Does anyone here have a better estimate than the following?

Complete 3 man Nalimov Tablebase.................80 kB (measured)
Complete 3+4 man Nalimov Tablebase...............30 MB (measured)
Complete 3+4+5 man Nalimov Tablebase............7.5 GB (measured)
"Complete" 3+4+5+6 man Nalimov Tablebase........1.146 TB (measured)[1]
Complete 3+4+5+6+7 man Nalimov Tablebase....200-600 TB (estimated)
Complete 3+4+5+6+7+8 man Nalimov Tablebase...40-180 PB (estimated)
[...]
Complete 3+4+5+6+[...]+30+31+32 man Nalimov tablebase...[UNKNOWN]

All sizes are for Nalimov Tablebases using standard compression.

(Note 1: As of this writing, nobody has bothered making tablesbases
for positions where a single king faces five pieces. It's hard to
imagine a non-stalemated 5 vs. 1 position where a computer can't
find a quick win without consulting a tablebase.)

I welcome any estimates that are better than the ones above, and
I especially welcome anyone who has the courage to give me an
estimate to replace the [UNKNOWN] above. I would also very much
like some data on how long it takes/took to generate each set.
-Guy Macon

------------------------------------------------------------

BTW, Here is a handy reference for expressing large numbers:

k kilo- 1.0E+3 x1,000 Thousand(US,UK)
M mega- 1.0E+6 x1,000,000 Million (US,UK)
G giga- 1.0E+9 x1,000,000,000 Billion(US)
T tera- 1.0E+12 x1,000,000,000,000 Trillion(US), Billion(UK)
P peta- 1.0E+15 x1,000,000,000,000,000 Quadrillion(US)
E exa- 1.0E+18 x1,000,000,000,000,000,000 Quintillion(US), Trillion(UK)
Z zetta- 1.0E+21 x1,000,000,000,000,000,000,000 Sextillion(US)
Y yotta- 1.0E+24 x1,000,000,000,000,000,000,000,000 Septillion(US), Quadrillion(UK)

Note: vendeka, xenna, xenno, and vendeko are bogus.
Before you try using them, please read these pages:
[ http://home.att.net/~numericana/answer/units.htm#prefix ]
[ http://physics.nist.gov/cuu/Units/prefixes.html ]
[ http://physics.nist.gov/cuu/Units/binary.html ]


G o o g l e F o o d : http://www.guymacon.com http://www.guymacon.com/
Guy Macon Guy Macon Guy Macon Guy Macon Guy Macon Guy Macon Guy Macon Guy Macon
Guy Macon Guy Macon Guy Macon Guy Macon Guy Macon Guy Macon Guy Macon Guy Macon
Guy Macon Guy Macon Guy Macon Guy Macon Guy Macon Guy Macon Guy Macon Guy Macon

On Fri, 22 Jun 2007 14:03:18 +0000, Guy Macon
http://www.guymacon.com/ wrote:


(Note 1: As of this writing, nobody has bothered making tablesbases
for positions where a single king faces five pieces. It's hard to
imagine a non-stalemated 5 vs. 1 position where a computer can't
find a quick win without consulting a tablebase.)


5 vs. 1 tablebases have been created, but not in Nalimov format.
Unfortunately, the code provided by Nalimov cannot as yet handle 5 vs
1 tablebases. You may want to check the CCRL tablebase forum for more
info.

http://kd.lab.nig.ac.jp/chess/discus...wforum.php?f=6

Tony

Tony M wrote:

Guy Macon http://www.guymacon.com/ wrote:

(Note 1: As of this writing, nobody has bothered making tablesbases
for positions where a single king faces five pieces. It's hard to
imagine a non-stalemated 5 vs. 1 position where a computer can't
find a quick win without consulting a tablebase.)

5 vs. 1 tablebases have been created, but not in Nalimov format.
Unfortunately, the code provided by Nalimov cannot as yet handle 5 vs
1 tablebases. You may want to check the CCRL tablebase forum for more
info.

http://kd.lab.nig.ac.jp/chess/discus...wforum.php?f=6

Interesting! I will dig deeper later today.

Do you by any chance know the size (in any format) with and without
the 5 vs. 1 tablebases?

On Sat, 23 Jun 2007 16:52:24 +0000, Engineer
wrote:




Tony M wrote:

Guy Macon http://www.guymacon.com/ wrote:

(Note 1: As of this writing, nobody has bothered making tablesbases
for positions where a single king faces five pieces. It's hard to
imagine a non-stalemated 5 vs. 1 position where a computer can't
find a quick win without consulting a tablebase.)

5 vs. 1 tablebases have been created, but not in Nalimov format.
Unfortunately, the code provided by Nalimov cannot as yet handle 5 vs
1 tablebases. You may want to check the CCRL tablebase forum for more
info.

http://kd.lab.nig.ac.jp/chess/discus...wforum.php?f=6

Interesting! I will dig deeper later today.

Do you by any chance know the size (in any format) with and without
the 5 vs. 1 tablebases?

http://kd.lab.nig.ac.jp/chess/tablebases-online/ lists the sizes of
each individual tablebase file in Nalimov format. I'm not sure about
the sizes in Chessmaster FEG format.

Tony