Manual de Fritz 15 2015 | Page 268

268

Fritz 15 Help

3.8.17 End Game Table Base support
The Program integrates the Syzygy table probing code © Ronald de Man, the Gaviota
EGTB probing code © Miguel A. Ballicora and the Nalimov EGTB probing code © Eugene
Nalimov.

Syzygy bases
The Syzygy 6-men EGTB system is the recommended table base solution for Houdini.
It features by far the best compression (smallest file size) and best performance,
especially when multiple threads perform simultaneous EGTB probing. Houdini has been
extensively tuned for this state-of-the-art end game table base system so as to
provide optimum precision and strength in the late endgame at any time control and
with any number of cores. While installing the files on a SSD disk is recommended, the
performance of the Syzygy on a normal hard disk is still very good.
The only drawback of Syzygy bases is that they do not contain the exact Distance to
Mate information. If you require an exact mating distance in a position you'll need an
alternative solution.
The Syzygy 6-men EGTB system is the recommended table base solution for Houdini.
For the Syzygy EGTB support you'll need to install the Syzygy EGTB files. At the time
of writing all the Syzygy files are available for torrent download at http://oics.
olympuschess.com/tracker/index.php. The 3-4-5-men Syzygy are also available for
direct download at http://olympuschess.com/egtb/sbases.
The total size of the 3-4-5-men files is 938 MB (290 files), the 6-men table bases
require 149 GB (730 files).

Nalimov EGTB
For the Nalimov EGTB support you'll need to install the Nalimov EGTB files.

Question: I don't understand the concept of setting a depth for EGTB probing.
Why wouldn't all positions that have 3-4-5 pieces probe the table bases to get
the evaluation?
EGTB probes are extremely slow compared to a normal evaluation by the engine.
Suppose you have a K+Q+P against K+N ending. Even without consulting the table
bases Houdini knows that this ending is easily won for the K+Q+P side. Consulting the
EGTB for this position would reduce Houdini's playing strength, as it could easily have
evaluated 1,000 other positions instead of making the rather useless EGTB probe.
Even the in-memory "soft" Gaviota or Scorpio probes are sometimes relatively slow
compared to a native evaluation of the position - especially when multiple threads are
running. If two threads simultaneously perform a soft probe, the second thread has to
wait for the first to finish its probe.
This is aggravated by the fact that Houdini contains a lot of end game knowledge in
its native evaluation function - a lot more than most other chess engines. For
example, even in a non-obvious ending like KBP v KB with bishops of the same color,

© ChessBase 2015