suggestion for a column topic from one of the experts: Penalty Cards
Situation That Increase EV <i>and</i> Chance of Hitting a Royal.
penalty cards
LHOOQ wrote:
suggestion for a column topic from one of the experts: Penalty Cards
Situation That Increase EV <i>and</i> Chance of Hitting a Royal.
I'd suggest, based on recent discussions here, that the rationale for
factoring penalty cards into play doesn't relate to either of these
aspects of play.
- H.
LHOOQ wrote:
suggestion for a column topic from one of the experts: Penalty Cards
Situation That Increase EV <i>and</i> Chance of Hitting a Royal.
Harry responded: I'd suggest, based on recent discussions here, that
the rationale for factoring penalty cards into play doesn't relate to
either of these aspects of play.
I disagree. Every penalty card correctly factored increases EV over a
penalty-free strategy. The question is "how much?". The numbers bandied
about here recently suggest 0.001%-0.002%. I don't have the tools to
seriously debate those figures at this point, but I expect to soon. I
believe that in many games, penalty cards contribute significantly more
than this. But even if it gets to 0.01% in some cases, it's a pretty
small factor for most players.
Many penalty card effect the chance of hitting a royal --- often
DECREASING that chance, but sometimes INCREASING that chance.
In 9/6 Jacks, for example, from "AQT6" J, for example (where the quote
marks signify suitedness), it's correct to hold the 4-card flush over
the 3-card royal. That will eliminate the chance for a royal on this
hand, rather than giving you a 1-in-1081 chance.
In FPDW, for example, from "KT"875, penalty-free strategy says to toss
the hand and accurate strategy tells you to keep the "KT" (and many
proponents of a penalty free strategy use a simplified "no straight or
flush penalty" rule for this particular type of hand.) This is a case of
simple penalty cards INCREASING the chances for a royal. In the same
game, on a hand like "KJ"986, penalty-free strategy tells you to pitch
the hand, the simple penalty rule says to pitch the hand, but "appendix
level" exceptions tell you to keep the "KJ" --- which allows you a
chance for a royal, which would be denied by both a penalty-free
strategy and simplified penalty considerations.
My basic argument for studying penalty cards (basically, the more you
study penalty cards the more you know all aspects of the game)
DEFINITELY affects EV because it eliminates a lot of the errors that
come from not knowing the penalty-free strategy well enough. These are
errors that are largely assumed away by proponents of penalty-free
strategies, but assuming them away in arguments doesn't make them
disappear from your game. Whether eliminating these errors affect your
chances for a royal or not depends on what kind of errors you were
making.
Although there has been a lot of recent discussion of the chance of
going broke before hitting a royal (which is a useful thing to look at
--- it definitely is a part of the game to study), hitting a single
royal flush is not such a big deal. Most players who play more than
casually, hit LOTS of royals over their career, and however many they
hit, they ALWAYS want more. The key to success at video poker (at least
for those with a profit-max sort of goal) is losing the minimum BETWEEN
royals. Increasing the rate at which royals occur is not a costless
process.
Bob Dancer
For the best in video poker information, visit www.bobdancer.com
or call 1-800-244-2224 M-F 9-5 Pacific Time.
[Non-text portions of this message have been removed]
LHOOQ wrote:
> suggestion for a column topic from one of the experts: Penalty
> Cards Situation That Increase EV <i>and</i> Chance of Hitting a
> Royal.
Harry responded: I'd suggest, based on recent discussions here,
that the rationale for factoring penalty cards into play doesn't
relate to either of these aspects of play.
Bob Dancer replied;
I disagree. Every penalty card correctly factored increases EV over
a penalty-free strategy. The question is "how much?". The numbers
bandied about here recently suggest 0.001%-0.002%. I don't have the
tools to seriously debate those figures at this point, but I expect
to soon. I believe that in many games, penalty cards contribute
significantly more than this.Many penalty card effect the chance of hitting a royal --- often
DECREASING that chance, but sometimes INCREASING that chance.My basic argument for studying penalty cards (basically, the more
you study penalty cards the more you know all aspects of the game)
DEFINITELY affects EV because it eliminates a lot of the errors that
come from not knowing the penalty-free strategy well enough.
Except for modest difference in nuance, I'd suggest we're in
agreement. I'd hope it was clear that I'd never suggest that penalty
cards have no impact on ER or Royal probability (surely you credit me
that much 
A strategy that factors penalties affects strategy ER
and hand probabilities, at least nominally.
But, as worded, what I suggest as being the primary argument
(RATIONALE) for factoring penalties into play is the same you identify
-- it hones basic strategy play and makes a signficant difference in
accuracy and the true ER of actual play (accounting for play errors).
I don't think in your statement that playing penalty situations
correctly, alone, is what you're referencing by the phrase "eliminates
a lot of the errors". Instead, I believe you're suggesting that given
most players don't assert they're 100% accuracy in actual play, those
who've studied penalty-related considerations are likely to more
closely approach 100% given additional game scrutiny. But, in truth,
I have considerable difficulty in ascertaining this from your posts.
