The Philosophy of Gambit Play

The Philosophy of Gambit Play


To play a gambit is to sacrifice material, usually a pawn, right in the opening. This is done to get something in return, what is known as your compensation

The concept of compensation is essential in understanding gambit play. The most usual form of compensation is faster development, leading to a potential attack. For example, this is the case in the Danish Gambit:

However, there are other forms of compensation. For example, consider the Cochrane gambit:

As we saw, here White's compensation consists instead in the weakened Black king and the ability to create a strong central pawn roller with d4-e4-f4.

In general, a good principle to keep in mind is:

To understand a gambit, you have to know what the compensation consists in!

Thus, to know how to play a gambit you need to know what your compensation consists in and how to exploit it. To know how to play against a gambit you need to know what their compensation consists in and how to negate it.

In this article I'll discuss gambit play both from an abstract perspective, but also more concretely by discussing different forms of compensation.

Before we get down to business, if you want to see what current day amateur gambit play looks like and how dangerous and fun it can be at fast time controls there is no better way than to check out my friend UAArtur's blog which contains many of his best games. For example, check out his:

Best Game of 2018

Best Game of January

Best Game of February


It's always fun to take a quick glance at the available literature. This time we'll look at general gambit books (there are many more on specific gambits).

First up are Nigel Davies's "Gambiteer I" and "Gambiteer II". The former suggested e4 with the Danish Gambit and Wing Gambit against the Sicilian and French, whereas the latter proposed the Schliemann against Ruy Lopez and Albin Counter-Gambit against the Queen's Gambit:


Boris Alterman has written three "Alterman Gambit Guide"-s, one for White and two for Black. The White one is a mixture of a quick overview of gambit lines, a teaching tool for developing players about development and tempi, and a repertoire book. It covers: Danish, Urusov, d4 against Philidor, Cochrane Gambit, Morphy Attack & Max Lange Attack, Evans, Panov Attack, Smith-Morra, and Milner-Barry. See Kenilworthian's review of it here.


The Black books are more like theoretical manuals. Black 1 covers: Benko, Blumenfeld, Vaganian Gambit (the Benoni, not the Trompowsky one), gambits against D-Pawn Systems and the English Defense Gambit while Black 2 covers: Marshall Attack, Hector's Gambit, Traxler, Frankenstein-Dracula, Falkbeer, and From's Gambit.


Finally, there's Yuri Razuvaev's "Key Concepts of Gambit Play", updated by Jacob Aagaard. While the previous two sets of books are more beginner- or intermediate-oriented, this book is different in looking at gambit lines played at the GM level. It discusses quite a different set of openings that you might expect.


Perhaps the best material on gambits has for years been published by Stefan Bücker and others in the German magazine Kaissiber.


There are also several anti-Gambit books available:





Finally, you could take a look at this website: Chess Gambits.

Why Gambits Work I: Sound Gambits

A gambit is sound when its compensation is objectively sufficient. What this means is that even with best play (correspondence play, computer analysis etc.) neither side is better or worse than usual. In other words, when you play a gambit with White it is sound when you retain at least your usual advantage. When you play a gambit with Black it is sound when you have no more than your usual disadvantage.

As a first example, consider the Two Knights Defence which pretty much everybody agrees is sound (notice that even though Black parts with a pawn, it's not usually even called a gambit...):

(See also my post on Two Knights, Ulvestad Variation which is not fully sound, but very dangerous for the unprepared White player)

As a second example, again from the Black side, consider the Marshall Attack which everybody agrees is sound (again, it's not usually even called a gambit...):

(See also my post on the Marshall Attack, Steiner Variation which is not fully sound, but, again, very dangerous for the unprepared White player)

Some gambits work simply because they're sound. However, most gambits are not sound. In fact, some are really bad and *shouldn't* ever work...

Interlude: Bad Gambits

Some gambits are basically just traps which work only if your opponent doesn't know the idea, or at very fast time controls where the point is to make the opponent lose the maximum amount of time. Out of those probably the most frequently played one is the Englund Gambit which is seen quite often in bullet. Here Black is playing for a crude trap without no real compensation and its best to know the refutation by heart:

There are some others. One of them, Reti: Tennison Gambit, will get its own post very soon.

Why Gambits Work II: Good Gambits

A gambit is good when it is not-fully-sound, but works well in practical play. Some of the reasons why any gambits work are the same as in the case of non-gambit sidelines or rare openings:

  • Your opponent won't know the frequently unintuitive best line and plays something intuitive, but 2nd-rate
  • You have more experience with the ensuing types of positions than your opponent
  • There is a psychologically unsettling effect in having to face the unfamiliar

    However, good gambits work for further sorts of gambit-specific reasons:

    • A good gambit puts your opponent under some sort of pressure, forcing them face concrete threats (White gambits) or to defend when they were expecting to slowly build or attack (Black gambits)
    • Best gambits feature an unintuitive & narrow path to equality/advantage. What this means is that even if there is a theoretical refutation, it's an unintuitive one and the path to it is narrow.
    • The unintuitiveness means that the relevant moves are a) harder to remember, b) hard or impossible to find over the board
    • The narrowness means that only specific moves will do, otherwise your opponent will be at a considerable disadvantage (against White gambits) or at the very least allow equality (against Black gambits)

    It should be obvious that all such effects are multiplied at shorter time controls which is why playing a Gambit repertoire with White and Black makes a lot of sense for Blitz and Bullet.

    There are plenty of good White gambits. For example, in e4-e5: the King's Gambit, Evans Gambit, Urusov Gambit; in the French: Milner-Barry, Alekhine-Chatard; in Sicilian: Smith-Morra, Wing Gambit, in d4-d5: Blackmar-Diemer Gambit, against the Dutch, Staunton Gambit etc.

    Here I want to look a little further at three black gambits, again with the idea of ascertaining the form of compensation.

    Icelandic Gambit

    If you want to know more then go on Chessable:

    ... and check out my Icelandic Gambit E-Book!

    Albin Counter-Gambit

    See more here:

    Von Hennig-Schara Gambit

    Again, there are many more. For example, Ruy Lopez: Jaenisch/Schliemann, Four Knights: Rubinstein's Gambit; Budapest Gambit, Benko Gambit; Bird's Opening: From's Gambit etc.

    To Accept or Decline?

    Let me conclude by discussing another aspect of gambit play, namely the choice between accepting and declining.

    We've all heard the adage that the refutation of a gambit lies in its acceptance. This might be correct from a "theoretical" point of view, but many gambits are simply too dangerous to be accepted in practical play and thus many people prefer to decline them. For example, I think most e5-players decline the Danish and the general recommendation is to decline the Urusov as well. Most Sicilian players probably decline the Smith-Morra. Many players decline the Blackmar-Diemer etc.

    The best gambits are sound or good gambits that can't be declined without making serious concessions. Out of those discussed above:

    • Cochrane Gambit - Nxf7 forks the Queen and Rook and so Black has to take it
    • Icelandic Gambit - d4 just leads to Exchange French: Monte Carlo after exd5 =
    • Albin Counter-Gambit - all ways of declining it leave one worse or equal
    • von Hennig-Schara Gambit - once one enters the line where it is possible it can't be declined at all

    Curiously, many of these gambits are black and I think Black gambits are in general underestimated from the practical point of view. More about Black and gambits in the coming weeks...