So much to do.

[hfwolfe]

Tonight hanging out with S---h and Y----f, probably at Aurel

Comments

in the likeness of the human mind

[random_walker]

Those card-shufflers have documented vulnerabilities. http://www.maa.org/mathland/mathtrek_11_18_02.html

Recommend you analyze the machine's patterns and exploit in active & passive ways. Genius to bring in this kind of white elephant, I bet none of them suspect.

Be well, hbar.

-RW

Re: in the likeness of the human mind

[hfwolfe]

Hey man! I was just about to write you traditionally.

Yeah, I read about that. my little plastic shuffler just does a riffle, but with a human triple cut and three passes, I don't know anyone smart enough to exploit it.

So the exploit you are referring to is normally executed with a bayes technique called "shuffle tracking" which I just baught a book on, but it's pretty usless in poker. I'm dreaming of quitting what I'm doing to do tracking blackjack full time. Who do we know that can hardcode me a computer into a liesure suit? I want you to write the inference ware, though. I want a big cut, since you won't risk being beaten up in a casino parking lot in Atlantic City.

Also, what is your impression of specific data mining texts, and the field in general? I've been thinking of branching out a bit.

Re: in the likeness of the human mind

[random_walker]

Interesting. Never fear, I've nothing but respect for the real work in gambling. In Nevada I think card-counting is only ejectable, whereas in Atlantic City it is considered cheating and therefore a crime. The trick I suspect is the realtime i/o. I wonder what the expected payoff-per-dollar from an optimal full-information strategy is

Re: in the likeness of the human mind

[hfwolfe]

No doubt. One would assume the phase space and entropy of most real datasets requires some experienced assumptions to limit the conjecture space.