Just another conspiracy theory

[Karlo Bala]

I know it is not likely, but maybe, who knows...

I'm looking at strong Fritz propaganda, and ... what if a first ippolit's source code was somehow infiltrated by chessbase?

[Eelco de Groot]

Not sure I understand exactly what this theory is Karlo... But I wouldn't be surprised if Frans Morsch never looked at the program, barely heard it exists somewhere on the Net. What evidence do we have that even he reads this board? Only as a guest I think. But he does not really need it and I think he prefers to find his own ways. He even turned down Ken Thompson when Ken came with his plans for Belle on a chip so Donninger went in his stead. Fritz is not a bitbase program I think although this was not explicitly stated in the interview, they avoid technical terms that, only scares the readers away, this story is more for people that barely know Fritz, never heard of Frans Morsch probably.

Frans does try to improve the results against Rybka and Fritz 11 clearly had adapted its style to compete more, but that is what you would expect. It is after all a competitive field. The most important Rybka influence mentioned is no doubt the testing method:

For the development of the playing strength we employ new automated test methods. A few million games are played on a series of computers. This shows us precisely the efects of changes in the search or evaluation functions.

How is this mass of data evaluated?

The use of advanced mathematical procedures on the base data eliminates statistical errors. This allows continuous further development.

Some would argue that Fabien Letouzey started that too with Fruit and they have some point, and programmers like Ed Schröder and Christophe Théron they already tested everything with batteries of PCs also long before Fruit entered and conquered, coming seemingly out of the blue but really also after much testing. Hardware advances, multiprocessing and lowering costs per CPU have taken this just a magnitude further now, human tuning by hand just can't produce results fast enough and basically can't compete. For sure there are no secrets given away in the interview but is nice to hear something from Frans himself. I believe I once saw Frans when he was at the "HCC computerdagen" with his new Mephisto chesscomputer giving demonstrations. That must have been the Mephisto Mondial with the program Nona inside, 6502 processor at 2 MHz and 16 kilobytes ROM, or the sucessor Mondial II with 32 kB and 5 MHz 6502. Frans comes from assembler days, he does not need Ippolit!

[Hood]

hi,

if the company will not look to the developments of others
it is going away from the market.

it is general rule.

if the engineer does not want to do it his boss will made him

rgds
chris

[Gian-Carlo Pascutto]

But I wouldn't be surprised if Frans Morsch never looked at the program, barely heard it exists somewhere on the Net.

Frans comes from assembler days, he does not need Ippolit!

That's true, somebody whose native language is assembler is probably better off reading the original rather than the decompiled C

(Sorry for the troll, couldn't resist)