The key to success at video poker (at least for those with a
profit-max sort of goal) is losing the minimum BETWEEN royals.
Increasing the rate at which royals occur is not a costless process.
Now, here's where I'm going to disagree with you. (And it wouldn't be
surprising if Steve Jacobs weighs in.)
A strategy to minimize loss between royals will generally extend the
royal cycle and will reduce ER. You have greater overall profit over
the course of the royal cycle, but because that profit extends for
more hands, it results in a smaller profit per hand (and EV per hand).
I won't touch on the magnitude of differences here since it might be
argued that they're relatively immaterial in the course of standard
play.
For an adequately bankrolled player, I'd assert that the goal of the
successful player should be to maximize ER.
There are exceptions to the desirability of a max-ER strategy. One is
that if you approach a very attractive progressive that presents a ROR
considerably higher than your standard play (say one at a higher
denomination, or one with a base ER excluding the royal that is
significantly inferior), a min-loss between royal strategy (which you
reference here) may be more desirable.
I acknowledge Steve for the education reflected here (and I'm waiting
correction/refinement ;).
- Harry
I wrote: > The key to success at video poker (at least for those with a
profit-max sort of goal) is losing the minimum BETWEEN royals.
Increasing the rate at which royals occur is not a costless process.
Harry responded: Now, here's where I'm going to disagree with you. (And
it wouldn't be surprising if Steve Jacobs weighs in.)
And well you should disagree. I wasn't speaking precisely. I was
speaking in general terms comparing max-EV strategy (which I support)
and a "get the royals quickest" strategy (which I don't support). I
meant to say the first strategy has a lower cost between royals than the
second. I did not mean to say that a strategy devised to reduce loss
between royals would be superior to a max-EV strategy.
I too have learned from Steve and hope to learn more.
And in general, Harry, I agree that we are mostly on the same page. And
thanx for the nice words on the recent article, and I forgive you for
switching rodents on me.
Bob Dancer
For the best in video poker information, visit www.bobdancer.com
or call 1-800-244-2224 M-F 9-5 Pacific Time.
[Non-text portions of this message have been removed]
Bob, if you truly believe that this is the "key to success at video poker"
then you should promote a "Min-Cost-Royal" strategy instead of the
max-EV strategy. For 9/6 JoB, the average loss between royals is
984.22 units, while the Min-Cost-Royal strategy reduces the average
loss to 975.99 units.
The strategy that minimizes the average loss between royals can be
found by pretending that the royal payoff is exactly large enough to
give a breakeven game. For 9/6 JoB this value is a tiny bit less than
976 units, and the resulting strategy reduces the royal cycle from
40390.55 to 35939.23, and reduces overall ER to 99.5103%.
Of course, this strategy will have its own set of penalty card situations.
···
On Tuesday 06 December 2005 11:28 pm, Bob Dancer wrote:
The key to success at video poker (at least
for those with a profit-max sort of goal) is losing the minimum BETWEEN
royals. Increasing the rate at which royals occur is not a costless
process.
Now, here's where I'm going to disagree with you. (And it wouldn't be
surprising if Steve Jacobs weighs in.)
Yah, sure.
A strategy to minimize loss between royals will generally extend the
royal cycle and will reduce ER.
Correct about reducing ER, of course, since any deviation from max-EV
strategy will reduce ER. The royal cycle isn't necessarily extended,
that only happens for those rare games that have ER > 100%. For
unfavorable games, the Min-Cost-Royal strategy will reduce the royal
cycle.
You have greater overall profit over
the course of the royal cycle, but because that profit extends for
more hands, it results in a smaller profit per hand (and EV per hand).
I won't touch on the magnitude of differences here since it might be
argued that they're relatively immaterial in the course of standard
play.For an adequately bankrolled player, I'd assert that the goal of the
successful player should be to maximize ER.
Standard disagreement on my part: the goal of the successful player
is whatever that particular player wants to achieve. If that means losing
the least between royals, then he/she should play Min-Cost-Royal
strategy. If he/she wants to maximize the probability of hitting a royal
before losing the current bankroll, then Min-RoRBR is the proper strategy.
If he/she wants to maximize ER then max-EV is correct.
There are exceptions to the desirability of a max-ER strategy. One is
that if you approach a very attractive progressive that presents a ROR
considerably higher than your standard play (say one at a higher
denomination, or one with a base ER excluding the royal that is
significantly inferior), a min-loss between royal strategy (which you
reference here) may be more desirable.
I don't think this is rational, because it isn't consistent. If ER is what
you truly want, then you should always strive to maximize ER, whether
you have a massive bankroll or you're down to your last unit. Similarly,
if one truly wants to maximize the probability of hitting a royal before
going broke, then one should play Min-RoRBR strategy at all times.
And, if Bob truly believes that minimizing losses between royals is the
key to success, then he should use the Min-Cost-Royal strategy.
