Semi-blindfold Chess

This blog will document my attempt to develop a moderate ability in blindfold chess. I have been playing chess regularly for the past two and a half years, mainly on the internet rather than at any particular club. In the only tournament I have entered (this after I had been playing around eighteen months) I managed a rating of about 1500. At time of writing on ICC I have a standard time control rating of about 1750 (ICC rating is of course quite inflated), and on Chess Tactics Server a rating of 1491 (peak of 1522).

I have always been interested in the idea of playing blindfold chess, partly because some (e.g. GM Christiansen) believe it teaches better visualisation which can help with calculation, and partly simply because it's a fun challenge (perhaps inspired also by the current coverage of the 2010 Amber tournament in Nice).

I have decided to work towards complete blindfold gradually, starting with semi-blindfold and gradually removing the pieces. To do this I play on ICC with the board layout set as all pieces and pawns invisible and with a chess board in front of me. On the board I have been using Othello tiles to represent all pieces and pawns (i.e. I can see whether the tile represents black or white, but nothing else - see image below).

In some ways this is a fairly modest handicap, as remembering which piece is which isn't too difficult. However seeing patterns is more of a challenge - a move which looks obviously good or bad at a glance on a normal board can be quite non-obvious in semi-blindfold. I created a new account on ICC named Semi-blindfold and played six games at 30 minutes per games + 15 seconds per move over the course of a few days. I hadn't known exactly what to expect, so I started by playing against opponents with a relatively lower rating.

In the course of the six games I won five and lost one, the loss being due to a mouse slip in an otherwise unclear position where I had felt slightly worse.

Above is an example showing the position as I could see it on the board in front of me (in this case before the mouse slip 24...Rb3??). I had meant to play Rb4... it seems a mouse slip is somewhat easier to do when the board on screen is blank. See below for the actual position.


So what was different from normal play? Certain moves seemed a lot more difficult to spot. For example in the position below I played 12.Qa4+?? to try to break the pin, overlooking the obvious reply 12...b4 (I had not seen that the d4 knight controlled b4), and should have instead played the necessary 12.Bxd4.

Another example of the problems I found with semi-blindfold occured in the position below.

Here I played Bh5??, intending Bxg6...f4 and picking up the e5 bishop. This overlooked the strength of an immediate f4. As it was Bh5 was rightly met by d6 (Bxg6, hxg6, f4, Be6). In fact, I didn't notice the threat of Bxb3 at all until he played the move - quite an unpleasent surprise.

Below is a sample of one of the semi-blindfold games. I'm not sure it's representative, but it's shorter and more interesting than most of the others. It also carries the lesson that just because you can make a check does not mean it is a good idea... though I haven't often seen it cause as much of a problem as it does in this game.

Play chess online

Broadly speaking these six games left me with the impression that I could manage semi-blindfold play at a level of perhaps 1400-1500 (standard ICC). At the end of these games my rating was 1552 but I think I was a bit too conservative in some of my choices of opponents. My play was fairly handicapped, but then I expect that to be the case with blindfold even if I do ever become reasonable at it. For that reason after these six games I have decided I am going to start removing pieces (that is, Othello tiles), starting with my king and queen. This shouldn't make it that much harder initially, but it's a start.