I finished equal 3rd-4th, but was placed third on tiebreak (average rating of opponents), winning 150 euros (fourth won 120 euros).

Bad Bertrich 50+ Game Nine

Heron (or similar)

I was upfloated two points to top board against the No1 seed, an international master who had won his first eight games.

Spanton (1997) - Dieter Pirrot (2321)

French Burn

1.e4 e6 2.d4 d5 3.Nc3 Nf6 4.Bg5 dxe4!?

This lags behind 4...Be7 in popularity in ChessBase's 2026 Mega database, but is fashionable, and is preferred by Stockfish18 and Dragon1.

5.Nxe4 Be7 6.Nxf6+!?

The main move in Mega26 is 6.Bxf6, after which Black chooses between 6...Bxf6, which can transpose to the game, and 6...gxf6, when Black keeps the bishop-pair, at the expense of a damaged kingside pawn-structure.

6...Bxf6

6...gxf6!? seems to make less sense now, as Black no longer gains the bishop-pair, but it has been played by grandmasters

7.Bxf6 Qxf6 8.Nf3 0-0 9.Qd3!?

The engines prefer this, marginally, over the much more common 9.Bd3.

9...c5 10.0-0-0!?

This is the point behind 9.Qd3!?

Now both sides have castled, how would you assess the position?

*****

*****

*****

*****

*****

White has a slight lead in development, which may be enough for an edge, although the engines' evaluations fluctuate, with Stockfish18 starting off by giving White a slight edge, whereas at the same moment Dragon1 calls the game completely equal. But slowly the two engines' evaluations converge, and briefly Dragon1 is more fond of White's position than is Stockfish18.

10...Rd8 11.Qe3?!

White has problems after this. Probably best is 11.Qe4, and if, as in the game, 11...Nc6, White has time for 12.Bd3, when Kramnik has tried 12...Nxd4!? 13.Qxh7+ Kf8, but the engines give White at least a slight edge. They prefer 12...Qh6+ 13.Kb1, and either 12...cxd4 or 12...Nxd4, winning a pawn, but with White having good compensation, eg 12...Nxd4 (Stockfish18's top choice) 13.Rhe1!? Rb8!? (apparently a novelty) 14.Nxd4 cxd4 (14...Rxd4? 15.Qe5) 15.Qe5 Bd7 16.Be4 Qf6!? 17.Qxf6 gxf6 18.Rxd4, when White has regained the pawn, and has the better structure.

11...Nc6 12.d5!?

Maksim Chigaev (2632) - Vladimir Kramnik (2753), Chess.com Blitz 2023, went 12.Bd3?! cxd4 13.Qe4 g6 14.Rhe1 Bd7, with the upper hand for Black, according to the engines (0-1, 42 moves).

12...exd5

How should White proceed?

*****

*****

*****

*****

*****

13.Qxc5

Possibly better is 13.Rxd5!? Be6 (13...Rxd5?? 14.Qe8#) 14.Rxd8+ Rxd8, and now either 15.a3 or 15.Bd3, but in each case the engines reckon Black has at least the better part of equality.

13...Bf5!?

This apparent-novelty, targeting c3, is the top choice of Dragon1, for a while, but it comes to concur with Stockfish18 in preferring the known 13...Bg4, when 14.Be2 Rac8 gave Black an initiative in Aspasia Giaidzi (2360) - Susan Lalić (2320), Athens Scheveningen (Women) 1992 (0-1, 47 moves), and Gyula Mészáros (2419) - Kornél Szalai (2315), Hungarian Team Championship 2006 (1-0, 44 moves).

14.Nd4!

If 14.Bd3, Black has 14...Bg4!?, when White has nothing better than 15.Be2, transferring to the previous note.

14...Nxd4 15.Qxd4 Rac8 16.c3 Qh6+

The engines suggest 16...Qg5+!? 17.f4, and now 17...Qh6, claiming Black is at least slightly better, one point being 18.Bd3 (this is best, according to the engines) can be met by 18...Be4!?, when f4 is hanging, along with g2.