There are some goals that require the strategy to change as the
bankroll fluctuates, but max-EV and Min-RoRBR and Min-Cost-Royal
strategies are all completely independent of bankroll size.
···
On Wednesday 07 December 2005 01:21 am, Harry Porter wrote:
I guess I have to disagree a wee bit with all of you (or at least it seems like a disagreement
at this time). While Steve considers non-MAX EV strategies ( and may go so far as to
propose that such stratgeis are appropriate for some players), I want to go farther: I don't
think "EV" per se is the only quantity-- or even the most important quanntity for us ( and
especially Bob's mythical pro-player ) to worry about (that is optimize).
I don't have a problem with the logic or math that folks are using-- it all seems quite
kosher-- my disagreement lies with what it all "means", how we tend to mis-intrepret the
statistical quantities when we don't understand their meaning (or haven't defined their
meaning), how we continue to argue about the numerical quantites rather than what the
quantites mean or say. Once we take a look at what these quantities mean, I think folks
may come to see the EV is not exactly what they think it is. Admittedly this is a difficult
subject to discuss in a tidy manner. So let me play a little fast and loose and use a story:
Assume Player 1 follows a strict max EV stratgey. And I mean strict, every penalty card.
Player 1 plays perfectly, just like a computer. (I'll call this player "Bob")
Now Player 2 comes along, sits down next to Bob, and plays a different Stratgey. Bob
knows its a different stratgey becuase he sees hands where he would have made a
different play. When Bob notices the different play, he finds that the EV of his play was
sometimes higher and sometimes the same as player 2's play. Since Bob plays (by
definition) the true max EV strategy, this other player must be playing a stratgey with a
lower EV.
So Bob says to Player 2, "you know you can do better if you use my strategy, since your
strategy isn't max-EV". Player 2 response, "Nope. I'm trying to get as many Royals as
possible at the lowest cost to me. My buddy says that if I get more Royals them him, he'll
give me a phat check. That's a lot of money to me. I think it is worth it for me to not play
max-EV. I am playing a max-RF strategy because with the potential of the phat prize, my
overall EV is higher with it than with the Max EV strategy" Player #2's logic seems to
make sense to Bob so he smiles, and each continues to play their stratgey.
But we are not done yet. Player 2's choice seems to make sense, but since the the terms
"my EV" (that is my EV from play) and the stratgies EV (that is, the EV from the stratgey,
the parent population EV or first moment) haven't been defined (we don't really know what
they "mean" or even if they should "equate"), we're not done. So I'd say we haven't done a
complete analysis of the situation. Let's start with EV, as an example. Most of would
define EV as the first moment of the PDF. In other words, we assume that EV is the simple
arithmetric average (in any units we choose). We tend to stop our analysis at this level
(without going deeper) since, knowing how to compuet EV, we feel confident that we
can-- and should-- use "EV" all over the place. But what are we using it for? The EV we
all love is the first moment of the "parent popluation" ( that is the theoretical return) and
it is the simple aritmeric average of our actual results. Are they the same thing? DO the
quantities have the same meaning? For VP, I already know the parent popluation mean is a
poor proxy for our actual returns. And I think most playerw would agree with that
statement. But How poor? That's a VERY important question, perhaps the key question
that provides meaning to the term "EV". The answer to that question is given by the PDF
(if the PDF is normal, the answer can be equivelently given by just the variance). The PDF
gives the actual spread of EV's that we would be expected to acheive in real play. We
know how to compute the PDF for any number of hands played, for any bankroll, and for
any (properly defined) strategy. And given the PDF, we can answer all kinds of rigorous
statistical questions, technical ones like what is the likelihood of so and so's results being
from this VP game (vs another), and more useful ones, like I've lost X, what is the
theortical likihood of this happening, and are my actual results consistant with the
strategy I think I am playing? [There are a lot of other PDF's we could compute, like the
PDF of the variance, etc, and many question we can answer with them]. So the PDF can
help understand what EV means to us. Take JoB and FPDW as examples. FPDW is a
positive game, while JoB isn't. That means the EV of FPDW is always greater than for JoB--
for any number of hands. Does that mean the the we are more likely to win playing FPDW?
or less likely to lose? To answer these questions, you can look at the PDF's on Jazbo's
website (under volatility) or the CDFs. You might notice that the peaks in the PDFs don't
cluster around the theoretical EV (which isn't indicated, but is close to zero). the peak is
always to the "left" or the EV. That strongly suggests that EV (the first moment of the
parent population) is not a particualrly good indicatior of our real-play "EV". In other
words, the simple arithmetric average we compute for our results will tend not to cluster
about the theoretical EV. That is, what we mean by "EV for our own play" might is
something a bit different from what we mean when we say "theoritical EV". This issue is
nothing knew to statisticians: The statistical quantity (of the parent population) that
"best" (which needs to be rigoursly defined) approximates the average of our actual results
may not be the simple arithmetric mean. Indeed for a reasonable number of VP hands (in
the 10K's), most people would say the "mode" better approximates our actual results (look
at Jazbo's PDFs), not the "EV" or even the "median". Wow. EV is not king. But rather than
argue about what statistical quantity is best (if any!) , why not look at the entire PDF? It
contains everything we do know, so it must contain the answers we want if the answer is
knowable (strictly speaking this is not true, but it is true enough)
So, what we need to do is compare the PDF for player 1 ( Bob), the max EV guy, to the PDF
for player 2 (the other stratgey) and look for the differences. The next step is to decide
which differences are important to us and statistically significant (This can be done
rigorously since we know the parent PDF). We can do this as a function of # hands played,
or better yet, time (given hands/hour). We can then see the real effect of different
strategies such as playing with and without penalty cards. Case closed.
