Author: Peter McKenzie
Date: 19:06:03 11/15/99
Go up one level in this thread
On November 15, 1999 at 14:48:04, Ratko V Tomic wrote: >> Other models are possible, but in any case I doubt very much that you >> could compress any given position into 64 bits. > >You can encode _any_ 2^64-1=1.84..*10^19 positions in 64 bits, and the code >11...1 (the 64 binary 1s) can be reserved as a prefix to all the remaining >positions. With the right "model" one can even make an arithmetic coder produce >this encoding. Hence, you could have for example all chess positions that were >ever published be in these 2^64-1 "high probability" positions, each taking >exactly 64 bits, and have room to spare for some centuries of play to come. Or, >the same way, you could encode, say a million positions of your choice in 20 >bits each, etc. I'd like to see such a model :-) If the model basically consists of a gigantic lookup table then it kind of defeats the purpose of compression don't you think? Of course none of what you say refutes what I said, because I said 'any given position' (which means any possible position) couldn't be compressed into 64 bits.
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.