vpFREE2 Forums

What is the "long term"?

#1

One of the issues debated now and then on this board is the meaning
of "long term", as applied to VP. Singer mentioned the lack of a
definition in his review of Linda's book.

There may be no "perfect" definition, but this might work. The long
term is the number of games where a normal approximation applies ( to
within +/- 1 SD of the EV). In other words, 84% of the sessions of
that size will lie within +/-1 SD of the EV of the game. It may be
necessary to attach a "degree of confidence" (not sure). This is
similar to, but not identical, to NO (which includes CB and assumes
the CLT applies). We can't use the CLT formula to compute the long
term (circular reasoning).

If this is a reasonable definition, the next step is to find
the formula to compute this "number of games". Using this definition
the long term will vary from game to game, depending on each games
variance (and maybe some other variable?). For example, I'm sure
9/6JB has a much smaller "long term" than DDB.

Does anyone know what the formula is? I've looked, but can't find
it. Perhaps someone has a different definition? I know this is
an "academic" issue of little or no interest to many, but it would be
nice to agree on a formal definition. If we had a definition, we
could compute it for every game.

#2

I remember one of my high school teachers telling us "In the long run,
we're all dead." My brain starts to hurt when I start thinking about
things like a precise formula for determining "long term." However, as
you say, I'm sure that the topic is of some interest to the more
academic types in the Group.

···

--- In vpFREE@yahoogroups.com, "brumar_lv" <brumar_lv@...> wrote:

One of the issues debated now and then on this board is the meaning
of "long term", as applied to VP.

#3

>
> One of the issues debated now and then on this board is the

meaning

> of "long term", as applied to VP.

I remember one of my high school teachers telling us "In the long

run,

we're all dead." My brain starts to hurt when I start thinking

about

things like a precise formula for determining "long term." However,

as

you say, I'm sure that the topic is of some interest to the more
academic types in the Group.

This board has a "philosophy' that VP has a mathematical basis, and
many of us commit thousands of $$$ based on this concept. Even the
casinos rely on the math too, especially over the "long run". So I
find it a tad embarassing (or should I say frustrating) the group has
not come come up with a good definition, nor any of the VP books for
that matter. Until we come up with an acceptable definition we are
an easy target for Rob. My gut feeling is the formula is out
there already (in some stat book) but not in a VP context. Why?
Because every VP game has a known mean (EV) and Variance.

···

--- In vpFREE@yahoogroups.com, "John" <bigjohnl@...> wrote:

--- In vpFREE@yahoogroups.com, "brumar_lv" <brumar_lv@> wrote:

#4

brumar, defining "longterm" is like defining "big". What is "big"? Is
a grapefruit big? How about a basketball? How about Earth? Jupiter?
The Solar System? The Milky Way? You'll notice that each question
makes the answer to the previous question a decided "no".

Longterm is a relative term, just like "big". Some might consider "NO"
hands to be longterm, while others would think various multiples
of "NO" are more appropriate. Still others might prefer to
define "longterm" as some specified (high) probability of being within
some specified (small) percentage of EV. There is not going to be a
single mathematical definition of "longterm".

--Dunbar

···

--- In vpFREE@yahoogroups.com, "brumar_lv" <brumar_lv@...> wrote:

This board has a "philosophy' that VP has a mathematical basis, and
many of us commit thousands of $$$ based on this concept. Even the
casinos rely on the math too, especially over the "long run". So I
find it a tad embarassing (or should I say frustrating) the group has
not come come up with a good definition, nor any of the VP books for
that matter. Until we come up with an acceptable definition we are
an easy target for Rob. My gut feeling is the formula is out
there already (in some stat book) but not in a VP context. Why?
Because every VP game has a known mean (EV) and Variance.

#5

Brumar_lv,
There was a discussion on this in late January of this year, but
maybe I can add another 2 cents worth. We calculated that N0 for
FPDW (100.76%) w/0 CB was 444,976 hands. Let's suppose that is what
someone plays in one year (about 2 hours/day, 600 hands/hr---seems
reasonable).
1) After 1 year of play, there would be a 15.87% (1sd) that you
would be losing money.

[Here we are using the normal distribution, which I think applies
after 1 N0. Note that on Jazbo's multiplay page, he says that
people misuse this---it would be really nice if he explained the
proper use. Now the sd is inversely proportional to the square root
of the number of hands played, so...]

2) After 4 years (not 2) of play, there would be a 2.28% (2sd) that
you would be losing money.

3) After 9 years of play, there would be a 0.13% (3sd) that you
would be losing money.

The numbers for Bonus Poker (99.17%) would be similar, but for
making money instead of losing money. I don't know of a definition
for long term. The formula for N0 is listed under bankroll links.
Jeff

#6

I remember one of my high school teachers telling us "In the long

run,

we're all dead."

John Maynard Keynes couldn't have said it better. ;}

···

--- In vpFREE@yahoogroups.com, "John" <bigjohnl@...> wrote:

#7

brumar_lv wrote:

This board has a "philosophy' that VP has a mathematical basis, and
many of us commit thousands of $$$ based on this concept. Even the
casinos rely on the math too, especially over the "long run". So I
find it a tad embarassing (or should I say frustrating) the group has
not come come up with a good definition, nor any of the VP books for
that matter.

You reject NO out of hand for a reason I can't fathom in the least and
propose your search for a "holy grail" definition that apparently no
one here has previously had either the ingenuity or brilliance to propose.

#8

Artmo replied:

···

--- In vpFREE@yahoogroups.com, "John" <bigjohnl@...> wrote:

I remember one of my high school teachers telling us "In the long

run,

we're all dead."

John Maynard Keynes couldn't have said it better. ;}

I would have to agree. I like the expression...