17.Qd2?

White has no problems after 17.Rd2, according to the engines.

Black to play and gain what the engines reckon is a positionally won game

*****

*****

*****

*****

*****

17...d4?
White does not have a satisfactory answer to 17...Qg6!, according to the engines, the threat being 18...d4. White can stop this with 18.Qd4, but 18...Bb1 threatens both mate and the a2 pawn. White can stop both of these with 19.Bd3, but 19...Bxd3 20.Qxd3 allows 20...Qxg2, when the half-open g file is not of much use to White, with only heavy pieces on the board. Other queen moves by Black are also better than the text.

18.Qxh6 gxh6 19.Bd3 Be6 20.Bb1

Not 20.Kb1 dxc3 21.bxc3, as Stockfish18 points out the possibility of 21...Rxd3!! 22.Rxd3 Bf5 23.Kc2 Rd8 24.Rhd1 b5, when it is only now Dragon1 sees the danger. The point is that the black rook and bishop are holding all three white pieces in place, meaning Black has a free piece in the shape of the king. A sample line runs 25.Rd1 Kg7 26.Rd2 h5 27.Rd2 h4 28.Rd1 Kh6 29.Rd2 Kg5 30.h3 Be4 31.Rd1 Kf4 f5 32.Rd1 Ke5 33.Rd2 f4 34.f3 Bxd3+ 35.Rxd3 Rxd3 36.Kxd3 Kd5, when Black's reserve tempi make this an easily won pawn ending, eg 37.a3 a5 38.Kd2 Kc4 39.Kc2 a4 etc.

20...b5 21.Rd3 dxc3

If 21...b4, then 22.Rhd1 bxc3 23.bxc3, when both black further captures on c3 lose.

22.Rxd8+ Rxd8 23.bxc3

Who, if anyone, has the advantage in this rook-and-bishop ending?

*****

*****

*****

*****

*****

Neither side, it seems - the engines call the game completely equal.

23...Bc4

The engines suggest 23...Rc8 or 23...Rb8!?

24.Re1

Both engines are at first are quite enthusiastic about 24.Rd1, but Stockfish18 soon changes its assessment to equal, although Dragon1 persists in awarding White a slight edge.

24...Rd6 25.Re3 Rf6 26.f3 Ra6 27.Kb2 Rd6 28.Kc1 Kg7 29.a3 Ra6 30.Kb2 Rd6 31.Kc1 Ra6 ½–½

Bad Bertrich 50+ Game Eight ADDENDUM

JOE Skielnik points out that in the line 18...Nxe4 19.fxe4 f5 20.Bxg7+! Kg8!, White is not forced to allow the perpetual that follows 21.Bxf8, and can instead play 21.Bf3!?, giving the following position.

Play should then continue 21...Qxf3+ 22.Qf2, after which best may be 22...Qh1+ 23.Qg1 Qxg1+ 24.Kxg1 Kxg7

*****

*****

*****

*****

*****

Black is two pawns up, but Stockfish18 and Dragon1 reckon White has full compensation, one line running 25.Nxb6 Rad8 26.Rxd7+ Rxd7 27.Nxd7 Rg8!? 28.exf5 exf5 29.Kf2.

Position after 29.Kf2

*****

*****

*****

*****

*****

White has three connected pawns against Black's three isolanis, but Black has the better minor piece for working with a rook, especially in an open game with play on both sides of the board.

The position is completely equal, according to the engines, but would be fascinating to play out over the board.

Bad Bertrich 50+ Game Eight

Beaver (I think) preening itself - three days after a wild boar rushed past within 20 yards of where I was walking on a forest track (but I did not even come close to snapping a pic of that animal)

I was downfloated.

Spanton (1997) - Horst Prüsse (2006)

Maróczy Bind

1.e4 c5 2.Nf3 a6 3.c4!?

