Tuning

[Lyudmil Tsvetkov]

this is precisely the type of position for which the white king should receive some nice penalty, as the general pattern favours black very much, especially what concerns possible king attacks.

Why it favors black and not white? What if position is flipped? It would favor white instead?

Once detected the closed position how can SF understand if it favors white or black?

Black's pawn chain is pointing towards the kingside. White's chain towards the queenside. If Black castles short, there is no impact from white's chain. Contrary to black's chain, if White castles short.
So another idea might be to give a penalty for White castling short in these KID like positions. Or, if White already castled short, give a penalty to king safety.
Since this is partially redundant with pawns attacking kingRing, maybe that penalty should be increased in such positions.
Of course, that's all only valid if I correctly understood Lyudmil's idea.

Hi Joerg.

As usual, you understood my idea better than myself.

If you allow me, I think the best thing with such structures would be to give penalty for the king in the general case when it is on the side where the chain is pointed at, i.e. for king on e1,f1,g1,h1, regardless of whether it has castled or not.
I do not know about the penalty, maybe 20cps for a start would be OK, but who knows. What were Joona's values?

Hi Joerg and Lyudmil,

You could also try a progressively larger penalty that increases from e1 to f1 to g1/h1. Maybe 10 cp for e1, 15 cp for f1 and 20 cp for g1/h1.

Regards,
CL

I think it is a viable idea, Carl.

The problem is that now every patch that adds knowledge fails in SF. Only simplifications and contempt work.

The closed patch for a big number of blocked files Marco tried might have been necessary for SF to avoid too much closure of the position, but you see, you need to tune it, and tune it, and tune it...

I think the second idea, somehow related to closed positions, namely the KID-test wtih chains pointed at the enemy king, is the more interesting one though. SF very much needs this knowledge, but, again, in order to succeed, you have to tune it, and tune it, and tune it...

Interesting if someone in the future, probably Joerg or Ralph, will push such a test though. (the KID test )

Apart from that, you know, half of SF is search. And SF search parameters seem to be specifically tuned to handling open positions...

[Michel]

Only simplifications and contempt work.

If you are referring to Leonid's patch: this is not contempt but a clever search trick which makes SF prefer certain forced draws over others. It does not alter the root score (not in a systematic way at least).

No genuine contempt implementation has been shown to work against equal engines yet. It is highly unlikely that this is possible.

The jury is still out if Leonid's patch is really an improvement against equal engines (we will probably never know for sure) but at least it is (almost) theoretically correct.