Annotated Atomic Games

Annotation is not a skill that I naturally have – in fact, I tend to avoid annotating or even examining games for the most part. But if you want to improve, you’ll eventually have to look at some games and understand what went on within them and why the moves were made.

As a sample of what to expect with annotated atomic games, I have three examples that were previously published – all were originally published on Leonard Blackburn (Rekursiv)’s atomic site.

tipau – siggemannen, 2005 Atomic World Championship Final annotated by siggemannen and with a preface by Rekursiv. You may download the PGN file here. Or I’ve adapted the game below:

This game was played at FICS on December 1st, 2005 in the finals of the Atomic World Championships organized by Chronatog. I’d like to thank siggemannen for this very instructive analysis, and I would like to thank both players for giving us so many high-quality games over the past few years. — Rekursiv


After a few years of idling, I finally was able to find a game worth enough to annotate. Although many great games have been played by others, I didn’t want to diminish their power with my comments. And for my own games, I just don’t find them good enough. Another thing that bothered me was that I wanted to show a game which could be educational for everyone who is familiar with atomic. There is no use of annotating a pretty attacking game with tons of variations, because it would be hard to follow. Also, the margin for annotation errors is much bigger, atomic is a complex game, and so far there are no computers to check your variations, like in normal chess. Unfortunately most of my games are so shallowly played, without any deep thinking that it would be hypocritical to annotate them. The next annotated game will feature my loss.

As some of you might know there was a World Atomic Championship last year. I thought I would annotate a game of mine against tipau in the final. It features by then a new plan in a fresh line which was developed by tipau (as well as 99% of other atomic lines). Nowadays most people play e6 in answer to Nc3, but back then it was all new and untested…

Before the match I had extensive preparation, after all tipau is the stronger atomic player in my opinion. Among other things I’ve been looking for a good answer to 1.Nc3, tipau’s pet move throughout his atomic career. 1…Nf6 has been under heavy fire and still is these days despite my best efforts to find defenses against white’s numerous threats, so I had to rely on other continuations like 1…e6, 1…e5 or even 1…c6. The last two replies are considered more or less unsound, and so was the first one, until tipau found the queen sac line which seems to give black good counterplay. Before the match we played a few training games where I tested out this new line, with mixed results. I thought I was getting good positions, but white’s queen managed to get free and it was over. It was certain that the same variation would occur in our match, so I had to prepare some kind of antidote…

1 … e6 2. Nf3 Qf6

This is the queen sac line, which was developed by the great man. Main point is to start the attack on the f-square together with moves like Nh6. As white’s knight is pinned against the f-pawn white needs to act quickly not to lose.

3. Ne4

The natural move. now white threatens simple Nxf6 or Ng5 winning the queen. It seems like black is doomed but he has a shot.

3. … Nh6

It would seem like black totally ignores white’s threats. But on a closer look, 4.Neg5 Kd8 5.Nxh7 Bd6 is winning because white cannot stop Bxg3 and eventually Qd4 doubleattacking. of course on 4.Nxf6 Ng4 wins. Lately even 3…Qd4!? right away been tested not without success.

4. g4

Now white still threatens Nxf6 so black must sac his queen.

4. … Qd4 5. Nxd4 Nc6

This is the point of black’s play. He now threatens Nd4 and Nb4 attacking the c-pawn. 6.c3? would be a mistake now because it doesn’t actually protects against either of the knight’s jumps. After 6.c3 Nd4 7.cxd4 Bb4 wins. Therefore:

6. e3

Before this move was found by tipau, it was thought that black was simply winning. 6.e3 is a bit paradoxical as it doesn’t protect against Nb4, but it clears the way for the king which will be important in the following variations.

6. … Nb4 7. Bb5 c6 8. c3

Now 8…Nc2+ isn’t lethal because white’s king has an escape square on f1. One of the first games in this variation went: 8…Nxa2 9.O-O Be7!? 10.h4 O-O 11.Bd3 f5 and white later sacked away his advantage and lost in siggemannen-tipau, 2005. Later 8…Nxa2 was more or less found unsound after a few TrojanKnight-tipau games and the whole variation was considered bad for black.

