Thursday, June 30, 2005

X-rake jogging

Let's have a look on an attacking piece, for instance the Bishop
A Bishop can reach every square on the board within two moves.
The question is, what potential targets are there within this reach?
Every hostile piece on the same color can be a possible target.
The next question is, what are the impediments between the Bishop and its target? And can those impediments be removed?

To see that quick you can imagine some sort of rake that comes out of the Bishop.
In fact this is an extended version of the X-ray scan from Mousetrapper.
Every Bishop has two outgoing rakes with a right angle in between (see both diagrams)

White to move and win.

I am experimenting with "X-rake jogging".
That goes as follows.
I adjust the thinkinglevel of Arena to 40 seconds per ply and let the computer play against itself.
If white is to move:
I imagine rake 1 for the white-squared Bishop of white and have a look at the targets within its reach.
I imagine rake 2 for the same Bishop.
Then I do the same for the black-squared Bishop of white.
When black is to move, I do the same for blacks Bishops.

At the moment I stil have to follow all the single paths from Bishop to target, which takes a lot of time.
Of course the final goal is that a whole rake pops out immediate when looking at a bishop.

At the same way you can look at the rooks, which is simpler. See a previous post of mine for an explanation of this phenomenon.



  1. Just look at the 32 light squares, they pop out as fast as you can imagine. Each of them can be reached by your bishop in 2 moves. So why you need rakes?

    Today I tried your 6-ply against my computer. I lost 4 times in a row and then drawed. All you must get is an endgame. In this stage a 6-ply deep search means complete blindness, the computer plays like a fool and you can win even when down in material. In my case the computer did not find a rook sacrifice and queening 6 full moves later and agreed 3 times repetition.

  2. Mouse,
    On the teeth of the rakes are your targets. I'm trying to relate the X-ray scan to the target scan.

    The 4 times you lost: weren't that simple tactics? The scanproces we are after has to prevent those. That's the test.

  3. Tempo,
    I am trying to filter the targets so that only the most promising pop out. Still working on this process, I'll post later...
    4 times lost, yes, to simpe tactics, but this emerged out of a worse position. I cannot play slower than say G15 against a computer because it plays so fast. Earlier you described your new strategy with a fast lane to endgame which is the perfect way against a computer.

  4. That was a really cool post... never really thought about it: "a bishop can reach any square in two moves."

  5. "BTW yesterday I realized all of a sudden that no one of the Knights knows that Margriet and I are married very happy for more than 20 years."

    8| what a surprise!

  6. I propose a new piece, the omega piece, one that can reach ANY square in one move. That is, that piece is one you can pick up, and put anywhere, except on the enemy king (well, and the opponent's omega piece). Wouldn't that be fucked up?

    It would probably help me to visualize the entire board.

  7. Blue,
    there is a form of chess where you sit next to another couple where you get their pieces when they trade and you can place the pieces on your own board whereever you want in stead of making a move.