hgm wrote:benstoker wrote:Second Place is just another way to say First Loser.
You don't seem to get it. The strongest program won't win at all in your tournament style. The program that enters most cheating copies will. In the current engine landscape Houdini would not stand a chance. 'Bodini' will finish some 200 points above it.
Who will become 'TopDog' might as well be decided by throwing dice. Is that what you want? The whole idea is a total bust.
Ok, if you want to steer the thread this way and re-define Top Dog that way, then I win. I'll enter in 1 million clones designed to lose to my Super Duper Cheat Machine (which runs on a Casio calculator watch from 1989) and not only that, buy off all the other competitors and tourney officials with expensive bribes.
Part of what you say seems to be that there MUST be at least SOME kind of filter for a tournament, otherwise, you get the above absurdities. I agree.
The question is what is that filter.
The argument is that clones must be filtered out even if the clone is stonger, because of some kind of "originality" criteria. I ask why would you filter out the STRONGEST engine? You say because it isn't "original". I ask again, what is "original" in chess engines today? What is NOT merely a tweak or optimization of public domain routines? Whatever Ippo* had that made it stronger than R3, was NOT some grand paradigm shifting new algorithm, right? You all looked at the code and should know by now. Ippo* just implemented already known and used routines in a more efficient or optimized manner, right? Ditto for Fruit to Rybka.
Now, the winner of the Free-For-All-No-Holds-Barred-Anything-Goes-Including-Murder-And-Bribery-And-Hostage-Taking World Chess Engine Tournament convening at the Bulldog in Amsterdam would never be a Top Dog.
Top Dog beats all other programs mano a mano. Simple as that.
The funhouse scenario described above is not even chess. It's some other game for thugs and miscreants.
The point I attempt to make is legalistic, or philosophical maybe.
There is so much focus on this idea of "originality" in this Attack of the Clones debate. I ask you programmers to identify one single chess engine algorithm that is PATENTABLE. Are bitboards patentable by someone? SEE? Does anybody even own any kind of patent on a chess engine? Identify it please. Are any of the following routines owned by anyone and patentable, or are they each and every one of them public domain:
- 8*8 Board
Mailbox
0x88
15*12 Board
16*12 Board
Pawn Fills
Pawn Spans
Attack Spans
Dumb7Fill
Kogge-Stone Algorithm
Fill by Subtraction
ABDADA - Alpha-Bêta Distribué avec Droit d'Anesse
All-Nodes
Biboards
BitScan
Butterfly Boards
Cut-Nodes
Depth-First
DTS - Dynamic Tree Splitting
EBF - Effective Branching Factor
ETC - Enhanced Transposition Cutoff
Exact Score
Fail-Hard
Fail-High
Fail-High Nodes = Cut-Nodes
Fail-Low
Fail-Low Nodes = All-Nodes
Fail-Soft
FHR - Fail-High Reductions
Frontier Node
Futility pruning
History Heuristic
IBV - Integrated Bounds and Values
ID - Iterative Deepening
IID - Internal Iterative Deepening
Killer Move - Killer Heuristic
LMR - Late Move Reductions
LS1B - Least Significant One Bit
Minimax
MS1B - Most Significant One Bit
MVV-LVA - Most Valuable Victim - Least Valuable Aggressor
Negamax
NegaScout
Nibble
NMH - Null Move Heuristic
NMO - Null Move Observation
Null Window
Passed Pawn
Pre Frontier Node
Pruning
PV - Principal variation
PV-Nodes
PVS - Principal Variation Search
PVS - Principal Variation Splitting
QS - Quiescence Search
R - Depth Reduction R
Ranks
Razoring
Scout - Scout-Algorithm
SE - Singular Extensions
SEE - Static Exchange Evaluation
Side to move
SIMD - Single Instruction Multiple Data
Skewer
SSE - Streaming SIMD Extensions
SSS = State Space Search like SSS* and Dual*
Stop square
SWAR - SIMD Within A Register
TD - Temporal Difference Learning
TT - Transposition Table
X-ray
YBW - Young Brothers Wait
Zobrist Hashing
Again, after all is said and done, forever and anon, from here to eternity, Top Dog will always be that one special engine that has that one itsy bitsy piece of originality that makes it beat all other engines.
Houdini is the Top Dog today. It doesn't matter one iota that it is a wholesale rip of ivanhoe. Of necessity, there's something ORIGINAL added to Houdini, because there ain't no other program as good as it is.
In fact, all programs OTHER than Houdini are but copies and clones of the same worn out chess engine algorithms. It is only Houdini that has achieved something TRULY ORIGINAL and PROVEN it by being STRONGER THAN ALL OTHER PROGRAMS.