When is a brilliant move not a brilliant move?

"That's a hell of a move, Alisdair". My opponent had visibly wilted as he spoke. I tried not to react; remain calm, keep focused. For I was not necessarily evaluating the position the same way. My move had had the desired effect: it was a psychological dagger to my opponent's heart, with only a few minutes left on his clock. I had made it hoping it was the double-exclam move, but objectively I thought I could see a way out. Before I moved, this was the position:

Can you find the move for Black? I didn't. But I couldn't resist 25...Rc3?? This, in fact, blunders away my advantage. Lucky for me, it caused my opponent the necessary problems. White can't capture the rook - if 26.Bxc3??, I win with 26...Qxe3+ 27.Kh1 and then Philidor's Legacy to finish it (if you don't know what that it, allow one of my students to demonstrate later). Meanwhile, since the Queen is under attack, and the Teflon rook prevents her from defending e3, it looks like White is in trouble. So much so that Iain (my opponent) actually resigned here.

What I was looking at was 26.Bd4, which just about holds the balance (slight advantage Black). But Iain and I both missed the critical move: 26.Qb2! He can actually afford to allow 26...Qxe3+, because with the Queen having retreated to the 2nd rank, he has 27.Kf1 available! I'm now losing the knight as I need to save my rook AND prevent a back rank mate. Returning readers will notice that this is the climax to one of the games I mentioned in last week's post. What I didn't know then is that Iain has his own blog, and he covered the whole game. You can check out his blog here.

Instead, what I should have played was 25...Nc3! This leaves me slightly ahead after 26.Re1 Nb5 27.Qb2 f6 28.a4 fxe5 29.axb5 Qb6. I had overlooked this, as I didn't see Qxe3+ as a threat if I didn't have the knight on e4 to threaten checkmate with. It turns out to be plenty threatening. In my hubristic time-pressured nervous calculation, what I'd hoped I'd uncorked with Rxc3 was a plachutta - the rarest of all tactics (so rare that my spell check has no idea what that word is!) - what I'd actually played was a common-or-garden interference, and not even a good one.

So far, there are no brilliant moves to be seen. Which is kind-of my point: sadly, at my level, moves like Rxc3 are more common than the "!!" moves - those which you think (or hope) at the time are going to earn the hallowed annotation - only to look at it later and realise it is objectively unsound. So here's an actual brilliant move, courtesy of one of my students. He was losing, but saved the game with a nice idea:

I'd only been working with him a couple of weeks at that point! Not that I really should take any credit. When he sent it to me, he said "you'll see it immediately" - and I did, but that doesn't take away from how pleased with himself he's entitled to be. Just because this particular brilliancy will be obvious to most experienced players, it doesn't take away the panache. In other words, I think my student's move met the criteria for a brilliant move - it was a piece sacrifice that led to a forced line that is favourably outcome-altering compared to any other legal move he had.

When I was growing up, one of the first chess books I got to read was Winning Chess Brilliancies by GM Yasser Seirawan. For a young aspiring player just starting out in the game, the concept of brilliant moves was encapsulating - something to aspire to. Maybe, one day, I would play a brilliant move. But it crossed my mind that my competence might never rise to the ambition. Brilliant moves were rare, requiring a stroke of genius to find in one-off games. Fast forward to 2024, and chess.com is handing them out like confetti. At least, that's something I'd lamented to another student prior to him finishing this game in style:

"Please don't take my exclamation marks away!" he joked. I don't think anyone would begrudge him two "brilliant"s here - firstly for offering up the Queen and then forcibly sacrificing her to complete Philidor's Legacy. (Lucky him; he actually got to put it on the board). Despite the fact that this is a known position and so the combination is arguably easier to find than when it was first played, I don't think my student is any less deserving of credit, particularly when he had to use a pseudo Queen sacrifice to deflect a rook off e2 to engineer the whole thing in the first place. He would like me to point out that he is a lowly 1400 and he found this plan! But then, earlier in the same game...

It is worth noting my student's opponent had a brilliancy too - although in his case, it was a matter of finding the only way out. I wouldn't have said sacrificing the Queen in itself merits the "!!" since they can win it straight back with a knight fork - but combined with it being the only move that doesn't lose, yes, it's brilliant. So now I've gone too much the other way; into the realms of moves that are worthy. So let's return to another position from last week's post, if I present it as a puzzle:

I mean, thank you for the double exclam, chess.com, but I don't think it was brilliant. The game is already decisively winning for me, and I'm not sacrificing the rook in the traditional sense, since I'll win the Queen to a tactic. If you didn't solve this puzzle first time, it's probably because there are a handful of moves that are winning. This I think is the cleanest, because had my opponent not resigned, after 29...Nxe5 30.Kxe5, I'm winning even more material with 30...c2+ Compare that to a few moves previously:

So there, I would maintain, brilliant moves are still possible (even to us mere mortals), even if the bar were to be a little higher. There's an article by Jeremy Kane that explains the basics of how Game review evaluates what's brilliant. If chess.com is indeed "giving out brilliant annotations like confetti", to paraphrase my former self, what changes would I make? This is hard to explain because the general points that Jeremy Kane describes are a good start. Here is my take:

1. If it is objectively the best move, and the computer doesn't understand this, the move is brilliant. Period. No other criterion needs met.