Conclusion: I beleive The PDF (or CDF) is the thing to look at. A change in strategy or
game may change the PDF. When choosing a game, I sometimes chose the game with the
lower EV becuase I like its PDF better: the "peak" of the PDF has a higher likely return (right
shifted is better) or the left "tail" is shifted to the right, or something else. Likewise, I
imagine that there are folks who choose a statgey in this same PDF based way, though the
differences in PDFs between strategies are often much smaller than those between games.
And of coarse, as others have pointed out, there are some out there who may (should?)
choose a strategy based upon its associated RoRBX. But IMHO, they ignore the PDF at
their own peril. We tend to throw around words like "superior" and "exact" and "better"
AND "my EV" as if WE all understand exactly (lol) what they mean and how they MUST be
used. There is nothing wrong with saying "this is the RoRBR of the lowest cost royal
strategy" or whatever (infact that may be valuable information) The problem apprears
when we say things like "playing strategy [x] is better than not playing [x]" without clearly
stating on what basis it is better and without demonstrating why (or how) that basis is
important (or statistically significant or meaningful, etc).
···
--- In vpFREE@yahoogroups.com, "Bob Dancer" <bob.dancer@c...> wrote:
I wrote: > The key to success at video poker (at least for those with a
> profit-max sort of goal) is losing the minimum BETWEEN royals.
> Increasing the rate at which royals occur is not a costless process.Harry responded: Now, here's where I'm going to disagree with you. (And
it wouldn't be surprising if Steve Jacobs weighs in.)And well you should disagree. I wasn't speaking precisely. I was
speaking in general terms comparing max-EV strategy (which I support)
and a "get the royals quickest" strategy (which I don't support). I
meant to say the first strategy has a lower cost between royals than the
second. I did not mean to say that a strategy devised to reduce loss
between royals would be superior to a max-EV strategy.Bob Dancer
Each time I see/read the topic on "penalty cards", I get this feeling
that I should write to Santa (Ho ho ho) and present my wish as
follows:
Dear Santa,
Could you please knock on IGT's door and ask the programmers/software
engineers to "Explain penalty cards in video poker games in details
and how much more a player can win if s/he fully understand the
correct HOLDS when they show up on the screen?".
Merry X'mas!
I guess I have to disagree a wee bit with all of you (or at least
it seems like a disagreement
at this time). While Steve considers non-MAX EV strategies ( and
may go so far as to
propose that such stratgeis are appropriate for some players), I
want to go farther: I don't
think "EV" per se is the only quantity-- or even the most important
quanntity for us ( and
especially Bob's mythical pro-player ) to worry about (that is
optimize).
I don't have a problem with the logic or math that folks are using--
it all seems quite
kosher-- my disagreement lies with what it all "means", how we tend
to mis-intrepret the
statistical quantities when we don't understand their meaning (or
haven't defined their
meaning), how we continue to argue about the numerical quantites
rather than what the
quantites mean or say. Once we take a look at what these
quantities mean, I think folks
may come to see the EV is not exactly what they think it is.
Admittedly this is a difficult
subject to discuss in a tidy manner. So let me play a little fast
and loose and use a story:
Assume Player 1 follows a strict max EV stratgey. And I mean
strict, every penalty card.
Player 1 plays perfectly, just like a computer. (I'll call this
player "Bob")
Now Player 2 comes along, sits down next to Bob, and plays a
different Stratgey. Bob
knows its a different stratgey becuase he sees hands where he would
have made a
different play. When Bob notices the different play, he finds that
the EV of his play was
sometimes higher and sometimes the same as player 2's play. Since
Bob plays (by
definition) the true max EV strategy, this other player must be
playing a stratgey with a
lower EV.
So Bob says to Player 2, "you know you can do better if you use my
strategy, since your
strategy isn't max-EV". Player 2 response, "Nope. I'm trying to
get as many Royals as
possible at the lowest cost to me. My buddy says that if I get
more Royals them him, he'll
give me a phat check. That's a lot of money to me. I think it is
worth it for me to not play
max-EV. I am playing a max-RF strategy because with the potential
of the phat prize, my
overall EV is higher with it than with the Max EV strategy"
Player #2's logic seems to
make sense to Bob so he smiles, and each continues to play their
stratgey.