I earlier tried a line with 8…Nc2+ but didn’t get a satisfactory result. But throughout the games I felt that black had some kind of hidden potential, despite material disadvantage… For me it was a new way of thinking about an atomic position. Back in the old days players used to slug it out in the opening, getting some kind of material advantage with crude attacks.

Well, I wasn’t different. Material was the most important, as endgames were just a walk to victory for the strongest side. You won a pawn or two, exchanged rooks in the heat of the battle, exchanged your pieces for pawns and marched to victory. Nowadays it’s not that easy. Many opening lines are built upon this “potential” of the pieces instead of the plain material count.

An extreme example is I guess:
1.Nf3 f6 2.Nd4 Nh6 3.e3 Ng4 4.f4 b5
in which black sacs a whole bunch of pieces just to have the possibility of catching the white’s queen. Another modern line is 1.Nf3 f6 e3 d5!? which was all rage a year or so back. It is a very clear example of using the potential of piece development and the control over a single semi-open file. Computers generally disregard the positional advantage of open files for the rooks, but I think it’s one of the most important factor in atomic, perhaps even more important than material… After all, a single rook is enough to draw.

To draw a parallel to chess, the old 1.e4 e5 and 1.d4 d5 openings have been
taken over by the dynamism of the Sicilian and the various Indian systems.
Although 1.Nc3 e6 is a highly positional line in contrast with wild slugging of 1.Nc3 Nf6 I, as an attacking player, prefer the former. For me Nc3 e6 is the ultimate Hedgehog of Atomic…

Anyways, back to the game. I didn’t want to get through the complications of 8…Nxa2 so i followed with my original plan:

8 … Nc2+ 9. Kf1 cxb5?!

The novelty by that time i think. So what did black achieve by these moves? White has lost the ability to castle, which is important, as white’s king seems quite unsafe in the center. On the material side black have 3 pieces for the queen minus a pawn, but that’s for the endgame, and we’re still in the middlegame. White’s queen has little to no mobility, and it will take a while to activate it. The activation of the queen will require pawn moves which will increase the “air” behind them, and black’s bishops might be able to use that. Also black might have a possibility of opening a file or two, especially the f-file. Another point is if black would decide to capture the rook at a1 and is able to make his rooks mobile, they can penetrate white’s position and exchange themselves for the queen and rook of the opponent. With all that in mind, I contemplated the current position.

10. d3!?

tipau doesn’t want to play into my plan with the more natural 10.d4, which would have weakened the light-squares around the king. On the other hand white will need to play d4 at some point, otherwise it’s hard to see a way to activate his queen. Nowadays i’ve mostly seen 10.h3!? as to prepare the move 11.g5 with the idea of 11…Nf5? 12.g6!.

10 … Bd6

Black wants to provoke even more pawn-pushing and also prepare castling.

11. f4 g6

I’m not sure this move is the best one as it seems rather slow. On 11…O-O I was afraid of 12.g5 Ng4 13.h3, but black looks good after 13…Nh2+!

12. e4

White starts his pawn-storm, with the idea of e5 next. Notice that other
pawn-pushes besides e5 aren’t as good. On 13.g5 Ng4 would follow and 13.f5 is met with 13…Bf4. 13.h4 with the idea of h4-h5-hxg6 looks more
menacing. Another possibility is perhaps 13.Be3 where white tries to
clear up the space for the queen to breathe. All those threats need to be met somehow…

12 … f5?!

Generally I dislike playing a pawn two squares at a time, but i think it’s justified this time. On 13.g5, I was planning 13…fxe4 (13…Ng4 14.exf5 O-O is interesting). Often it’s good if black is able to stir up complications in atomic, especially if you like that sort of position… The general idea behind f5 is to clear the f-file, then castle quickly, and finally sac on f4 and mate the king.

13. g5 b6

I deviated from my original plan because I sensed that black needed to develop the other wing for the best effect. Now 14.e5 might be dangerous due to 14…Bb7 or even 14…Bc5. The knight-“sac” isn’t really a sac because in sharp positions each tempo matters.