This is obviously rare, but occasionally the engine you are using for analysis doesn't appreciate the true value of a move until after it has been played. The move is made, and with best play from both sides, the evaluation goes up and up - better than what the original mover could have achieved with any of the alternative moves. When I was explaining to my wife what the various annotations in analysis meant, I did describe "brilliant" as the engine saying: "f*****g h***, not even I saw that!" - which is of course inaccurate, seeing as the engine will find 99% of brilliancies without the electronic equivalent of breaking a sweat. But a good way of describing to a total beginner (sorry dear). Since I (obviously) don't have any examples of my own, here's an example of Aman Hambleton out-analysing the engine:

In this famous example, the Game Review awarded GM Hambleton an inaccuracy and a mistake because it couldn't understand his plan - or analyse it through to the conclusion that Hambleton (rightly) reached. Between Nd3, exd3 and Qxe1, Hambleton deserves at least four exclamation marks - distribute them how you like; you have to describe at least one of these moves as brilliant.

2. It's not a sacrifice if the player is recovering the material IMMEDIATELY

This is one reason why I say 29...Rxe5 in my example above was NOT brilliant. I was immediately winning more material than I originally "sacrificed", through a simple tactic. Give me an exclamation mark, sure, but not two. Since I so far seem to be bringing in games I've used before, here's a new one, recently finished:

3. The evaluation of the position should have a significant dependence on whether or not the "brilliant" move is found.

I don't mean it has to be the only move that is winning, or that any other move would probably lead to a different outcome. Although that, of course, would meet the condition. But finding the move should award the player a significant improvement on any position they might otherwise have had. See this example:

4. The more legal moves the opponent has, when none of them help the opponent, the more likely the move under consideration is likely to be brilliant.

Forcing combinations are all very well, and they give reassurance to the committed calculator. However, the more romantic brilliancies are the one where the player must rely on their assessment of the position - that the scene that transpires must pay off no matter what the opponent does. I would say the case above works well here - my move is not check; there are a number of legal moves my opponent could make - although obviously, most of them are bad. But that's the nub of it not being a forcing line - there is greater jeopardy - more of a chance that your opponent has a resource you haven't seen.

5. It doesn't HAVE to be a sacrifice; it could just be a counterintuitive or otherwise hard-to-find move.

Endgame studies are riddled with such moves. Here's one that comes from Neustadtl in 1890:

Why am I setting out my stall for what I think makes a move brilliant? I don't think it does anything for the good of the game to dilute the meaning. When I was starting out in chess, a brilliant move was something a novice wondered if they'd achieve in their lifetime; now you can find one every other week. I saw one forum this week where a learning player had played a brilliant move by accident! If they'd seen that they were hanging a rook, thus making the move "brilliant", they wouldn't have gone for it - and lost the game. I think they lost the game anyway, but not because of that. The possibility of maybe making a brilliant move one day could be used as motivation, giving a player a drive to have a deeper understanding of the game.

Here is one more case where the exclamation marks might be taken away:

White has just played 11.Bb5!! This discovers an attack on the Queen without threatening to "win" another piece. In fact, White has declined to play the obvious 11.Nxa8, recovering the material while saving the knight, which was under attack. Oh, and there's nothing guarding the h1 rook, so Black could take that next move. He shouldn't, though. This is a brilliant move by White if ever I did see one. If you had the Black pieces here, and you hadn't seen this position before, you might just align yourself with Iain's comments at the top of this post when I played it against you.

BUT... What if I told you that this move is straight out of my opening repertoire? Which it is. If we reach this position, I'll probably blitz out my 11th move. I didn't even find 11.Bb5 - Stockfish did. I'm not saying I wouldn't have found it if I'd adopted this line in the olden days of having to study positions by pushing pieces around a physical board for ages, but I didn't have to. I followed this line down the rabbit hole, and Stockfish was ready with the answer. Would I really deserve a "!!" for playing out a move that I'd pre-prepared? If the answer is "no", this suggests a psychological element to the award of exclamation marks. They are earned partly through the spontaneity and invention of the player making it. It's an odd one though, because the cold objective quality of the move doesn't change. I've read the claim (I am not sure I believe it) that the threshold for lower rated players on chess.com is lower when awarding brilliants. I hope this is not true. The same move might be awarded exclamation marks or not depending on what their rating is - and a weaker player might be awarded a "!!" where the stronger player might not. Again, where is the incentive to improve?

Just to be clear: I'm not here to take away anybody's "brilliant" annotations. That's not for me to decide. All I'm saying is it probably wouldn't hurt us to have to work for our meal. In my blog a couple of weeks ago, I mentioned my game in Round 2 of the Scottish Chess Championships. I told you that I hadn't checked with the computer, but I believe I didn't find the right idea, and that my opponent sacrificed a rook. Well, guess what? My sac-and-attack was working fine; my opponent's rook sacrifice was unsound. But I didn't believe it and fell apart at the critical point.

So, it was game on, until we reached this position:

What lessons do I learn from this? At my level, most sacrifices are unsound. The ones that aren't might not be followed up on correctly. Calculation is key - which is a shame, because I'm not great at it, and when I am, I eat up too much time. But hey-ho. Let's finish off with a miniature that actually had some brilliant moves: