For sake of simplistic simplicity, I suggest that a good definition of long
run is: the point at which it is most probable that ER has emerged, based
on perfect play. SD=most probable (simplistically). Note the distinction
between "has" emerged and "will" emerge.
lb
For sake of simplistic simplicity, I suggest that a good definition
of long
run is: the point at which it is most probable that ER has
emerged, based
on perfect play. SD=most probable (simplistically). Note the
distinction
between "has" emerged and "will" emerge.
And just how do you suggest (in your most simplistic way, of course)
that one determine that "point"?
It's one thing to establish a definition, but quite another for the
definition to mean something concrete and useful. Seems to me that
the trouble with "long term," "long run," or whatever we call it is
that it's an idea, an abstraction, a concept, rather than something
concrete, something that can be nailed down. I think that I know what
I mean personally by the long term, but darned if I have any
confidence in its actually meaning anything useful. Dan speaks of
pulling a number out of the air early in his vp interest (200000
hands) but later deciding that 4 million is perhaps better. Why?
Don't both those numbers come from the air? Intuitively, I would be
inclined to believe that the larger number is better. But is it
better than, say, 2 million, or 6? And just what, exactly would make
it so?
Pete
--- In vpFREE@yahoogroups.com, "Lawrence Boxer" <ljboxer@e...> wrote:
I like Tom Ski's method for determing the confidence level of a game
on our bankroll. Long-term for most of us is 95% confidence. This
is my prefered way to think of for me and I'm willing to risk my
bankroll on 95%. To a fundementalist long-term might be 99.9%
confidence.
Long-term isn't an exact point it is a subjective term like "This
room is warm". How warm? Well we measure that in degrees and can
measure long-term with the confidence factor. These are objective
terms used for a more scientific measurement.
I mean personally by the long term, but darned if I have any
confidence in its actually meaning anything useful. Dan speaks of
pulling a number out of the air early in his vp interest (200000
hands) but later deciding that 4 million is perhaps better. Why?
Don't both those numbers come from the air? Intuitively, I would
be
inclined to believe that the larger number is better. But is it
better than, say, 2 million, or 6? And just what, exactly would
make
Agreed! Paymar (see post 21332) suggests essentially the same thing
as you with a 5% ROR and applying Sorokin's formula.
Pete
-- In vpFREE@yahoogroups.com, "Doug Smith" <oztun_2000@y...> wrote:
I like Tom Ski's method for determing the confidence level of a game
on our bankroll. Long-term for most of us is 95% confidence. This
is my prefered way to think of for me and I'm willing to risk my
bankroll on 95%. To a fundementalist long-term might be 99.9%
confidence.Long-term isn't an exact point it is a subjective term like "This
room is warm". How warm? Well we measure that in degrees and can
measure long-term with the confidence factor. These are objective
terms used for a more scientific measurement.--- In vpFREE@yahoogroups.com, "paisonvp" <paison@i...> wrote:
> I mean personally by the long term, but darned if I have any
> confidence in its actually meaning anything useful. Dan speaks of
> pulling a number out of the air early in his vp interest (200000
> hands) but later deciding that 4 million is perhaps better. Why?
> Don't both those numbers come from the air? Intuitively, I would
be
> inclined to believe that the larger number is better. But is it
> better than, say, 2 million, or 6? And just what, exactly would
make
> it so?
>
> Pete
My simplistic answer is that long term is reflected in standard
deviation. Since ER is based on perfect play over the theoretical
long term, at what point does sd satisfy you that play will result in
a win (or that results will at least fall within a range that makes a
win most probable)? Simplistically, that is the long term. And it
satisfies the need for the expression to be more than merely an
abstract.
Interestingly, ER computations, at least in Winpoker, are not based
on any kinds of multiple long term runs. Still, ER assumes both
perfect play and a sample size large enough for randomness to "sort
itself out." So, assuming perfect play, and assuming that a larger
sample size increases confidence in outcome, both requirements are
met. imho, it's as good a definition of long term as any other.
lb
And just how do you suggest (in your most simplistic way, of
course)
that one determine that "point"?....
I think when the original poster poised the question, "how many
hands in the long run?" they wanted a more numerical answer.
My simplistic answer is that long term is reflected in standard
deviation. Since ER is based on perfect play over the theoretical
long term, at what point does sd satisfy you that play will result
in
a win (or that results will at least fall within a range that
makes a
win most probable)? Simplistically, that is the long term. And it
satisfies the need for the expression to be more than merely an
abstract.Interestingly, ER computations, at least in Winpoker, are not
based
on any kinds of multiple long term runs. Still, ER assumes both
perfect play and a sample size large enough for randomness
to "sort
--- In vpFREE@yahoogroups.com, "lawrenceboxer" <ljboxer@e...> wrote:
itself out." So, assuming perfect play, and assuming that a larger
sample size increases confidence in outcome, both requirements are
met. imho, it's as good a definition of long term as any other.lb
>
> And just how do you suggest (in your most simplistic way, of
course)
> that one determine that "point"?....
>