"Till death do us part."

Dennis
vp-connoisseur

[Non-text portions of this message have been removed]

#9

Hey Dubar;
I can't believe this. Big is the word I was going to compare it to. (Saw the question last night - thought I'd answer in the morning). Obviously crazed minds think alike.
There is no way to define long-term without some context and I'm not sure why there is so much concern about the definition of long-term, when we can get really detailed numbers about probabilities based on the number of hands played, bankroll, variance and so on. One of the best sources is your VP analyzer.
You can't get a precise answer to an ambiguous question.
Skip

Dunbar_dra wrote:

···

brumar, defining "longterm" is like defining "big". What is "big"? Is a grapefruit big? How about a basketball? How about Earth? Jupiter? The Solar System? The Milky Way? You'll notice that each question makes the answer to the previous question a decided "no".

Longterm is a relative term, just like "big". Some might consider "NO" hands to be longterm, while others would think various multiples of "NO" are more appropriate. Still others might prefer to define "longterm" as some specified (high) probability of being within some specified (small) percentage of EV. There is not going to be a single mathematical definition of "longterm".

--Dunbar

--- In vpFREE@yahoogroups.com, "brumar_lv" <brumar_lv@...> wrote:

This board has a "philosophy' that VP has a mathematical basis, and many of us commit thousands of $$$ based on this concept. Even the casinos rely on the math too, especially over the "long run". So I find it a tad embarassing (or should I say frustrating) the group has not come come up with a good definition, nor any of the VP books for that matter. Until we come up with an acceptable definition we are an easy target for Rob. My gut feeling is the formula is out there already (in some stat book) but not in a VP context. Why? Because every VP game has a known mean (EV) and Variance.

vpFREE Links: http://members.cox.net/vpfree/Links.htm

Yahoo! Groups Links

---------------------------------------------------------------------------------------------------
Text inserted by Platinum 2006:

This message has NOT been classified as spam. If it is unsolicited mail (spam), click on the following link to reclassify it: http://127.0.0.1:6083/Panda?ID=pav_1493&SPAM=true
---------------------------------------------------------------------------------------------------

--
Thanks!
Skip
http://www.vpinsider.com

#10

assuming a normal distribution is reasonably accurate for the core
+/-1sd, the errors are out on the right side tail

if you don't want to assume a normal distribution, you'd have to run a
pdf sim which involves matrix multiplication or the fourier transform

···

--- In vpFREE@yahoogroups.com, "brumar_lv" <brumar_lv@...> wrote:

One of the issues debated now and then on this board is the meaning
of "long term", as applied to VP. Singer mentioned the lack of a
definition in his review of Linda's book.

There may be no "perfect" definition, but this might work. The long
term is the number of games where a normal approximation applies ( to
within +/- 1 SD of the EV). In other words, 84% of the sessions of
that size will lie within +/-1 SD of the EV of the game. It may be
necessary to attach a "degree of confidence" (not sure). This is
similar to, but not identical, to NO (which includes CB and assumes
the CLT applies). We can't use the CLT formula to compute the long
term (circular reasoning).

If this is a reasonable definition, the next step is to find
the formula to compute this "number of games". Using this definition
the long term will vary from game to game, depending on each games
variance (and maybe some other variable?). For example, I'm sure
9/6JB has a much smaller "long term" than DDB.

Does anyone know what the formula is? I've looked, but can't find
it. Perhaps someone has a different definition? I know this is
an "academic" issue of little or no interest to many, but it would be
nice to agree on a formal definition. If we had a definition, we
could compute it for every game.

#11

Hey Dubar;
There is no way to define long-term without some context and I'm

not

sure why there is so much concern about the definition of long-

term,

when we can get really detailed numbers about probabilities based

on the

number of hands played, bankroll, variance and so on. One of the

best

sources is your VP analyzer.
You can't get a precise answer to an ambiguous question.
Skip

The definition I was suggesting is the point where all the detailed
numbers about probablilities have a firm mathematical foundation
based on the normal distribution and CLT. At what point (how many
games) does that happen? Surely it varies from game to game,
depending on the variance and possibly other factors.

Stat books are full of examples on how to compute "sample
sizes" to get a specified level of confidence in an estimate about
the population based on a sample. For VP we know all about the
population in advance. So I figured there must be a way to use the
these population parameters to compute the smallest sample size
where its valid to apply the normal distribution concepts. My
suggestion is that the long run, by definition, be that number of
games.

As some have noted, however, there may not be a way to do a
calculation like this, and even if there is, people will have
different opinions as to its validity. It just seems fairly
reasonable to use the point where the normal approximations can be
used would be a good definition for long term.

I hope nobody is upset about this. I realize a lot of useful methods
already exist, like NO, ROR calculations, etc. It just seemed it
might be possible to add another facet to these useful concepts by
coming up with a basic definition of "long term" for VP.

···

--- In vpFREE@yahoogroups.com, Skip Hughes <skiphughes@...> wrote:

#12

I disagree.

The point where the normal distribution can be used depends on how far
from the mean you are looking. See Central Limit Theorem.

The ROR number doesn't depend on assuming a normal distribution
approximation at all.

The N0 number only depends on -1SD which is fairly close even for just
4 royal cycles. And technically, you don't need the normal
approximation to calculate N0. N0 can simply be defined as the number
of hands required to get an 84% chance of winning.

If you want more, check the Nevada Gaming Regs, I think they specify
Chi-square testing. This could be a measure of the long term, but I
don't think it's a particularly useful measure for a gambler.

···

--- In vpFREE@yahoogroups.com, "brumar_lv" <brumar_lv@...> wrote:

It just seems fairly
reasonable to use the point where the normal approximations can be
used would be a good definition for long term.