[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [gnugo-devel] trevor_1_25.2
From: |
Trevor Morris |
Subject: |
Re: [gnugo-devel] trevor_1_25.2 |
Date: |
Fri, 08 Feb 2002 11:35:15 -0500 |
At 05:01 PM 2/8/2002 +0100, Arend Bayer wrote:
>
>> Items of particular note:
>> Mkpat generates forced anchors for Q & Y (at colors O & X).
>
>That's very nice, I will probably use this for the intrusion patterns
>as well; it will give both a speed up and resolve an inconsistency
>in my experimental influence code (I need to be able to associate the
>intrusion as originating from a well-defined O-stone). Can I do this
>with your implementation without forcing anchors for the rest of the
>database?
You can certainly force the anchor, yes. Now, I haven't tested
this with the DFA "-m" option, so I'm not sure what would happen
there. In fact, I've not tested it at all except w/ read_attack.db,
so beware, but I think it will work elsewhere.
>That sounds quite good. I wonder whether it could be most successful to
>combine the two approaches. It might be that direct attacking moves like
>liberties of the string are more quickly generated directly than via a
>pattern; also, one can then continue to use the move order heuristics in
>order_moves, which seem to be quite useful. This centralized handling of
>move ordering might be more efficient than writing intelligent
>value helpers as you have done it e.g. with RA000a.
>
>On the other hand, the special_attack/special_rescue functions more or less
>try to play the role of a pattern (their check for surrounding conditions
>really looks like pattern matching). By generating those special_... moves
>by patterns, one could
>- very easily do tuning of these moves (of course it is more comfortable
> to create a pattern than to write a new special_rescue11 routine)
>- give them intelligent values so that they sort into the other moves in
> a sensible order.
This all sounds reasonable. I'm still groping around in the dark with
this too. My biggest question now, is how to find backfilling moves.
I think patterns just may not be sufficient.