Upgrade to Chess.com Premium!

Defending Against Gambits

  • GM BryanSmith
  • | Oct 18, 2012

The game begins, and within a few moves your opponent pitches a pawn. It’s not considered the most correct gambit, and might even fall into the category of “cheesy”. But by taking the pawn, you are playing “his” game. You have a pawn, but you have to defend, for a while. Your dreams of a crushing attack concluded with sparkling sacrifices have to be delayed while you consolidate your position. You worry that you won’t be able to contain events and your scrappy opponent will land a decisive blow. What do you do?

Defending against an opening gambit is not a simple matter. You have to play solidly, but at the same time avoid playing too passively and allowing long-term compensation. You have to feel the right moment to return the pawn or go onto the counterattack. The extra pawn can skew your positional ‘barometer’. It confuses your sense of who stands better, and how big of an advantage is enough.

There are thousands of different opening gambits – from the subtle pawn sacrifice introduced by a top grandmaster on move 18 to the brutal Englund Gambit, introduced as early as move one. There are gambits for development, for positional gains, to weaken the opponent’s position, to institute a blockade, to gain control of the center. But today we are going to focus on a typical kind of gambit that an amateur player might meet - where one side sacrifices a pawn very early for open lines and one or two tempi of development. We will look at the Blackmar-Diemer gambit; but the concepts are applicable to many other such gambits - the Englund gambit (1.d4 e5), the Goring/Danish gambit (1.e4 e5 2.d4 exd4 3.c3), the Evans Gambit, and the Smith-Morra gambit. The last two are a class above the previous ones, and deserve the utmost respect. Of course there are many others.

Let’s say your opponent begins with the moves 1.d4 d5 2.e4.

What do you do? If you normally play the French or Caro-Kann against 1.e4, you can “chicken out” by playing 2…e6 or 2…c6. Then you are back to your familiar opening. Of course, you don’t get an extra pawn and – objectively speaking – an advantage as black as early as move two. You also don’t prevent your opponent from later offering another gambit or making another unsound sacrifice.

So the first step is to actually take the pawn. It is the principled thing to do, and by not taking the pawn you are already defeated psychologically in some way. Now, it is a good idea to have some idea about a good continuation after taking the pawn. The Blackmar-Diemer gambit isn’t good, but it still requires some respect. So – while you certainly shouldn’t spend too much of your life preparing a response to an opening you rarely see – it would be a good idea to search in an opening database or in books to see how good players respond to it. There are several good systems, just pick the one that you like.

Black can play 1.d4 d5 2.e4 dxe4 3.Nc3 Nf6 4.f3 exf3 5.Nxf3 Bf5, which is probably the most popular. Here is an example:

Also quite enough for Black to claim an advantage is 5…Bg4 6.h3 Bxf3, followed by playing like in the “Fort Knox” line of the French (…e6, …c6, …Nbd7, ….Be7, etc) except with an extra pawn.

A surprisingly strong system is just basic development with 5…e6, 6…Be7, and 7…0-0. Black blocks in the light squared bishop temporarily, but avoids allowing White to gain time by attacking it. This system is probably neglected because it looks a little passive. True, if Black sits around then White will build up his standard kingside attack – which is the reason he plays the Blackmar-Diemer. So the crucial break for Black is …c5, which he can do quite early, as in the following game:

White pretty much threw everything he had at Black in this game. Did Fedorchuk face any real troubles? Sure, he had to calculate a few variations, but that can happen when playing chess.

One of the keys to successfully playing against these kinds of gambits is to know what kind of attack is really dangerous to you and what is just your opponent's suicidal final flailing. Notice that in the above game, after Fedorchuk defeated the attack, there was no long endgame exploitation of the extra pawn. This is because the missing pawn induced White to make even more sacrifices. When the smoke cleared Fedorchuk was actually up three pawns. So, allowing your opponent to attack a bit can work out well for you. 

So what is the key to playing against these kind of borderline-sound gambits? There is no magic bullet, and you can find examples of even quite strong players going down to weaker players in the Cheesy Gambits. The main thing is research. Just as in other openings, you need to study a little bit - it's that simple. When you have some idea what to do, a reasonable scheme in mind, you have much more confidence, and you can avoid falling into your opponent's typical plans. One of the reasons that some higher-rated players have lost to these gambits occasionally is that they simply paid them no attention whatsoever. 

Besides having done some basic research, it is important to understand the psychological pitfalls of playing against a gambit. The extra pawn makes it hard to objectively evaluate the position. In a normal position, the side who has the initiative is usually preferred. But in these gambits, the gambiteer always has some kind of initiative early on - the question is whether it is enough. This makes evaluation more difficult.

The number one error is probably cowardice. Your opponent sacrificed a pawn - he gets a temporary "high" from that. Nevertheless, that doesn't necessarily make him brave - he was simply reckless enough to play something that other people invented and which does not have a good reputation. So it is important that you do not fear ghosts. Trying at all costs to avoid even the hint of an attack is one sure way to face a real attack a little later.


  • 21 months ago


    While I think the claims of a definite advantage for Black are overstating things, I think the general point that Bryan makes is fair, as theoretically the onus is on White to prove enough compensation (i.e. equality) and so Black is better off after taking on f3 than in the original starting position.  Of course that doesn't stop it from being a fully viable weapon at club level- I have certainly played it a fair number of times with success.

    The other common recommendation against gambits is to "pragmatically" decline them and get gambiteers out of their comfort zones (e.g. 1.e4 e5 2.d4 exd4 3.c3 d5, or 1.d4 d5 2.e4 dxe4 3.Nc3 e5, and hoping to lure White into a level endgame).  That approach also works for some people, but personally, I prefer Bryan's recommended approach- it involves more risk, but if you know what you're doing then you get much better winning chances, especially against gambits that are only borderline-sound.  The positions in "accepted" lines of gambits are usually unbalanced and offer large scope to outplay the opponent.

  • 21 months ago


    I disagree with you saying that the Blackmar Diemer is weak. There is no advantage to be found for the defender. The two examples you give are not very serious, since white is rated 600 point below black - blakc would win in any opening ! Please refer to blackmardiemergambit.blogspot.com to convince yourself that white is doing great !

  • 21 months ago


    If you look at the games of Diemer you get the impression, that his opponents were simply confused by his crude attacking style. Defensive technique is normaly better developed today. His opponents threw away winning positions quite regularly. I studied the BDG for a while and play it sometimes in Blitz against 1.e4. d4 2. d5 the Scandinavian defense as a surprise weapon. It works sometimes quite well and it is definitely funny to play - but to give away a central pawn like this for nothing is not healthy. Diemer himself played it one time against a really strong opponent - Bogoljubov - and lost. Bogo himself played it after this game one time - and won - maybe its a gambit to have fun against weaker opponents Wink

Back to Top

Post your reply: