| View previous topic :: View next topic |
| Author |
Message |
Sven Schüle
Joined: 15 May 2008 Posts: 2242 Location: Berlin, Germany
|
Post subject: Re: KQKP and KRKP endgames Posted: Sun Jun 24, 2012 10:14 am |
|
|
| zamar wrote: |
I've done some experiments about perfect rule based classifiers.
Kqk, krk, kbk, knk are straightforward.
kpk is difficult, required about 15 rules.
KQQK, KQRK, KQBK, KQNK, KQPK, KRRK, KRBK, KRNK, KRPK are straightforward.
KBBK is relatively straightforward
KBNK is surpisingly difficult (around 10 rules)
KBPK, KNPK, KPP are too difficult
KQKQ is very difficult, likely too difficult
KQKR is doable, but there are some exceptional perpetual stalemate chases which are very difficult
KQKN, KQKB are straightforward
KQKP is very difficult, but I believe is doable (30 rules)
KRKR, KRKB, KRKN are very (maybe too) difficult
KRKP is very difficult but I believe is doable
KBKP complex, but should be doable
KNKP, KPKP are too difficult
There is not AFAIK complete theory about many of these ending, but if you want to build your own classifiers, simply generate all positions, check their status from tablebases and build your own rules with trial and error.
However I must warn you that the task is huge and ELO gain is going to be minimal. |
Let me first express my highest appreciation of your work on that topic! I have tried only a very tiny subset of that and can confirm that it can be really a huge task to write good classifiers for such a wide range of endgames as given in your list.
There is only a minor point I would like to comment on: from the list above which also contains several cases of mating the lone king, may I assume that your classifiers do more than just deciding about WDL (maybe even WD only), by also returning an evaluation that helps to actually play a "good" move, as in KBNK for instance?
As to KBNK specifically, I am surprised as well about the number of rules you mentioned since I think that a rule set like the one below were sufficient for mating the black king, with decreasing priorities:
1. no stalemate (can be skipped since it is usually detected by search)
2. wB and wN not hanging (can be skipped since it is usually detected by search)
3. maximize distance of bK to center
4. minimize distance of wK to bK
5. minimize distance of bK to a corner with same square color as wB
I have successfully applied it in KnockOut, using exactly rules 3-5 in that order.
Sven |
|
| Back to top |
|
 |
|
| Subject |
Author |
Date/Time |
KQKP and KRKP endgames |
Lucas Braesch |
Fri Jun 22, 2012 2:30 pm |
Re: KQKP and KRKP endgames |
Mincho Georgiev |
Fri Jun 22, 2012 4:10 pm |
Re: KQKP and KRKP endgames |
Sven Schüle |
Fri Jun 22, 2012 4:11 pm |
Re: KQKP and KRKP endgames |
Lucas Braesch |
Sat Jun 23, 2012 7:12 am |
Re: KQKP and KRKP endgames |
Kevin Hearn |
Sat Jun 23, 2012 8:45 am |
Re: KQKP and KRKP endgames |
Sven Schüle |
Sat Jun 23, 2012 12:08 pm |
Re: KQKP and KRKP endgames |
Sven Schüle |
Sat Jun 23, 2012 12:24 pm |
Re: KQKP and KRKP endgames |
Lucas Braesch |
Sat Jun 23, 2012 12:36 pm |
Re: KQKP and KRKP endgames |
Sven Schüle |
Sat Jun 23, 2012 3:19 pm |
Re: KQKP and KRKP endgames |
Ronald de Man |
Sat Jun 23, 2012 1:49 pm |
Re: KQKP and KRKP endgames |
Miguel A. Ballicora |
Sat Jun 23, 2012 2:45 pm |
Re: KQKP and KRKP endgames |
Ronald de Man |
Sat Jun 23, 2012 4:05 pm |
Re: KQKP and KRKP endgames |
Sven Schüle |
Sat Jun 23, 2012 4:19 pm |
Re: KQKP and KRKP endgames |
Ronald de Man |
Sat Jun 23, 2012 5:11 pm |
Re: KQKP and KRKP endgames |
Miguel A. Ballicora |
Sat Jun 23, 2012 5:37 pm |
Re: KQKP and KRKP endgames |
Michel Van den Bergh |
Sat Jun 23, 2012 6:09 pm |
Re: KQKP and KRKP endgames |
Joona Kiiski |
Sat Jun 23, 2012 6:16 pm |
Re: KQKP and KRKP endgames |
Daniel Shawul |
Sat Jun 23, 2012 6:27 pm |
Re: KQKP and KRKP endgames |
Tord Romstad |
Sun Jun 24, 2012 6:32 pm |
Re: KQKP and KRKP endgames |
Daniel Shawul |
Sun Jun 24, 2012 11:17 pm |
Re: KQKP and KRKP endgames |
Robert Hyatt |
Mon Jun 25, 2012 12:21 am |
Re: KQKP and KRKP endgames |
Ronald de Man |
Sat Jun 23, 2012 6:21 pm |
Re: KQKP and KRKP endgames |
Miguel A. Ballicora |
Sun Jun 24, 2012 7:19 pm |
Re: KQKP and KRKP endgames |
Robert Hyatt |
Mon Jun 25, 2012 12:32 am |
Re: KQKP and KRKP endgames |
Robert Hyatt |
Mon Jun 25, 2012 12:17 am |
Re: KQKP and KRKP endgames |
Thomas Petzke |
Sat Jun 23, 2012 4:21 pm |
Re: KQKP and KRKP endgames |
Joona Kiiski |
Sat Jun 23, 2012 4:44 pm |
Re: KQKP and KRKP endgames |
Lucas Braesch |
Sun Jun 24, 2012 3:14 am |
Re: KQKP and KRKP endgames |
Sven Schüle |
Sun Jun 24, 2012 10:14 am |
Re: KQKP and KRKP endgames |
Ronald de Man |
Sun Jun 24, 2012 11:05 am |
Re: KQKP and KRKP endgames |
Sven Schüle |
Sun Jun 24, 2012 11:51 am |
Re: KQKP and KRKP endgames |
Ronald de Man |
Sun Jun 24, 2012 12:39 pm |
Re: KQKP and KRKP endgames |
Sven Schüle |
Sun Jun 24, 2012 2:15 pm |
Re: KQKP and KRKP endgames |
Ronald de Man |
Sun Jun 24, 2012 2:24 pm |
Re: KQKP and KRKP endgames |
Uri Blass |
Sun Jun 24, 2012 1:24 pm |
Re: KQKP and KRKP endgames |
Sven Schüle |
Sun Jun 24, 2012 1:46 pm |
Re: KQKP and KRKP endgames |
Uri Blass |
Sun Jun 24, 2012 1:30 pm |
Re: KQKP and KRKP endgames |
Thomas Petzke |
Mon Jun 25, 2012 7:27 am |
Re: KQKP and KRKP endgames |
Joona Kiiski |
Sun Jun 24, 2012 5:13 pm |
|
You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum
|
|