Difference between revisions of "TDChess"

From Chessprogramming wiki
Jump to: navigation, search
(Created page with "'''Home * Engines * TDChess''' '''TDChess'''<br/> a chess program by Jonathan Baxter based on KnightCap and its Temporal Difference Learning|TD le...")
 
 
(2 intermediate revisions by the same user not shown)
Line 2: Line 2:
  
 
'''TDChess'''<br/>
 
'''TDChess'''<br/>
a chess program by [[Jonathan Baxter]] based on [[KnightCap]] and its [[Temporal Difference Learning|TD learning]] adaption for [[Minimax|minimax]] search in chess, [[Temporal Difference Learning#TDLeaf|TDLeaf]]. As mentioned by Jonathan, TDChess is about 5-6 times faster in [[Nodes per second|nodes per second]] than KnightCap. KnightCap had a really complicated [[Evaluation|evaluation]] which he had "dumbed down" a lot to get more speed. He also changed the [[Search|search]] a fair bit, including how it does [[Move Ordering|move ordering]] now on the basis of a [[Static Exchange Evaluation|static exchange evaluation]], and how [[Quiescence Search|quiescence]] works. The sorting of moves has become lot faster due to sorting on demand. Also the [[Book Learning|book learning]] is very different from the original KnightCap <ref>[https://www.stmintz.com/ccc/index.php?id=42099 Re: TDchess discussion] by [[Jonathan Baxter]], [[CCC]], February 03, 1999</ref>. TDChess won the [[NC3 1999]] [[Australasian National Computer Chess Championship]] <ref>[https://www.stmintz.com/ccc/index.php?id=87251 Re: Australian-ch, questions !] by [[David Blackman]], [[CCC]], January 09, 2000</ref>.
+
a chess program by [[Jonathan Baxter]] based on [[KnightCap]] and its [[Temporal Difference Learning|TD learning]] adaption for [[Minimax|minimax]] search in chess, [[Temporal Difference Learning#TDLeaf|TDLeaf]]. As mentioned by Jonathan, TDChess is about 5-6 times faster in [[Nodes per Second|nodes per second]] than KnightCap. KnightCap had a really complicated [[Evaluation|evaluation]] which he had "dumbed down" a lot to get more speed. He also changed the [[Search|search]] a fair bit, including how it does [[Move Ordering|move ordering]] now on the basis of a [[Static Exchange Evaluation|static exchange evaluation]], and how [[Quiescence Search|quiescence]] works. The sorting of moves has become lot faster due to sorting on demand. Also the [[Book Learning|book learning]] is very different from the original KnightCap <ref>[https://www.stmintz.com/ccc/index.php?id=42099 Re: TDchess discussion] by [[Jonathan Baxter]], [[CCC]], February 03, 1999</ref>. TDChess won the [[NC3 1999]] [[Australasian National Computer Chess Championship]] <ref>[https://www.stmintz.com/ccc/index.php?id=87251 Re: Australian-ch, questions !] by [[David Blackman]], [[CCC]], January 09, 2000</ref>.
  
 
=See also=
 
=See also=
* [[Various Classifications#Acronym|Acronym]]
 
 
* [[KnightCap]]
 
* [[KnightCap]]
 
* [[Temporal Difference Learning]]
 
* [[Temporal Difference Learning]]
Line 15: Line 14:
 
=References=  
 
=References=  
 
<references />
 
<references />
 
 
'''[[Engines|Up one Level]]'''
 
'''[[Engines|Up one Level]]'''
 +
[[Category:Acronym]]
 +
[[Category:Chess Suffix]]

Latest revision as of 22:19, 13 January 2019

Home * Engines * TDChess

TDChess
a chess program by Jonathan Baxter based on KnightCap and its TD learning adaption for minimax search in chess, TDLeaf. As mentioned by Jonathan, TDChess is about 5-6 times faster in nodes per second than KnightCap. KnightCap had a really complicated evaluation which he had "dumbed down" a lot to get more speed. He also changed the search a fair bit, including how it does move ordering now on the basis of a static exchange evaluation, and how quiescence works. The sorting of moves has become lot faster due to sorting on demand. Also the book learning is very different from the original KnightCap [1]. TDChess won the NC3 1999 Australasian National Computer Chess Championship [2].

See also

Forum Posts

Re: TDchess discussion by Jonathan Baxter, CCC, February 03, 1999

References

Up one Level