This is the third-most popular response to the O'Kelly Variation of the Sicilian, behind 3.c3 and especially 3.d4 in ChessBase's 2026 Mega database, but it is the top choice of Stockfish18 and Dragon1.

3...e6 4.Nc3 Qc7!?

This committal move with the queen is easily the commonest continuation in Mega26, and was the choice of Albéric O'Kelly de Galway when he reached the position in 1948, although 15 years later he preferred 4...Nc6 (both games were drawn).

5.d4 cxd4 6.Nxd4 b6!?

This was O'Kelly's choice is the 1948 game, his opponent replying 7.Bd3.

7.Be3 Nc6 8.Rc1!? Nf6

Branko Damljanović (2504) - Igor Miladinović (2536), Serbian Team Championship 2021, went 8...Bb7 9.a3!? Nf6 10.Be2 Bc5 11.b4 Nxd4 12.Bxd4 Bd6!? 13.Qd3 Rc8!? 14.g3 h5!? 15.h4, with at least a slight edge for White, according to the engines (½–½, 37 moves).

9.Be2 Bb4 10.f3 0-0 11.Qd2!? Kh8!? 12.0-0

Now both sides have castled, how would you assess the position?

*****

*****

*****

*****

*****

The opening has been a success for White. Black has not challenged the Maróczy Bind with any of the three classic pawn-breaks, ...b5, ...d5 and ...f5, and White is ahead on development. The engines give White the upper hand.

12...Nxd4 13.Bxd4 Bc5 14.Rfd1 Bb7 15.Na4 Bd6 16.g3?

White has at least the upper hand after both 16.h3 and 16.e5!?, according to the engines, eg 16.e5!? Bxe5 17.Bxe5 Qxe5 18.Nxb6.

16...Bxg3! 17.hxg3 Qxg3+ 18.Kf1

Black to play and (at least) draw

*****

*****

*****

*****

*****

18...Bxe4?

Correct is 18...Nxe4 19.fxe4 f5 20.Bxg7+! (everything else loses) Kg8! (White wins after 20...Kxg7? 21.Qc3, and is on top after 20...Qxg7?! 21.Qe3!) 21.Bxf8 Qh3+ 22.Kf2 Qh4+ 23.Kf1 (not 23.Ke3? Qxe4+ 24.Kf2 Qg2+ 25.Ke3 Qg3+ 26.Bf3 Qxf3+ 27.Kd4 e5+ 28.Kxe5 Qe4+, when mate follows in short-order) Qh3+ etc.

19.fxe4 Qh3+ 20.Ke1 Nxe4 21.Qf4

Not the only winning move.

21...Qh1+ 22.Qf1 Qh4 23.Bf2

White's king has no pawn cover, but Black's attack is over, and White has emerged with two bishops for four pawns

*****

*****

*****

*****

*****

23...Qe7 24.Nc3 f5 25.Qg2 d5!?

Seeking to open lines to the white king, or create a big pawn-roller, but only a major blunder can change the outcome.

26.cxd5 exd5 27.Nxd5 Qd6 28.Bd4 Rg8 29.Nc3 Qh6 30.Nxe4 fxe4 31.Rc3 e3!? 32.Bxe3 Qh4+ 33.Bf2 Qb4 34.Rd2 Rad8!? 35.Rxd8 Rxd8 36.Qg5 Re8 37.Qd2 Qe4 38.Re3 Qb1+ 39.Qd1 Qg6 40.Rxe8+ Qxe8 41.Bxb6 1-0

Bad Bertrich 50+ Game Seven

Looking down on Bad Bertrich

Michael Hillenbrand (2113) - Spanton (1997)

Réti

1.c4 e6 2.Nf3 d5 3.e3!?

The main continuations in ChessBase's 2026 Mega database are 3.d4 and especially 3.g3, but the text is a fashionable alternative, one of the points being it prevents 3...d4.