14. b4!

Wisely, tipau stops any ideas of Bc5 once and for all. White wants to suffocate black’s bishops with his pawns, and take territory for the queen. Unfortunately it would seem that this plan has its drawbacks as future e5 will allow Ng4 once again…

14. … Bb7

Seems that black is quicker here. The bishop looks very strong and immediate threat is 15…fxe4 and black penetrates white’s camp. Only chance for white is to close up the diagonal with d4-d5, so it would seem that white did waste a tempo with 10.d3… I was satisfied with my position, and tipau was running out of time as well. Sacrifices were in the air…

15. d4

If only white could have played d5 now… But it’s black to move.

15. … Ng4

I didn’t give much thought on this sacrifice as i thought that 16.exf5 would leave white’s position shattered. But the threats of N(any)e3 and Nxh2 cannot be ignored lightly.

16. exf5

The best chance i think, but white’s position is already very difficult.

16. … Bg2+ 17. Kf2

Now black needs to castle and eventually sac the bishop on f4, but white still has a few defensive resources left.

17. … 0-0 18.Qe2

Not sure I like this move. White tries to activate his queen, but it only allows black to regroup his forced better. On the other hand this move stops the idea of 17…Rae8 and 18…e5. It’s already hard to suggest anything better for white.

18. … Bf1 19. Qd1 Ne3

Black reactivates the knight that has been idle for the most part of the game.


I felt relieved at this move because I saw that white was losing his queen without compensation. I was afraid of 20.Qe1 but black had good moves with 20…Ng2 and 21…Bxf4. Also 20.Qd3 was interesting, but then prosaic 20…Nd1! double-attacking the king and c3 won.

20… Nd1+ 21. Ke1 Bxf4 22. Bf4 e5, and White resigned, 0-1

Strangely white king ended up on e1 with the “caring” knight and bishop next to him. I thought this game was a good example of modern atomic strategy, and hopefully the reader was able to learn something from these notes. I would like to thank my great opponent tipau for providing me the chance to play this game, and I hope I didn’t put him in a bad light here… I can say that he has lots of good wins against me which are even better examples of how atomic is supposed to be played at the highest level.

I hope that I didn’t make any huge mistakes in my analysis which was pretty light to say the least, but I’ve tried to stay clear of definite evaluations because basically, I’m not good enough to provide them…

The Beauty of 1 0 Atomic Chess

siggemannen (2503) vs Chronatog (2104)
Annotated by Chronatog – September 13, 2006
Originally written for Leonard Blackburn’s website

For most people, part of what they consider mastery of something includes the spatial value of time itself. You don’t expect a great chef to labor ten hours on making a simple tomato soup. And yet the only time you ever hear people whining about greatness taking time is with chess. Go figure. After thousands of games and obscene amounts of time spent on studying chess, you’d expect that people could reel off moves much faster than they do.

But how do I illustrate the beauty of 1 0 atomic chess? When siggemannen and I finished playing a series of 1 0 games, I knew I had a candidate in one of the games, the only draw in our series that day. And it is that game that I present to you today.

1.d4 d5 2.Nh3 Bg4 3.f3 Bxh3 4.g3 e5 5.h4 e4 
6.e3 f5 7.Na3 b5 8.c4 Bxa3 9.c5 c6 10.g4 g6 
11.g5 h6 12.h5 hxg5 13.f4 Nf6 14.hxg6 Qh4+ 15.Rxh4 Rh1 
16.Kd2 Rh2+ 17.Bg2 Na6 18.b4 O-O 19.a4 Nxb4 20.axb5 Rg8 
21.Rb1 Rh1 22.Qf1 Rxf1 23.Rb8+ Kd7 24.Rd8+ Ke7 25.Rd7+ Ke8 
26.Rd8+ Ke7 27.Rxg8 a5 28.Ba3 a4 29.Bb4 a3 30.Bxa3 Kf7 {draw agreed}

Bear in mind that this was played at 1 0 time controls. I’m fairly sure that some of the moves won’t hold up under exhaustive analysis but in order for a theoretically drawn game to be won, somebody has to blunder. 1 0 just speeds this process up by adding limited time to the equation.

