Rybka 4 beaten blindfold by neurosurgeon

[zullil]

proving that he does not remember 30 million digits of PI is easy

choose 1000 random numbers 1-30,000,000 and test him if he can write the digits in the relevant 1000 places.

I seem to recall that it is possible to calculate any specified digit of pi without knowledge of any of the preceding digits. I don't recall the complexity of the calculation, and I seem to think one typically works in base 16!

[Onno Garms]

I also think this is a delayed April Fool's joke.

However, if a neurosurgeon is crazy enough, he might have a developed a way to combine a cochlear implant with a communication device and make it possible to hide the whole thing in his skull. Scary idea. Wonder if he can repeat the trick when not playing blindfold, i.e. without the moves being announced.

EDIT: Did anybody check the moves of the neurosurgeon for matches with moves of Houdini?

[Daniel Shawul]

His claim is actually 200 million based on his wiki page. The world rankings do not list him but the top guy there can recite about 70000 digits
http://pi-world-ranking-list.com/lists/memo/index.html

[rbarreira]

I also think this is a delayed April Fool's joke.

However, if a neurosurgeon is crazy enough, he might have a developed a way to combine a cochlear implant with a communication device and make it possible to hide the whole thing in his skull. Scary idea. Wonder if he can repeat the trick when not playing blindfold, i.e. without the moves being announced.

If he keeps coming up with more and more unbelievable feats, your idea might become the most plausible explanation...

[M ANSARI]

With enough time and effort and a large budget, I am sure someone could design a system that would pass all scrutiny and still allow some form of electronic device to be used. The data transfer required for chess moves or for knowing pi numbers is not that much. Just like some magicians can make buildings or passenger airliners "disappear convincingly" in front of you... I am sure that beating R4 or remembering pi numbers is not much more difficult.

[Cubeman]

Thank you Graham!!

It seems Father that your exploits are now surpassed, no longer will your posts match those of this topic.From now on if ever the world needs to defeat the aliens in a game of chess I now choose this neuro-surgeon.

[BubbaTough]

Is it possible he just memorized a game that was a win against Rybka without book, and then reproduced it? Do people actually think an unrated player did this without a trick?

-Sam

[Albert Silver]

Is it possible he just memorized a game that was a win against Rybka without book, and then reproduced it? Do people actually think an unrated player did this without a trick?

-Sam

Unfortunately you would be surprised at the number of gullible people dying to believe it. That is why the Chessbase article warns and of course makes ample fun of it. Those who check out the suggested further reading should have a good laugh.

[Father]

Hello Alex Brown,
Finally I could be in an absolutle silence and rest in pease. q.e.p.d.
A new sistem playing computers in Man versus Machine we will be seeing, against and again.
Antichess will be an obsolete and old technique of XXI century.
I am happy.

[Onno Garms]

Is it possible he just memorized a game that was a win against Rybka without book, and then reproduced it?

Did they play without book?

Engines tend not to reproduce games exactly due to tiny timing differences. This might work in some cases, but normally does not.

If you set up a single position, the probability that the same move gets played again on identical hardware with identical time is large. But over a game the probability of getting a different move somewhere is large because of different timing in the computer. Also it is virtually impossible to feed the computer with exactly the same time (1 millisecond accuracy). Finally, once the engine has computed a move (with slightly different timing) its hash table content differs.