syzygy wrote:Larry's statement is innocent but....
I worked with Chris Whittington on the chess program
Chess system tal.
The program played like tal, it really sacced pieces and attacked opponent.
I remember playing at the aegon tournament with chess system tal against bronstein or Sofia polgar, putting the "attack with mickhail tal" book cover close to the PC screen.
What seems like a fairy tale today may be reality tomorrow.
Here we have a fairy tale of the day after tomorrow....
Indeed. Like claiming that putting second-hand Chess knowledge from Fischer in the evaluation is the same thing as claiming the engine plays like Fischer, and defending that silly thought by claiming most people would misunderstand that in the same way.
hgm wrote:Indeed. Like claiming that putting second-hand Chess knowledge from Fischer in the evaluation is the same thing as claiming the engine plays like Fischer, and defending that silly thought by claiming most people would misunderstand that in the same way.
Most customers of chess software are not such talented and keen-witted programmers like you. And most customers will conclude that an engine with "second-hand Chess knowledge from Fischer in the evaluation" will play like Fischer or at least "human". Publishing such comments is called marketing.
syzygy wrote:Larry's statement is innocent but meaningless.
This, basically.
Btw, the evaluation function of my own patzer engine can be said to be second-hand knowledge from Max Euwe ("Praktische Schaaklessen", not "Oom Jan leert zijn neefje schaken").
So is mine, what are the odds?
(Pretty good, I imagine.)
Actually, it has second-hand knowledge from Nimzowich too.
Alexander Schmidt wrote:Most customers of chess software are not such talented and keen-witted programmers like you. And most customers will conclude that an engine with "second-hand Chess knowledge from Fischer in the evaluation" will play like Fischer or at least "human". Publishing such comments is called marketing.
So most customers are stupid, and a fool and his money are soon parted. That doesn't make it wrong to say things they will misunderstand, or make it right for them (or us) to misunderstand such things.
As Canada Bill used to say:
It's morally wrong to allow suckers to keep their money.
I don't think in this case the customers have much right to complain, though: the engine in fact plays much stronger than Fischer.
Nobody here is complaining about the statement, let alone suggesting that anyone should get his money back. So relax.
I agree with Lucas that they "just did what worked in testing, like everyone else".
The fact that Larry is an excellent chess player may certainly have helped in the selection of ideas to be tested. Nobody is denying that. It does not contradict what Lucas said.
That, as a result of this, "a lot of Rybka's evaluation" can be qualified as Fischer-specific chess knowledge (one way of understanding "second-hand knowledge from Fischer") is not something that should be taken very seriously.
Of course given that Larry learned a lot from Fischer's columns early in his chess career, there might well be something in Rybka that can somehow be traced to some of those columns, and in that sense it might be called "second-hand knowledge from Fischer", but it will more likely be general chess ideas that were communicated to Larry by Fischer than specific chess knowledge that is characteristic of Fischer.
If I have learned how to play chess by reading Max Euwe's books, do I possess (first-hand? second-hand?) knowledge from Euwe? Sure, why not.
Well, the statement does not even claim it was Fischer-specific Chess knowledge. I am sure Fischer did have a lot of general Chess knowledge too. So conluding from this that the engine must play like Fischer just makes no sense.