But we are not done yet. Player 2's choice seems to make sense,
but since the the terms
"my EV" (that is my EV from play) and the stratgies EV (that is,
the EV from the stratgey,
the parent population EV or first moment) haven't been defined (we
don't really know what
they "mean" or even if they should "equate"), we're not done. So
I'd say we haven't done a
complete analysis of the situation. Let's start with EV, as an
example. Most of would
define EV as the first moment of the PDF. In other words, we
assume that EV is the simple
arithmetric average (in any units we choose). We tend to stop our
analysis at this level
(without going deeper) since, knowing how to compuet EV, we feel
confident that we
can-- and should-- use "EV" all over the place. But what are we
using it for? The EV we
all love is the first moment of the "parent popluation" ( that is
the theoretical return) and
it is the simple aritmeric average of our actual results. Are they
the same thing? DO the
quantities have the same meaning? For VP, I already know the parent
popluation mean is a
poor proxy for our actual returns. And I think most playerw would
agree with that
statement. But How poor? That's a VERY important question,
perhaps the key question
that provides meaning to the term "EV". The answer to that
question is given by the PDF
(if the PDF is normal, the answer can be equivelently given by
just the variance). The PDF
gives the actual spread of EV's that we would be expected to
acheive in real play. We
know how to compute the PDF for any number of hands played, for any
bankroll, and for
any (properly defined) strategy. And given the PDF, we can answer
all kinds of rigorous
statistical questions, technical ones like what is the likelihood
of so and so's results being
from this VP game (vs another), and more useful ones, like I've
lost X, what is the
theortical likihood of this happening, and are my actual results
consistant with the
strategy I think I am playing? [There are a lot of other PDF's we
could compute, like the
PDF of the variance, etc, and many question we can answer with
them]. So the PDF can
help understand what EV means to us. Take JoB and FPDW as
examples. FPDW is a
positive game, while JoB isn't. That means the EV of FPDW is
always greater than for JoB--
for any number of hands. Does that mean the the we are more likely
to win playing FPDW?
or less likely to lose? To answer these questions, you can look at
the PDF's on Jazbo's
website (under volatility) or the CDFs. You might notice that the
peaks in the PDFs don't
cluster around the theoretical EV (which isn't indicated, but is
close to zero). the peak is
always to the "left" or the EV. That strongly suggests that EV (the
first moment of the
parent population) is not a particualrly good indicatior of our
real-play "EV". In other
words, the simple arithmetric average we compute for our results
will tend not to cluster
about the theoretical EV. That is, what we mean by "EV for our own
play" might is
something a bit different from what we mean when we
say "theoritical EV". This issue is
nothing knew to statisticians: The statistical quantity (of the
parent population) that
"best" (which needs to be rigoursly defined) approximates the
average of our actual results
may not be the simple arithmetric mean. Indeed for a reasonable
number of VP hands (in
the 10K's), most people would say the "mode" better approximates
our actual results (look
at Jazbo's PDFs), not the "EV" or even the "median". Wow. EV is
not king. But rather than
argue about what statistical quantity is best (if any!) , why not
look at the entire PDF? It
contains everything we do know, so it must contain the answers we
want if the answer is
knowable (strictly speaking this is not true, but it is true enough)
So, what we need to do is compare the PDF for player 1 ( Bob), the
max EV guy, to the PDF
for player 2 (the other stratgey) and look for the differences.
The next step is to decide
which differences are important to us and statistically significant
(This can be done
rigorously since we know the parent PDF). We can do this as a
function of # hands played,
or better yet, time (given hands/hour). We can then see the real
effect of different
strategies such as playing with and without penalty cards. Case
closed.
Conclusion: I beleive The PDF (or CDF) is the thing to look at. A
change in strategy or
game may change the PDF. When choosing a game, I sometimes chose
the game with the
lower EV becuase I like its PDF better: the "peak" of the PDF has a
higher likely return (right
shifted is better) or the left "tail" is shifted to the right, or
something else. Likewise, I
imagine that there are folks who choose a statgey in this same PDF
based way, though the
differences in PDFs between strategies are often much smaller than
those between games.
And of coarse, as others have pointed out, there are some out there
who may (should?)
choose a strategy based upon its associated RoRBX. But IMHO, they
ignore the PDF at
their own peril. We tend to throw around words like "superior"
and "exact" and "better"
AND "my EV" as if WE all understand exactly (lol) what they mean
and how they MUST be
used. There is nothing wrong with saying "this is the RoRBR of
the lowest cost royal
strategy" or whatever (infact that may be valuable information)
The problem apprears
when we say things like "playing strategy [x] is better than not
playing [x]" without clearly
stating on what basis it is better and without demonstrating why
(or how) that basis is
important (or statistically significant or meaningful, etc).
>
> I wrote: > The key to success at video poker (at least for those
with a
> > profit-max sort of goal) is losing the minimum BETWEEN royals.
> > Increasing the rate at which royals occur is not a costless
process.
>
> Harry responded: Now, here's where I'm going to disagree with
you. (And
> it wouldn't be surprising if Steve Jacobs weighs in.)
