Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: SEE for forward pruning in Q. Search and nullmove cooking problems

Author: Robert Hyatt

Date: 11:33:59 08/12/99

Go up one level in this thread


On August 12, 1999 at 13:32:06, Vincent Diepeveen wrote:

>On August 12, 1999 at 09:47:51, Robert Hyatt wrote:
>
>>On August 12, 1999 at 08:32:02, Vincent Diepeveen wrote:
>>
>>>On August 12, 1999 at 01:33:35, Dave Gomboc wrote:
>>>
>>>>On August 12, 1999 at 01:10:14, blass uri wrote:
>>>>
>>>>>This example shows that the null move is not a very good idea.
>>>>>If you need more 3*2=6 plies to see the right move and there is no
>>>>>zunzwnag(playing no move cannot help black) then you are not close to see
>>>>>everything to depth n-r with null move.
>>>>>
>>>>>I thought depth n when you use null move with R=2 means that except for
>>>>>zunzwangs you analyse everything to depth n-2 and I see that it is not the case.
>>>>>
>>>>>Uri
>>>>
>>>>Well, null-move is as you described, but what is shown above is "recursive
>>>>null-move", which seems to be pretty popular, and some people abbreviate
>>>>this as simply "null-move".  It is the sort of search algorithm that gives
>>>>people a happy feeling inside about how deep they are searching, but leaves
>>>>holes for programs like DB to drive a truck through.  Of course, Bob has pointed
>>>>this out more than once before.  Consider how much is being chopped out of some
>>>>13 ply search, and you might agree that Bob isn't just being stubborn: there's
>>>>actually quite a big difference in coverage.
>>>>
>>>>Of course, those who use recursive null-move are making the reasonable gamble
>>>>that the extra coverage isn't beneficial at their search speeds.
>>>>
>>>>Dave
>>>
>>>Now this is pure BS. Even at 5 0 Deep blue gets kicked silly by a laptop.
>>
>>
>>total bullshit.
>
>>no counter-argument needed.  When was the last time you beat my program?  My
>>program can't beat Cray Blitz.  Cray Blitz couldn't beat Deep Thought.  Blitz
>>or tournament.
>
>Well there are a lot of eyewitnesses of the following:
>
>Rebel beating Anand

Rebel lost 1.5-.5 in 40/2 time controls.  Deep Blue beat _Kasparov_ in a
6 game match at 40/2 time control.  What does this have to do with a micro
beating deep blue?



>Rebel beating deep blue junior at a small laptop.

Rebel didn't play "deep blue junior" as I have already explained.  It played
a stateless web-based version that was playing _lots_ of simultaneous games
all over the world.  It is a favorite IBM demo site that is always busy.  Hsu
thought that it might be a 2200-level player at first, but later revised that
estimate downward because of all the simultaneous users hitting on it at the
same time via the web.

Get off of this kick.  I have seen them play micros at ACM events.  I have
seen them play GMs at ACM events.  I watched them put together the best-ever
win/lose ratio on ICC, playing the strongest GM players there, at _any_ blitz
time control they wanted (ie 2 12 which is far from blitz, yet they put everyone
away, maybe losing 1 game at blitz out of 141, with no weak players in the mix).




>Shredder beating Rebel in paderborn.
>
>Now deduction:
>
>DB junior is a laptop version of deep blue. Rebel played at
>a laptop against DB.


DB junior isn't a 'laptop'.  Where do you get your information?  It is a
single-processor SP workstation, accessed via a web browser.  It is fed
positions, it responds with moves.  It knows nothing about 'games' but only
about positions it is given.  It was put together as a demo.  That is all that
it is.  It was _not_ advertised as the normal deep-blue junior machine.  The
real DB junior has had quite a career beating GM after GM in demonstration
matches around the world.  In fact, no GM has beating it in a 'match' so far
that I am aware of.





>
>Shredder beating Rebel chanceless in world champs,
>So shredder beating Deep Blue chanceless.



what utter nonsense.  The Yankees beat the Braves, The Braves beat the
Phillies, the Phillies beat the Yankees. Your theory is shot...




>
>Not very surprising for 'depth' fanatics like you either,
>as Deep blue just gets 11 or 12 ply, and shredder regurarly got
>to 15 ply in paderborn.

different plies.  No null move.  No forward pruning.  No tactical mistakes.
I'd take 12 plies with no null move any day I could get it.  Because I have
_seen_ them play chess.  I have seen them beat a program (Cray Blitz) that
can clean your clock any day of the week.  I have seen them do this 3 times.
Seeing tactical shots 10-20 plies _before_ CB would see them.

You can claim that they are inferior all you want.  But you can count the
games they have lost to micro computers on one hand, and have plenty of fingers
left over to scratch your head in amazement with.





>
>So it just gets outsearched positional completely.


Yes...  so badly that it doesn't lose to other computers, to humans,  heck,
it might not even be able to beat or draw itself...



>
>Secondly i played myselve against deep blue junior, andit
>played positional so bad (tactical it played perfect),
>that with a lot of manoeuvring i got it play all kind of
>crap pawn moves, which even a 3 ply search in DIEP would not
>have make it play such horrible moves.


you played against the web-based toy.  Ask Dlugy about DB junior.  He worked
with them for a good while.  Ask Benjamin.  Or Byrne.  Etc.  You might discover
they have a _totally_ different opinion.




>
>Anyway. All facts together leads to deduction:
>
>Shredder beats Deep Blue chanceless.


Sometimes you make a lot of sense.  Other times you make a lot of nonsense.
Unfortunately, you are doing the latter and not the former here.  You've never
seen them 'in person'.  So you have no idea what you are talking about.  I'm
reminded of a famous quote:

  "it is better to remain silent and be thought a fool, than to open your
   mouth and remove all doubt..."




This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.