1.d4! d5

This was the first time siggemannen played 1.d4 versus me. I prefer to play systems that open lines up so I played the very natural looking d5, hoping that it didn’t lose outright like 1.e4 e5 does.

2. Nh3 Bg4 3. f3 Bxh3 4. g3

After 2. Nh3 was played, I immediately wanted to shut down any potential for his Queen to cause havoc by the very simple delayed exchange of my Bishop for his Knight. 4. g3 was a surprise though!

4. … e5 5. h4 e4 6. e3 f5 7. Na3 b5

I began grabbing space with e5 and after h4 was played, forced 6.e3 by siggemannen to close in his lines, otherwise, 7. … e3 would completely block him in with a huge lead in development and available tactics for me. Following the same principles, 7. … b5 shuts down Na3 and grabs more open space for me.

8.c4 Bxa3 9.c5 c6 10.g4 g6 11.g5 h6

8. … Bxa3 because c4 prepared c5 to close in my Bishop and I thought it best to swap it off for his other Knight. Then 10. g4 was a probing move, seeing if I’d overreact by playing Qc7, which wastes tempo and allows him to seal off the dark squares with f4. I thought g6 was best, defending against any Queen hijinks while not allowing him to completely shut off my lines. g5 was played to shut off my black diagonal for the queen, so I played h6 to get rid of the pawn.

12.h5 hxg5 13.f4 Nf6 14.hxg6 Qh4+ 15.Rxh4 Rh1 16.Kd2 Rh2+ 17.Bg2 Na6

I decided to play a gambit with 13. … Nf6, offering him pawn and knight for his pawn and opening up my lines – a further sacrifice becomes necessary. 14. … Qh4+ sacrifices the Queen and prepares my queenside castling to transfer over my other rook. I’m sure siggemannen could have traded off his Queen for my Rook and won easily with a simple advantage. Unfortunately, he didn’t see what I was threatening as it was 1 0. This is the beauty of 1 0!

18.b4 O-O-O 19.a4 Nxb4 20.axb5 Rg8!

siggemannen elected to shut off my knight’s space with b4, allowing me to castle long, unfortunately, I have to sacrifice yet another piece. 19. … Nxb4 so if he fails to recapture, I can play b4 and my king will be safe temporarily as he cannot transfer his pieces over to the open black diagonal fast enough to defeat the rooks on the other side of the board.

20. … Rg8 21.Rb1 Rh1!

siggemannen completely missed Rh1! which saved the game for me. Rb8+ loses due to Rxd1 atomic-explosion. The Queen is now lost, as I threaten atomic explosion. Both Qg1 and Qf1 produce the same result, however, it is important that you capture with the correct rook, or you will have thrown away the game.

22.Qf1 Rxf1 23.Rb8+ Kd7 24.Rd8+ Ke7 25.Rd7+ Ke8 26.Rd8+ Ke7 27.Rxg8 a5

Now the game is drawn after the rooks are exchanged off. My pawns are blockading siggemannen’s pawns and none of them are the color his bishop is, therefore no progress can be made with the game due to this pawnization. Even if I didn’t have the extra rook pawn, this position would still be a draw. And we had a total time of approximately 10 seconds left for each player remaining.

28.Ba3 a4 29.Bb4 a3 30.Bxa3 Kf7, draw agreed, 1/2-1/2

And this fine 1 0 game concludes in a gentlemanly fashion, an agreed draw. I’m completely sure that siggemannen could have won rather easily if he had traded his queen off immediately for my rook, as he would have had enough time to play b4 in addition. But that’s the beauty of 1 0. It allows games to reach superb finishes like this one. That’s the beauty of 1 0.

siggemannen annotated another game for Rekursiv’s site, which is again in pgn format. You can download that file here. That game was played September 30, 2007 between HentyReader and SPogoreckiBlind (a blindfolded account for SPogorecki).

Further annotated games will be presented as their own posts – these were samples of annotated atomic chess games to see how two different players annotate a game, and all three examples were played and analyzed without the assistance of computer engines.

Other Posts