>
>
> And well you should disagree. I wasn't speaking precisely. I was
> speaking in general terms comparing max-EV strategy (which I
support)
> and a "get the royals quickest" strategy (which I don't support).
I
> meant to say the first strategy has a lower cost between royals
than the
> second. I did not mean to say that a strategy devised to reduce
loss
···
--- In vpFREE@yahoogroups.com, "cdfsrule" <groups.yahoo@v...> wrote:
--- In vpFREE@yahoogroups.com, "Bob Dancer" <bob.dancer@c...> wrote:
> between royals would be superior to a max-EV strategy.
>
>
> Bob Dancer
>
Kelly bet, the no bust strategy. Just imagine the money Dancer could
have made if he used that instead!
http://members.cox.net/vpfree/Bank.htm
http://www.jazbo.com/videopoker/kelly.html
···
--- In vpFREE@yahoogroups.com, Steve Jacobs <jacobs@x...> wrote:
There are some goals that require the strategy to change as the
bankroll fluctuates, but max-EV and Min-RoRBR and Min-Cost-Royal
strategies are all completely independent of bankroll size.
Kelly betting is based on maximizing the expected value of the log of
bankroll. A true log-optimal strategy is a good example of a strategy
that varies with the number of units in the player's bankroll. When the
bankroll is very large, the log-optimal strategy becomes identical with
max-EV strategy. When the bankroll becomes small (one unit, for example)
the log-optimal strategy becomes identical with min-RoR strategy.
When the bankroll is just the right size to give a maximum bankroll
growth rate the log-optimal strategy is between max-EV and min-RoR,
yielding a strategy that has lower RoR than the max-EV strategy
but better EV than the min-RoR strategy.
···
On Wednesday 07 December 2005 12:23 pm, nightoftheiguana2000 wrote:
--- In vpFREE@yahoogroups.com, Steve Jacobs <jacobs@x...> wrote:
> There are some goals that require the strategy to change as the
> bankroll fluctuates, but max-EV and Min-RoRBR and Min-Cost-Royal
> strategies are all completely independent of bankroll size.Kelly bet, the no bust strategy. Just imagine the money Dancer could
have made if he used that instead!
If you are "adequately bankrolled" (what? 1% ror?) then you are
playing too conservatively, i.e. you are not achieving anywhere near
optimum bankroll growth. To get optimum bankroll growth, use Kelly
strategy.
···
--- In vpFREE@yahoogroups.com, "Harry Porter" <harry.porter@v...> wrote:
For an adequately bankrolled player, I'd assert that the goal of the
successful player should be to maximize ER.
frugal video poker will answer this question, and more
···
--- In vpFREE@yahoogroups.com, "gilbert_616" <gilbert_616@y...> wrote:
Each time I see/read the topic on "penalty cards", I get this feeling
that I should write to Santa (Ho ho ho) and present my wish as
follows:Dear Santa,
Could you please knock on IGT's door and ask the programmers/software
engineers to "Explain penalty cards in video poker games in details
and how much more a player can win if s/he fully understand the
correct HOLDS when they show up on the screen?".Merry X'mas!
--- In vpFREE@yahoogroups.com, "cdfsrule" <groups.yahoo@v...> wrote:
>
> I guess I have to disagree a wee bit with all of you (or at least
it seems like a disagreement
> at this time). While Steve considers non-MAX EV strategies ( and
may go so far as to
> propose that such stratgeis are appropriate for some players), I
want to go farther: I don't
> think "EV" per se is the only quantity-- or even the most important
quanntity for us ( and
> especially Bob's mythical pro-player ) to worry about (that is
optimize).
>
> I don't have a problem with the logic or math that folks are using--
it all seems quite
> kosher-- my disagreement lies with what it all "means", how we tend
to mis-intrepret the
> statistical quantities when we don't understand their meaning (or
haven't defined their
> meaning), how we continue to argue about the numerical quantites
rather than what the
> quantites mean or say. Once we take a look at what these
quantities mean, I think folks
> may come to see the EV is not exactly what they think it is.
Admittedly this is a difficult
> subject to discuss in a tidy manner. So let me play a little fast
and loose and use a story:
>
> Assume Player 1 follows a strict max EV stratgey. And I mean
strict, every penalty card.
> Player 1 plays perfectly, just like a computer. (I'll call this
player "Bob")
>
> Now Player 2 comes along, sits down next to Bob, and plays a
different Stratgey. Bob
> knows its a different stratgey becuase he sees hands where he would
have made a
> different play. When Bob notices the different play, he finds that
the EV of his play was
> sometimes higher and sometimes the same as player 2's play. Since
Bob plays (by
> definition) the true max EV strategy, this other player must be
playing a stratgey with a
> lower EV.
