Author: Uri Blass
Date: 11:11:47 01/29/02
Go up one level in this thread
On January 29, 2002 at 13:58:20, Dann Corbit wrote: >On January 29, 2002 at 10:08:46, Robert Hyatt wrote: > >>On January 28, 2002 at 19:41:07, Les Fernandez wrote: >> >>>On January 28, 2002 at 19:25:46, Vincent Diepeveen wrote: >>> >>>>On January 28, 2002 at 19:07:21, Les Fernandez wrote: >>>> >>>>>On January 28, 2002 at 19:01:26, Vincent Diepeveen wrote: >>>>> >>>>>>On January 28, 2002 at 18:54:07, Dann Corbit wrote: >>>>>> >>>>>>>On January 28, 2002 at 18:44:15, Vincent Diepeveen wrote: >>>>>>> >>>>>>>>On January 28, 2002 at 18:17:11, Dann Corbit wrote: >>>>>>>> >>>>>>>>that only shows how to store KRK as far as i see Dann, >>>>>>>>not a random position with nearly all pieces on the board. >>>>>>>> >>>>>>>>Not a single example of a full board position is inside the document. >>>>>>>> >>>>>>>>please encode next position for me, ignore castling rights doing it: >>>>>>>> >>>>>>>>nr3qrk/2QRp1Np/2p1Pp1n/2Pp3P/pp1P1K1P/3B1P2/PP1BNbp1/R7 w - - 0 1 >>>>>>> >>>>>>>I assume that you can read his simple encoding system. Now, take that method >>>>>>>and compose the position for it. Then consider that that position (together >>>>>>>with its eval, ce, pv, etc) are identical to these, if you have read "Through >>>>>>>the Looking Glass": >>>>>> >>>>>>>7r/1PBnb1pp/2p1b3/p1k1p1PP/p3Pp2/N1Pp1P2/Pn1Prq2/KRQ3RN b - - >>>>>>>krq3rn/pN1pRQ2/n1pP1p2/P3pP2/P1K1P1pp/2P1B3/1pbNB1PP/7R w - - >>>>>>>nr3qrk/2QRp1Np/2p1Pp1n/2Pp3P/pp1P1K1P/3B1P2/PP1BNbp1/R7 w - - >>>>>>>r7/pp1bnBP1/3b1p2/PP1p1k1p/2pP3p/2P1pP1N/2qrP1nP/NR3QRK b - - >>>>>> >>>>>>>Since we encode 4 positions and need to store only 1 (the one that is lexically >>>>>>>smallest on top) we divide the number of bits needed by 4. It is a trick so >>>>>>>simple that I am surprised anyone would not grasp the notion instantly. >>>>>> >>>>>>that reduces the thing by 2 bits. Somehow i get impression you guys >>>>>>confuse bits with bytes. You store positions in 162 bytes? >>>>> >>>>>Hi Vincent, >>>>> >>>>>No I think we are speaking of bits but my examples are representations of bits >>>>>but are actaully ascii at the moment but concept wise the same. Hey check this >>>>>out!! >>>>> >>>>>1.4 bits/position All positions have been extracted from a 63 bit binary key. >>>>>63/44= 1.4 bits/position. Thought you might enjoy this <S> >>>> >>>>your approach is worth nothing. because with my compression i get >>>>under 1 bit a position then. >>>> >>>>all you do is: "oh we need 250 bits to store a position, and we >>>>can mirror it 4 times ==> 250 / 4 = 60 bits a position needed". >>>> >>>>That is not funny of course. >>>> >>>>You need 250 bits in that case *not* 60 bits. >>>> >>> >>>Hello Vincent, >>> >>>I actually was not trying to be funny. The novelty is the fact that variants >>>can be extracted from one position, not just mirrored images. Therefore all you >>>need to have in your database one of these key positions and you get all the >>>rest for free. Anyway from a storage point of view it could be of interest >>>since storage is becoming a concern as new egtb's are generated. I played >>>around with 175,168 positions represented in binary (only 3 pieces KRk) and >>>after pkzip was used file size went from about 12MB down to about 700K. Granted >>>this was a text file at the moment, not long integer, but the compression >>>reduction I expect to be very good due to the fact the file is made up of only >>>1's and 0's. >>> >>>Les >>> >>> >>>> >> >>OK. >> >>First you are saying that you have proven that there are no more than 2^81 >>unique positions where neither side can castle... > >No. His notion is that if you mirror using every symmetry, the total number of >those positions (including ALL reflections) would be less than 2^81 in that >category. If this is his point then he is wrong. The number of reflection of the same position is only 4 so the total number of classes of 4 positions is only 1/4 of the number of positions and it cannot be less than 2^81. Uri
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.