3...Nf6

3...c5 renews the 'threat' of playing ...d4, so White usually continues with 4.d4 or 4.cxd5 exd5 5.d4, in both cases with equal chances, according to Stockfish18 and Dragon1.

4.b3!?

This is easily the most popular move in Mega26.

4...Bd6!?

The bishop more often goes to e7.

5.Bb2 0-0 6.g3!? b6!?

The engines prefer 6...dxc4!? or 6...Nc6!?

7.Bg2 Bb7 8.0-0

Now both sides have castled, how would you assess the game?

*****

*****

*****

*****

*****

The position occurs 39 times in Mega26, including eight games in which Nakamura was White. His score from the diagram is +5=2-1, but, according to the engines, White has no more than the better part of equality.

8...Nbd7 9.d4

Nakamura has played this, but also Qe2, cxd5 and d3.

9...c5!?

This may be slightly premature. The engines suggest 9...dxc4!? 10.bxc4, and now 10...c5.

10.cxd5 Bxd5

The engines agree this is better than 10...exd5.

11.Nc3 Bc6!?

Keeping a watch on the b5 square, but the engines prefer the conventional retreat 11...Bb7.

12.Qe2 Ne4!?

Black should probably play 12...cxd4, obliging White to accept an IQP or to allow an exchange of light-square bishops.

How should White proceed?

*****

*****

*****

*****

*****

13.dxc5

Stockfish18 gives on 13.Rad1 Qe7 14.Qd3 Nxc3 15.Bxc3 Bxf3!? 16.Bxf3, claiming a slight edge for White. Taking on f3 looks strange, but if, say, 15...Rfd8, White has 16.Ng5! Qxg5 17.Bxc6, gaining the bishop-pair anyway.

Dragon1 suggests 13.Nb5 Bb7 (13...Bb8? 14.Nd2) 14.Rfd1 a6 15.Na3!?, also claiming a slight edge for White.

13...Nxc3 14.Bxc3 Bxc5

The engines prefer 14...Nxc5.

15.Rfd1 Qe7 16.Rac1 Rac8 17.Nd4!?

The engines give 17.b4!?, one point being 17...Bxb4?! runs into 18.Bxb4 Qxb4 19.Rxc6! Rxc6 20.Rxd7, although this is not as strong as it first appears, as Black has 20...Qb1+ followed by ...Rc2.

17...Bxg2 18.Kxg2 Nf6 19.Qf3 Nd5

One line given by the engines runs 19...e5!? 20.Nf5 Qe6, with roughly equal chances.

What should White play?

*****

*****

*****

*****

*****

20.Nf5!?

This is Dragon1's top choice; Stockfish18 prefers 20.Bb2.

20...exf5 21.Qxd5

The engines slightly prefer 21.Rxd5.

21...Ba3 22.Rc2 Rfd8

What would you play as White?

*****

*****

*****

*****

*****

23.Qf3?

Contrary to first impressions, perhaps, White is at least equal in the diagram position, thanks to Black having back-rank mating problems. Best play, according to the engines, runs 23.Qe5! (23.Ba1!? is also fine) Qxe5 24.Bxe5 f6 25.Ba1 Rxc2 (the engines also like 25...Bc5) 26.Rxd8+ Bf8 27.a4, with complete equality, according to Stockfish18, although Dragon1 gives White a slight edge.

23...Rxd1 24.Qxd1 Qe4+

The immediate 24...Bb4 may be even stronger.

25.Kg1

How should Black proceed?

*****

*****

*****

*****

*****

25...h6

White is off the hook after 25...Bb4? 26.Rc1, and if 26...Ba3, then 27.Bb2!? or 27.Qd7.

26.Qd7

There is nothing better, eg 26.Qd2 is met by 26...Bb4.

26...Rc5 27.Qd8+?!