>
> So Bob says to Player 2, "you know you can do better if you use my
strategy, since your
> strategy isn't max-EV". Player 2 response, "Nope. I'm trying to
get as many Royals as
> possible at the lowest cost to me. My buddy says that if I get
more Royals them him, he'll
> give me a phat check. That's a lot of money to me. I think it is
worth it for me to not play
> max-EV. I am playing a max-RF strategy because with the potential
of the phat prize, my
> overall EV is higher with it than with the Max EV strategy"
Player #2's logic seems to
> make sense to Bob so he smiles, and each continues to play their
stratgey.
>
> But we are not done yet. Player 2's choice seems to make sense,
but since the the terms
> "my EV" (that is my EV from play) and the stratgies EV (that is,
the EV from the stratgey,
> the parent population EV or first moment) haven't been defined (we
don't really know what
> they "mean" or even if they should "equate"), we're not done. So
I'd say we haven't done a
> complete analysis of the situation. Let's start with EV, as an
example. Most of would
> define EV as the first moment of the PDF. In other words, we
assume that EV is the simple
> arithmetric average (in any units we choose). We tend to stop our
analysis at this level
> (without going deeper) since, knowing how to compuet EV, we feel
confident that we
> can-- and should-- use "EV" all over the place. But what are we
using it for? The EV we
> all love is the first moment of the "parent popluation" ( that is
the theoretical return) and
> it is the simple aritmeric average of our actual results. Are they
the same thing? DO the
> quantities have the same meaning? For VP, I already know the parent
popluation mean is a
> poor proxy for our actual returns. And I think most playerw would
agree with that
> statement. But How poor? That's a VERY important question,
perhaps the key question
> that provides meaning to the term "EV". The answer to that
question is given by the PDF
> (if the PDF is normal, the answer can be equivelently given by
just the variance). The PDF
> gives the actual spread of EV's that we would be expected to
acheive in real play. We
> know how to compute the PDF for any number of hands played, for any
bankroll, and for
> any (properly defined) strategy. And given the PDF, we can answer
all kinds of rigorous
> statistical questions, technical ones like what is the likelihood
of so and so's results being
> from this VP game (vs another), and more useful ones, like I've
lost X, what is the
> theortical likihood of this happening, and are my actual results
consistant with the
> strategy I think I am playing? [There are a lot of other PDF's we
could compute, like the
> PDF of the variance, etc, and many question we can answer with
them]. So the PDF can
> help understand what EV means to us. Take JoB and FPDW as
examples. FPDW is a
> positive game, while JoB isn't. That means the EV of FPDW is
always greater than for JoB--
> for any number of hands. Does that mean the the we are more likely
to win playing FPDW?
> or less likely to lose? To answer these questions, you can look at
the PDF's on Jazbo's
> website (under volatility) or the CDFs. You might notice that the
peaks in the PDFs don't
> cluster around the theoretical EV (which isn't indicated, but is
close to zero). the peak is
> always to the "left" or the EV. That strongly suggests that EV (the
first moment of the
> parent population) is not a particualrly good indicatior of our
real-play "EV". In other
> words, the simple arithmetric average we compute for our results
will tend not to cluster
> about the theoretical EV. That is, what we mean by "EV for our own
play" might is
> something a bit different from what we mean when we
say "theoritical EV". This issue is
> nothing knew to statisticians: The statistical quantity (of the
parent population) that
> "best" (which needs to be rigoursly defined) approximates the
average of our actual results
> may not be the simple arithmetric mean. Indeed for a reasonable
number of VP hands (in
> the 10K's), most people would say the "mode" better approximates
our actual results (look
> at Jazbo's PDFs), not the "EV" or even the "median". Wow. EV is
not king. But rather than
> argue about what statistical quantity is best (if any!) , why not
look at the entire PDF? It
> contains everything we do know, so it must contain the answers we
want if the answer is
> knowable (strictly speaking this is not true, but it is true enough)
>
> So, what we need to do is compare the PDF for player 1 ( Bob), the
max EV guy, to the PDF
> for player 2 (the other stratgey) and look for the differences.
The next step is to decide
> which differences are important to us and statistically significant
(This can be done
> rigorously since we know the parent PDF). We can do this as a
function of # hands played,
> or better yet, time (given hands/hour). We can then see the real
effect of different
> strategies such as playing with and without penalty cards. Case
closed.
>
> Conclusion: I beleive The PDF (or CDF) is the thing to look at. A
change in strategy or
> game may change the PDF. When choosing a game, I sometimes chose
the game with the
> lower EV becuase I like its PDF better: the "peak" of the PDF has a
higher likely return (right
> shifted is better) or the left "tail" is shifted to the right, or
something else. Likewise, I
> imagine that there are folks who choose a statgey in this same PDF
based way, though the
> differences in PDFs between strategies are often much smaller than
those between games.
> And of coarse, as others have pointed out, there are some out there
who may (should?)
