Derek Nalls wrote on Fri, May 23, 2008 12:47 AM UTC:
Originally, I planned two 'internal playtests'. [By this self-invented
term I mean playtests of the standard model of a person against a special
model that I have compelling reasons to think may be superior by a
provable margin.]
The first planned test involves the standard CRC model of Muller against a
special CRC model with a higher, closer-to-conventional rook value. Upon
closer examination, I suspected that the discrepancy was possibly too
small to be detected even with very long time controls. So, I announced
that this test was cancelled.
Notwithstanding, I may change my mind and return to this unsolved mystery
if Joker80 demonstrates unusually-high aptitude as a playtesting tool.
This might require very deep runs of moves with a completion time of a few
weeks to a few months per pair of games to achieve conclusive results.
The second planned test involves the standard CRC model of Scharnagl
against a special CRC model with a higher, unconventional archbishop
value.
Scharnagl currently assigns the archbishop with a material value of appr.
77% that of the chancellor in his standard CRC model.
Muller currently assigns the archbishop with a material value of greater
than 97% that of the chancellor in his standard CRC model.
Nalls currently assigns the archbishop with a material value of lesser
than 98% that of the chancellor in his standard CRC model.
I devised a special CRC model using identical material values for every
piece in the standard CRC model by Scharnagl except that it assigns the
archbishop with a material value of exactly 95% that of the chancellor
(18% or 1.65 pawns higher). [Note that this figure is slightly more
moderate than those by Muller & Nalls.] A discrepancy this large should
be detectable at short-moderate time controls. This test is now
underway.
If either of these tests are successful at establishing or implicating a
probability that the special models play stronger than the standard
models, then revisions to the standard models may occur. At that
juncture, we would be ready to begin 'external playtests'. [By this
self-invented term I mean playtests of the standard models of different
persons against one another.]