Objectively better is 27.Qd4 Qxd4 28.exd4 Rc8 29.d5!, when 29...Bb4?! 30.d6 Bxd6 31.Rd2 Rc6 leaves Black a doubled pawn up, and winning, according to the engines, but stronger is 29...f4!, eg 30.gxf4 Bd6 31.Kg2 b5 32.a3 Bxf4 33.d6 Bxd6 34.Rd2 Bxa3, with a two-pawn advantage, or 30.Kf1!? Bb4 31.d6 Bxd6 32.Rd2 Rxc3 33.Rxd6 Rc2 34.Rd8+ Kh7 35.gxf4 Rxa2, when Black is a pawn up, and all of White's pawns are isolated.

While the text is objectively weaker, keeping queens on gives Black more chance to go wrong.

27...Kh7 28.Bxg7 Kxg7 29.Rxc5 Bxc5

How large is Black's advantage in this ending?

*****

*****

*****

*****

*****

Black's kingside pawns are weak, but Black is up bishop for pawn, and has an advantage worth comfortably more than a minor piece, according to the engines.

30.Qd7 a5 31.Qd1 Kg6 32.Qe2 Qd5 33.Qc2 Qf3 34.Qd3 Qe4 35.Qe2 Qe5 36.Kg2 Qd5+ 37.Kg1 Bb4 38.Qb2 Qd2 39.Qh8 Qc3 40.Qg8+ Qg7 41.Qd8 Bc5 42.Kg2 Qf6 43.Qg8+ Qg7 44.Qd8 Kh7

Black has not made much progress after much manoeuvring, but also has not jettisoned a pawn, or otherwise reduced the winning chances

*****

*****

*****

*****

*****

45.Qd7 Qg6 46.Qd3!? f4!? 47.e4 h5 48.Qf3 Bd6 49.Kf1 Kh6 50.Qd3 Qf6 51.Kg2 Be5 52.Kg1 Qc6 53.Qe2 Qc5 54.Kg2 Qd4 55.Qf3 Qd6 56.Qe2 Qf6 57.Qd3 Qg5 58.Qd7 Kg7 59.Qd1 h4

At last I felt able to move in for the kill

*****

*****

*****

*****

*****

60.Qf3 Kg6 61.Qd3

61.g4 possibly holds out longer, but makes little difference to the engines' assessment of the position.

61...Qh5 62.Qd7 h3+!? 63.Kf1

Not 63.Qxh3? f3+.

63...fxg3 64.hxg3 h2 65.Qf5+ Qxf5 66.exf5+ Kxf5 67.Kg2 Ke4 0-1

Bad Bertrich 50+ Game Six

                                                             Klidinger waterfall

I was upfloated.

Spanton (1997) - Martin Schopf (1909)

Veresov

1.d4 d5 2.Nc3 Nf6 3.Bg5 h6!?

Putting the question to the bishop is somewhat uncommon - sixth-most popular in ChessBase's 2026 Mega database, behind ...c5, ...c6, ...Bf5, ...e6 and especially ...Nbd7. But it is Stockfish18's top choice, fractionally ahead of the mainline...Nbd7, while Dragon1 comes to like it almost as much as ...Nbd7.

4.Bh4

The Trompowsky-like 4.Bxf6 is more usual, when the recapture 4...exf6 is the top choice of humans and engines.

4...Bf5

Developing the light-square bishop outside of the pawn-chain becomes a serious option when White cannot attack Black's centre/queenside with a combination of c4, Qb3 and Nc3.

5.e3 Nbd7 6.Bd3 Bxd3

6...e6? can be met by 7.Nxd5!

7.cxd3!?

The engines marginally prefer this over 7.Qxd3.

7...c6 8.Nf3 e6 9.0-0 Be7 10.Rc1 0-0

Now both sides have castled, how would you assess the position?

*****

*****

*****

*****

*****

White is slightly more developed, and has a half-open c file, but the game is equal, according to the engines.

