I was alerted to the potential of this variation when looking for something negative to play in next season's Aragón Championships, in which, as the minimum entry requirement is an Elo of 2100, I am liable to find myself seriously outgunned.
It's not the first time I've taken an interest in the Queen's Indian, though previously I've been more attracted to the lines involving 4...Ba6: but they appear to be a bit too lively for my present purposes. It was with no small pleasure, then, that reading Yrjölä and Tella's Queen's Indian (Gambit, 2003) I came across the following comment:
8. Nxe4 Bxe4 is a totally drawing variation. There are more than a thousand games with this move [sic] in our database but during the last ten years there have only been a few games that have lasted more than 15 moves.Go on, you've got my attention...
There is nothing White can do to change the drawing nature of this line.This is what we want! So let's look at the stats.
Using our usual criteria and searching Chessbase Online Database for games between opponents rated 2400 and above, we find 73 games, of which 68 were drawn, the remaining five being split with three wins for Black and only two for White. This is a draw percentage of 93.15, only marginally behind the 93.28 achieved in the Cozio variation of the Petroff. Winners with Black include Colin Crouch, Keith Arkell failing where Tom Chivers succeeded: those drawing with White include not only Ivan Farago - six times, plus another one on the Black side - but Bellón, who obviously isn't too proud to have a nine-move draw with the White pieces provided the opponent (Evgeny Ermenkov) is the one to make the offer.
Oddly, the search results only include games going back to 1998. It's possible that Elo grades weren't available - or just weren't entered - for games before this date, as I imagine that (to take one example of many) both Lajos Portisch and Anatoly Karpov were some distance above that figure when they treated themnselves to a seventeen-move draw at Skara in 1980. (They'd managed twenty-five in Tilburg earlier the same year, perhaps the more aggressive positioning of the rook on f1 having encouraged White to continue.)
Not bad seeing as the queens are still on the board and there's no open files. Still, 8.Nxe4 isn't compulsory - Black also has to be aware of 8.Qc2 and 8.Bd2. I notice that discussing the former move, the authors give three games in the same variation played between Andersson and Karpov in 1983 and 1984. They were drawn in 21, 25 and 17 moves respectively.