Prev: misc: how random is 'random'?...
Next: Record stack back traces efficiently on space
From: Malcolm McLean on 5 May 2008 05:07

"rossum" <rossum48(a)coldmail.com> wrote in message
> Or alternatively the programmer has decided to make life easier for
> the undertrained maintenance programmers who will follow her in
> maintaining the code. A few extra parentheses can make code clearer
> and easier to maintain.
>
There's a strong case for an extra set of parentheses in that one line, I
agree, even though I left them out myself.

--
Free games and programming goodies.
http://www.personal.leeds.ac.uk/~bgy1mm

From: Malcolm McLean on 6 May 2008 15:44

"user923005" <dcorbit(a)connx.com> wrote in message
>The bias has to be fairly large. Suppose that there is a red/black
>bias of 1% {doesn't matter which way}. I guess that it would be good
>for the casino if it became known, since many customers would go broke
>trying to exploit it.
>The house still wins because they get 1/37th, which is well over 1%
>{assuming that the house 0 slot gets 1/37th give or take -- they will
>definitely notice right away if this is not happening -- it's the only
>thing that they really care about}.
>
The bias needs to be about 5%, because you are losing money on black,
supposing the bias to be to red, and also on high/low and even/odd.

>E.g.: number of slots of biased positive side is 18/37, 1% advantage
>for that side, less 1/37th for the house slot gives us:
>(18/37) * 1.01 - 1/37 = 0.46432 win expectancy.
>So for a wheel biased by 1% you win 46.432 percent of the time and
>lose 1-.4632= 53.567% of the time
>Does anyone seriously believe a casino will never test their wheels?
>Or that an imbalace of as small as 1% is hard to detect?
>
The point is you pick up the bias before statistically significant results
showing which way it is set are in. The system only works if the wheel is so
manufactured that it isn't possible to be unbiased, even though it might be
possible to reset the bias to random now and again.

>Casinos are run by people with poor reputations. I guess that even if
>a problem were somehow discovered it would be a very dangerous game to
>play.
>Thanks, but I think I'll keep my kneecaps.
>
Oh yes. But gangsters might be reading this, and they know how to look after
themselves. All part of the comp.programming service.

--
Free games and programming goodies.
http://www.personal.leeds.ac.uk/~bgy1mm

From: user923005 on 5 May 2008 21:39
On May 3, 5:54 am, Richard Heathfield <r...(a)see.sig.invalid> wrote:
> Malcolm McLean said:
>
>
>
> > "Richard Heathfield" <r...(a)see.sig.invalid> wrote in message
>
> >> If you know that the wheel is biased to black so much that it overcomes
> >> 0 and, on some wheels, 00 as well, then just bet on black on every spin,
> >> taking care not to push too hard. (Cash in at about 11pm, get a good
> >> night's sleep, and head back for the same table the following night.
> >> With luck, the house won't notice the bias for a week or two.)
>
> >> You don't need a computer program to tell you that.
>
> > But what if know that some wheels are biased, some are fair, and you
> > don't know which are which and which way the biased wheels are biased?
>
> If you don't know how a biased wheel is biased, your computer program won't
> know either.
>
> To find out whether a wheel is biased, track its results and see how far
> they depart from random. If the difference is statistically significant,
> by all means consider risking a flutter. Any casino that runs a biased
> wheel is bound to lose in the end.

The bias has to be fairly large. Suppose that there is a red/black
bias of 1% {doesn't matter which way}. I guess that it would be good
for the casino if it became known, since many customers would go broke
trying to exploit it.
The house still wins because they get 1/37th, which is well over 1%
{assuming that the house 0 slot gets 1/37th give or take -- they will
definitely notice right away if this is not happening -- it's the only
thing that they really care about}.

E.g.: number of slots of biased positive side is 18/37, 1% advantage
for that side, less 1/37th for the house slot gives us:
(18/37) * 1.01 - 1/37 = 0.46432 win expectancy.
So for a wheel biased by 1% you win 46.432 percent of the time and
lose 1-.4632= 53.567% of the time
Does anyone seriously believe a casino will never test their wheels?
Or that an imbalace of as small as 1% is hard to detect?

Casinos are run by people with poor reputations. I guess that even if
a problem were somehow discovered it would be a very dangerous game to
play.
Thanks, but I think I'll keep my kneecaps.
First  |  Prev  | 
Pages: 1 2 3 4 5
Prev: misc: how random is 'random'?...
Next: Record stack back traces efficiently on space