11.Na4 Nh5!? 12.Bxe7 Qxe7 13.Nc5!?

This may be new. Héctor Rossetto - Miroslav Filip, Mar del Plata (Argentina) 1961, went 13.Qb3 Rab8 14.Qc3!? Nhf6 15.b4 Rfc8 16.Rb1 Nb8!? 17.Rfc1 Nd6 18.h3!?, when Dragon1 prefers White, but Stockfish18 calls the game equal (½–½, 38 moves).

13...Nxc5 14.Rxc5 Nf6 15.Ne5 Nd7 16.Nxd7 Qxd7

How would you assess this heavy-piece middlegame?

*****

*****

*****

*****

*****

The game is completely equal, according to Stockfish18, although Dragon1 awards White the better part of equality.

17.Qa4 a6 18.Rfc1 Rfd8!?

The engines are fine with this, but it does seem a mysterious rook move.

19.b4 g6 20.h3 Kg7 21.Qa5 Rdc8 22.a4 Qd8 23.Qxd8 Rxd8

How would you assess this rook-and-pawn ending?

*****

*****

*****

*****

*****

Stockfish18 sticks with complete equality, while Dragon1 gives White a miniscule edge of under 0.1.

24.Kf1 Rd7 25.Ke2 Kf6 26.Kd2 Ke7 27.Kc3 Kd6

It is still possible for either side to blunder, eg 28.h4?? loses 28...b6.

28.Kb3 h5 29.h4 Ra7!?

Another somewhat strange-looking rook move, but again the engines are not bothered.

30.e4 f6 31.Re1 Re7 32.Rc2 b6!?

The engines react badly to this at first, but Stockfish18 comes to view it as OK, although Dragon1 is less sure.

33.e5+ fxe5

Stockfish18 suggests 33...Kd7!?, the engines agreeing 34.exf6 Rf7 is equal. However, Dragon1, but not Stockfish18, reckons 34.f4!? gives White a slight edge.

34.dxe5+ Kd7 35.Rec1 Rc7 36.d4

Black to play and avoid getting a difficult game

*****

*****

*****

*****

*****

36...b5

Black needed to play 36...Re8!, the point being to meet 37.b5 axb5 38.axb5 with 38...Rec8, giving complete equality, according to the engines.

37.f4?!

Better is 37.axb5 axb5, and now 37.f4, with play similar to the game.

37...Re8?

This is now bad, whereas 37...bxa4+ 38.Kxa4 Rf7 gives Black sufficient counterplay to compensate for having weak a and c pawns, according to the engines.

38.a5

Played to prevent Black getting possible counterplay down the a file after the action switches, as I intended, to the kingside. But the engines much prefer 38.axb5!? axb5 39.Ra2, taking the open file, a sample line continuing 39...Rf8 40.g3 Ke7 41.Kc2 Rd8 42.Kd3 Kf8 43.Ke3 Kg7 44.Rca1 Kf7 45.Ra7 Rdc8 46.Kf3 Kf7 47.Rxc7+ Rxc7 48.g4 Kf7 49.gxh5 gxh5 50.Ra8 Re7 51.Rh8, after which White wins either the h or c pawn.

38...Rf8 39.Rf2 Rcc8 40.Kc3 Rf7 41.Kd3 Rcf8 42.Ke3 Rf5 43.Rf3 R5f7

Black can do nothing but await White's activation of the 4-3 kingside majority, in sharp contrast to Black's immobile 4-3 queenside majority

*****

*****

*****

*****

*****

44.Rc2 Rf5 45.Rcf2 R5f7 46.Rg3 Rg8 47.Rff3 Rf5 48.Rg5 Rxg5?

This makes White's task easier.

49.fxg5 Ke7 50.Rf6 Rg7 51.Kf4 Rg8 52.g4 hxg4 53.Kxg4 Rg7 54.h5 gxh5+ 55.Kxh5 Rh7+

All three legal continuations win

*****

*****

*****

*****

*****

56.Kg4 Rh1 57.g6 Rb1 58.Rf7+ Ke8 59.Kg5 c5!?

