Sunday, September 09, 2007

Intuition

I have taken a closer look at my "chess intuition".
In the position below I dismissed the move 1. ... Bxe4 because I had the feeling "that can't be good".























Black to move and win.

My 3 points analysis:
  • Invasion squares: e4 and a2
  • Overworked piece: Nc3 defending e4, e2 and a2
  • Possible counter attacks by: Rg1 and Bd3
I tried to formulate where that feeling that 1. ... Bxe4 was bad stemmed from. This was approximately what I came up with:
Since the threat is stronger than the execution, I must not take away the threat of the bishop by trading it.

That is pretty vague, but we are talking about feelings here. When I started to analize the move, I could proof that the move 1.Bxe4 was bad indeed, but for a different reason!
The point is that after 1. ... Bxe4 2.Nxe4 Nxe4 3.Rxg7+! white has a decisive counter attack. For instance 3. ... Kxg7 4.Qe3 attacking h6 and pinning Ne4
This counterattack never crossed my mind when "intuitively" dismissing 1. ... Bxe4. If I take the same position as above but I put the rook at h1 instead of g1, the counterattack is ruled out. All of a sudden the move 1. ... Bxe4 becomes winning too!

This means that I'm just deceiving myself when I talk about "my chess intuition". At least in this position.
Painful yet instructive.

Maybe I must investigate the statement
When your intuition strikes up but you can't back it up with a narrative, you are probably wrong:)

12 comments:

  1. please put me down for a new playchess high. 1677.
    thanks,
    Jim

    ReplyDelete
  2. Regarding narratives, I think I do roughly the same thing with endgames. For example, a month ago my narrative script for knight endgames included Three rules. Now my narrative script includes Seven rules and I feel like I know them much better... though I'm far from mastery.

    The funny thing is, giving your script to someone else doesn't really transfer the knowledge. It may ease their acquisition of the knowledge, but only studying many concrete examples transfers it.

    I could tell a beginner that Knight Endings Are Pawn Endings and All Rook Endings Are Drawn and that "knowledge" would be silly and quite useless. Heck, I didn't really understand Botvinnik's Rule until this week.

    "Invasion is the key to most complex tactical positions". Great. Maybe some day I will really acquire that knowledge and find it useful. :)

    ReplyDelete
  3. LF,
    you touch an important point. The past months I was hesitant to publish my findings several times because I know I'm really on to something. But then I realized that my blog drowns between all the new chessblogs and that my writings are incomprehendsible for those who aren't prepared to put considerable time and effort into it. It is the process of acquiring knowledge that does the work and not the knowledge itself.

    ReplyDelete
  4. The good news is our posts are saved. Even if only one or two others are studying knight endgames now, at some point others will study them, and then my posts on that subject will be helpful... just like your pawn endings posts were useful to me and others months after you wrote them, and I'm sure your posts on complex tactical positions will be more helpful when I study complex tactics.

    ReplyDelete
  5. 1Nxe4 Nxe4 2. Rxe4 and winning

    maybe I am too optimistic

    but the rook cannot be pinned by Bd3 nor recaptured with the Nxe4. goodie

    ReplyDelete
  6. and rxBg6 does not work although best defense.

    ReplyDelete
  7. Montse,
    that's why this is such interesting postition. It adresses move order problems which are extremely common.
    Rybka gives in this position:
    1.Rxe4 = +2.86
    1.Nxe4 = -0.19
    1.Bxe4 = -1.21

    If you put the rook of white on the harmless h1 in stead of the active g1 square in this very same position, then Rybka gives:
    1.Rxe4 = +5.22
    1.Nxe4 = +5.91
    1.Bxe4 = +2.69

    So the element of counterattack by rook g1 is very important in how this position should be evaluated.

    I attempt to formulate this in a way that it will help in move order problems in general. But that is difficult.

    ReplyDelete
  8. I am not sure if I would of seen it without prompting but when you mentioned not considering e4 I could quickly see why it might work. That the h7 square is prevented by the knight as the only guard. In Ct-art there are a whole string of tactics where a rook or queen is sacked on the seventh rank on the f7 or h7 square and the other major piece delivers check on the square across from it either f7 or h7. normally with help from other pieces there is a mating net to follow. a chance for a major piece delivering check on the seventh rank is a clue that it should at least be investigated.

    ReplyDelete
  9. Here is an idea. Intuition is a vague pattern match to a tactic we have seen before. It reminds one of something else but done (and I hate this term )at a subconcious level.

    ReplyDelete
  10. An interesting post topic tempo, especially the part about intuition and the idea we read about in chess books that "threats are stronger than there execution". From my own chess experience, I have to say that I'd add a caveat to this guiding principle, which would be "threats are stronger than there execution if you're able to follow up with more threats." I've found that if I can't increase the pressure already existing, it usually pays to take advantage of what's there already.

    ReplyDelete
  11. Tempo,

    I have had a deeper look to this position. This time I used the computer to find out why 1...Rxe4 is stronger than 1... Nxe4. The initial sequences are the same. So what is the clue? White comes into a position that with 1. ...Nxe4 he does not have to recapture on the rook. It is a position that white can play and improve the settings of his pieces. The recapture would have resulted in a smashing loss. So what did i investigate. Recaptures, pins, counterstrikes with pieces, threats against black's king, and movements of the Q. What did I miss - King safety increasement which stalls balck's position. 4 Kb1 kills 1...Nxe4. King goes off the c-file,deadly long diagonal and defends a2.

    1..Nxe4 2.Nxe4 Rxe4 3.Rxg6! fxg6 4.Kb1!!! and not Nxe4??

    1..Rxe4 2.Nxe4 Nxe4 3.Rxg6 fxg6 or Nxc5 wins Why not a queen move on 3.Q-move ? reason the discovered check on the king is more dangerous as it opens the way for black to invade white's position. It gives black an extra tempo. It is a nice pin example which leads to an exploit of other weak squares a2 as some defenders are overloaded.

    ReplyDelete