White threatens 22.gxh4.

It is difficult imagine what he did to keep your <treasure>: he misplaced the queen, moved many times by bishop, forgot about the development ,weaked king's position, even he didn't want to exchange it for e6.

But each position must be judged on its merits rather than assessed on generalities. The idea behind Qe2-Qd3 is to protect the d5 pawn. If it works, white stays a whole pawn ahead. And that is worth a little effort and temporary discomfort, no?

According to opening explorer, the Qe2-Qd3 idea occurs in just three games in the database. Two draws and one black win. So that's not a great showing for the idea (admittedly from not a huge sample). But then let's take a look at the three white players who thought this was worth a try: Vlassov, we already know. Federov, right at the beginning of his career in 1974.

And the latest showing for this move? Ahem. Viktor Korchnoi in 2005 (a draw):

Korchnoi vs V Beim, 2005

all gold for Terekhin a reel it you in rope re strike at similar to

yesterdays puzzle a tag heuristic in tub thump queen ti ar 22.Kxh2

Perhaps the results — even granted the limited sample size — should be allowed to speak for themselves.

From my own experience, I know that in playing Slav lines as Black I have sometimes been very successful in taking and holding the pawn on c4, with seemingly minimal positional distortions required to achieve it, but some intangible factor in the resulting positions has almost always led me to a bad outcome.

Granted: computers are changing the way we think about chess, and they sometimes do succeed with moves that human players would find ugly and artificial. But principles exist for a reason, and one day, stronger programs may find reasons why the current ones are on the wrong track.

The other intangibles in the position are that black has already taken liberties by playing the scandinavian. He has given up a pawn presence in the centre and, in the "modern" variation, temporarily sacrificed a pawn which he may or may not get back.

If it is sound (and although this remains inconclusive, it seems probable), Black should be able to recover the pawn. Typically, the modern "best" way to play against such a gambit is to maneuver patiently and return the pawn as advantageously as possible. The worst thing to do, as several of our recent POTDs and GOTDs have illustrated, is to make heroic, positionally distorting (even if only "temporarily" so) attempts to retain the extra pawn. It is then that brilliancies ensue for the gambiteer.

My counsel is to accept it, and then return it at the proper moment to obtain a positional edge in exchange for it.

What I consider suspect is not taking the pawn, but making heroic efforts to hold onto it.

The distinction is what we see between modern variations of the Queen's Gambit Accepted (although, in fairness, they still crave but have not obtained full equality) and the attempts of yesteryear to keep a firm grop on the gambit pawn at any cost.

Of course, there is always room for differing understandings of this principle and how to apply it. In that diversity of interpretation lies much of what makes chess human and interesting, rather than a sawdust-dry recitation of "perfect" play.