This is Stockfish18's top choice, at least for a while, and Dragon1 comes to agree.

60.dxc5 d4 61.c6 Rc1 62.Rd7 Rxc6 63.g7!? Kxg7 64.g8=Q Kc7 65.Qf7+ 1-0

Bad Bertrich 50+ Game Five

World's longest two-seater bench?

Faced a Fide master.

Norbert Heck (2235) - Spanton (1997)

Queen's Pawn Game

1.d4 e6 2.Nf3 d5 3.g3 c5 4.Bg2 cxd4 5.Nxd4 Nf6 6.0-0 e5!?

This is the commonest continuation in ChessBase's 2026 Mega database, and is the top choice of Stockfish18 and Dragon1.

7.Nb3 Be6 8.c4

This and 8.Bg5 are the most popular moves.

8...Nc6 9.cxd5 Bxd5

The engines prefer 9...Nxd5.

10.Bh3!? Be7 11.Nc3 Bc4

How should White proceed?

*****

*****

*****

*****

*****

12.Bg2!?

Marius Manolache (2520) - Andrei Filip (2222), Tășnad Orașului Cup (Romania) 2021, went 12.Bg5 0-0 13.Nd2!? Ba6 14.Bg2, with an equal game, according to the engines (1-0, 37 moves).

Mikhail Demidov (2523) - Danii Voloshkin (2081), Russian Team Blitz Championship (Sochi) 2025, saw 12.Be3 0-0 13.Qxd8!? Rfxd8 14.Rfd1 b6!? 15.Bg2, again with an equal game, according to the engines (1-0, 39 moves).

The text, which may be a novelty, is the engines' top choice, albeit by a small margin.

12...0-0 13.Bg5 h6

The engines suggest 13...Qb6 or 13...Rc8.

14.Bxf6 Bxf6 15.Nd5 Rc8

Here the engines like 15...Nd4 or 15...e4!?

16.Rc1 Bxd5 17.Qxd5 Qb6 18.Qe4 Nd4!?

Dragon1 suggests 18...Rfd8 or 18...Be7, but Stockfish18 likes the text, even though after ...

19.Nxd4 exd4 20.Qxb7 Qxb7 21.Bxb7 Rab8 22.Rc7

... White has won a pawn.

How should Black continue?

*****

*****

*****

*****

*****

22...Rfd8?

The engines reckon Black should activate the bishop, eg 22...d3!? 23.exd3 Be5 24.Rd7 Rfd8 25.Rxd8+ Rxd8, after which Black wins back one pawn, and has enough activity to compensate for the other lost pawn, according to Stockfish18, although Dragon1 is less sure.

23.Ba6 Rxb2

Material is equal, but the engines reckon White is winning

*****

*****

*****

*****

*****

24.Bc4 a5 25.Rxf7 Kh8

The engines marginally prefer 25...Rc2!?, and if 26.Rc7+, now 26...Kh8, although they still reckon White is winning.

26.Rd1 Rc8 27.Be6 Re8 28.Bc4 Rc8 29.Ba6 Ra8 30.Bd3 Rxa2

Again material is level, but White has a winning attack

*****

*****

*****

*****

*****

31.Rb1 a4

There is no satisfactory defence. The best the engines can come up with is 31...Ra3 32.Ra7! Rc8 33.Kg2 a4 34.Rb5 Rb4 35.Rbxa5, after which White is a pawn up, but White's advantage is worth about a minor piece, according to the engines.

32.Rfb7! g5?

The engines suggest giving up the exchange with 32...Rg8!?, but after 33.Bc4 Rc2 34.Bxg8 Kxg8 35.Ra7 Rxe2 36.Rxa4 Black has no compensation, despite the passed d pawn, eg 36...d3?! 37.Rd1 d2 runs into 38.Kf1.

