I really don't wish to belabor PST to death [and am not really "answering" anyone in this paragraph], but I might point out that the Fruit arrays came "with the instructions included". That is, there is not really a blind comparison of array X in Fruit to array Y in Rybka, but a comparison of array X
which came from method M in Fruit with array Y in Rybka. The fact that array Y can be generated most directly
via method M then becomes more relevant, again in an "Occam's Razor" sense.
Furthermore, the evidence is (almost necessarily) merely
prima facie in nature, but wasn't disputed/rebutted by Rajlich, and would seem to meet a "likelihood of probabilities" evidentiary standard (say 75%+ for each relevant array) if there is no further comment, particularly if all PSTs are considered as a group [by which point it could be "beyond a reasonable doubt"]. Any individual example of the Rybka/Fruit PST situation could be attributed to chance, or some other process that produces the same results, which is one reason why I prefer to present them as a whole, and for that matter, to present the PST matter "as part of the whole": first recalling the prior incident of Crafty copying with an earlier Rybka version, then adducing the (more) literally copied parts from Fruit in the iterative search and
go parsing (including time management), adding to that the evaluation comparison overlap, then the hash table idiosyncracies, etc.
Much of the debate (so far) on the Rybka forum seem to consider various prospective explanatory rebuttals Rajlich might possibly make (either with PST or in general), which seems to miss the point of a
prima facie argument [not to mention that none of these explanations were actually made in the ICGA process]. This can also tend to ignore a pertinent legalistic principle, that any accused must present a
specific exposition [to be weighed by the adjudicator, after additional interrogation and/or criticism perhaps], and not rely on the sum-total of all possibilities. Thus, as a personal consideration (if nothing else), I don't have much interest in bickering about what "might have been" (with PST or otherwise), at least when such concoctions are (notably) less probable than any
prima facie and/or "Occam's Razor" argument.
Finally, I must be missing something about Miguel's accounting, as he seems to imply that the
-3 -1 0 1 array/vector is a parameter, whereas I would say that it is specifically fixed for the Fruit/Rybka discussion [or alternatively, that the divisibilities in his method must be kept]. It's somewhat clear that if you allow this array to vary then more configurations are possible PST-wise, but OTOH it's exactly the point that this array
didn't vary (and similarly for most other pieces) when discussing the Fruit process and the Rybka numbers. If the argument is then that Rybka only re-used "half" the numbers of Fruit, I would take this to be fairly conclusive that Rybka is (at least) "half" unoriginal herein.