33.Rh7+ Kg8 34.Rxh6 Kg7 35.Rb7+

If now 35...Kxh6, then 36.Rh7#.

35...Kg8 1-0

Bad Bertrich 50+ Game Four

Green lake

Spanton (1997) - Joachim Hiller (1879)

Spanish Exchange

1.e4 e5 2.Nf3 Nc6 3.Bb5 a6 4.Bxc6 dxc6 5.Nc3!?

This trails 5.d4 and especially 5.0-0 in popularity in ChessBase's 2026 Mega database.

5...f6 6.d4 exd4 7.Nxd4!?

7.Qxd4 would likely transpose to 5.d4 lines.

7...c5 8.Nde2 Qxd1+ 9.Nxd1 Bd7 10.Bf4 0-0-0

Stockfish18 and Dragon1 much dislike 10...Bd6!?, claiming 11.Bxd6 cxd6 leaves White with the upper hand.

11.Ne3 Ne7 12.0-0-0

Now both sides have castled, how would you assess the position?

*****

*****

*****

*****

*****

Black's bishop-pair and White's superior pawn-structure balance each other out, leaving the game completely equal, according to the engines.

12...g5!?

This is narrowly more popular than 12...Re8!? and 12...Ng6 in Mega26.

13.Bg3 h5 14.h4 Bh6?

New and active, but bad. Known moves are 14...g4 and 14...Bg7.

15.hxg5 Bxg5

Or 15...fxg5 16.Rxh5.

16.f4 Bh6 17.Rxh5 b6 18.Rdh1 Bg7 19.Rxh8 Rxh8 20.Rxh8+ Bxh8

How large is White's advantage in this minor-piece ending?

*****

*****

*****

*****

*****

White is 'only' a pawn up, but Black's 4-3 queenside majority cannot create a passer by itself, while White's 3-1 kingside majority should have little trouble doing exactly that. Black still has the bishop-pair, but the engines reckon White's advantage is the equivalent of being more than a piece up.

21.Kd2

The engines suggest 21.f5!?

21...f5!?

This is also the engines' top choice, but after ...

22.e5

... the position is hardly opening for the bishops.

22...Be6 23.a3 Kd7 24.Bh4 Nd5 25.Bg5 b5 26.Ng3 Nxe3 27.Kxe3 Ke8 28.Nf1 Bg7 29.Kf2 c4 30.c3

The engines prefer 30.g4!?

30...c5?

This is a serious positional error. On c5 the pawn hampers the future scope of Black's dark-square bishop, and becomes a target for the white pieces.

31.Ne3 Kf7 32.g4!? Kg6!?

32...fxg4 33.f5 Bc8 keeps alive Black's faint hope of bringing about an ending of opposite-colour bishops.

However, after ...

33.gxf5+ Bxf5 34.Nxf5 Kxf5

... Black can still hope for a draw by exchanging off as many pawns as possible, and then somehow sacrificing the black bishop to leave White with rook's pawn and wrong-colour bishop.

35.Kf3 Bf8

White to play and win

*****

*****

*****

*****

*****

36.e6!

Sacrificing White's extra pawn seems the only way to make progress.

36...Kxe6 37.Ke4

Thanks to Black's stifled 4-3 majority, White is effectively still a pawn to the good. The only fly in the ointment, from a White perspective, is that all the white queenside pawns are on the same-colour squares as Black's bishop, and therefore potentially vulnerable.

37...b4 38.f5+ Kf7

Or 38...Kd6 39.Bf4+ etc.

39.Kd5 Bg7!?

This is Stockfish18's top choice, but the position is hopeless for Black. Dragon1 prefers 39...a5.

The game finished:

40.Kxc4 bxa3 41.bxa3 a5 42.a4 Be5 43.Kd3 Bg3 44.Bd8 Be5 45.Bxa5 Kf6 46.Ke4 c4 47.Bd8+ 1-0