> choose a strategy based upon its associated RoRBX. But IMHO, they
ignore the PDF at
> their own peril. We tend to throw around words like "superior"
and "exact" and "better"
> AND "my EV" as if WE all understand exactly (lol) what they mean
and how they MUST be
> used. There is nothing wrong with saying "this is the RoRBR of
the lowest cost royal
> strategy" or whatever (infact that may be valuable information)
The problem apprears
> when we say things like "playing strategy [x] is better than not
playing [x]" without clearly
> stating on what basis it is better and without demonstrating why
(or how) that basis is
> important (or statistically significant or meaningful, etc).
>
> --- In vpFREE@yahoogroups.com, "Bob Dancer" <bob.dancer@c...> wrote:
> >
> > I wrote: > The key to success at video poker (at least for those
with a
> > > profit-max sort of goal) is losing the minimum BETWEEN royals.
> > > Increasing the rate at which royals occur is not a costless
process.
> >
> > Harry responded: Now, here's where I'm going to disagree with
you. (And
> > it wouldn't be surprising if Steve Jacobs weighs in.)
> >
> >
> > And well you should disagree. I wasn't speaking precisely. I was
> > speaking in general terms comparing max-EV strategy (which I
support)
> > and a "get the royals quickest" strategy (which I don't support).
I
> > meant to say the first strategy has a lower cost between royals
than the
> > second. I did not mean to say that a strategy devised to reduce
loss
> > between royals would be superior to a max-EV strategy.
> >
> >
> > Bob Dancer
> >
>
for example, perfect 9/6 JOB strategy, found here:
http://wizardofodds.com/jacksorbetter
has er=0.9954, var=19.78
spreadsheet wizard's numbers:
er=0.995439044, var=19.51467643
this simple strategy:
5SF>4RF>PAT>4SF>HP>3RF>4FL>KQJT>LP>4STo>3SF0>AKQJ>2RF2H>4STi3H>3SF1>KQJ>2H>JT>QT>1H>3SF2
has er=0.9954, var=19.40
--- In vpFREE@yahoogroups.com, "nightoftheiguana2000"
<nightoftheiguana2000@y...> wrote:
···
frugal video poker will answer this question, and more
--- In vpFREE@yahoogroups.com, "gilbert_616" <gilbert_616@y...> wrote:
>
> Each time I see/read the topic on "penalty cards", I get this feeling
> that I should write to Santa (Ho ho ho) and present my wish as
> follows:
>
> Dear Santa,
>
> Could you please knock on IGT's door and ask the programmers/software
> engineers to "Explain penalty cards in video poker games in details
> and how much more a player can win if s/he fully understand the
> correct HOLDS when they show up on the screen?".
>
> Merry X'mas!
>
>
> --- In vpFREE@yahoogroups.com, "cdfsrule" <groups.yahoo@v...> wrote:
> >
> > I guess I have to disagree a wee bit with all of you (or at least
> it seems like a disagreement
> > at this time). While Steve considers non-MAX EV strategies ( and
> may go so far as to
> > propose that such stratgeis are appropriate for some players), I
> want to go farther: I don't
> > think "EV" per se is the only quantity-- or even the most important
> quanntity for us ( and
> > especially Bob's mythical pro-player ) to worry about (that is
> optimize).
> >
> > I don't have a problem with the logic or math that folks are using--
> it all seems quite
> > kosher-- my disagreement lies with what it all "means", how we tend
> to mis-intrepret the
> > statistical quantities when we don't understand their meaning (or
> haven't defined their
> > meaning), how we continue to argue about the numerical quantites
> rather than what the
> > quantites mean or say. Once we take a look at what these
> quantities mean, I think folks
> > may come to see the EV is not exactly what they think it is.
> Admittedly this is a difficult
> > subject to discuss in a tidy manner. So let me play a little fast
> and loose and use a story:
> >
> > Assume Player 1 follows a strict max EV stratgey. And I mean
> strict, every penalty card.
> > Player 1 plays perfectly, just like a computer. (I'll call this
> player "Bob")
> >
> > Now Player 2 comes along, sits down next to Bob, and plays a
> different Stratgey. Bob
> > knows its a different stratgey becuase he sees hands where he would
> have made a
> > different play. When Bob notices the different play, he finds that
> the EV of his play was
> > sometimes higher and sometimes the same as player 2's play. Since
> Bob plays (by
> > definition) the true max EV strategy, this other player must be
> playing a stratgey with a
> > lower EV.
> >
> > So Bob says to Player 2, "you know you can do better if you use my
> strategy, since your
> > strategy isn't max-EV". Player 2 response, "Nope. I'm trying to
> get as many Royals as
> > possible at the lowest cost to me. My buddy says that if I get
> more Royals them him, he'll
> > give me a phat check. That's a lot of money to me. I think it is
> worth it for me to not play
> > max-EV. I am playing a max-RF strategy because with the potential
> of the phat prize, my
> > overall EV is higher with it than with the Max EV strategy"
> Player #2's logic seems to
> > make sense to Bob so he smiles, and each continues to play their
> stratgey.
> >
> > But we are not done yet. Player 2's choice seems to make sense,
> but since the the terms
> > "my EV" (that is my EV from play) and the stratgies EV (that is,
> the EV from the stratgey,
> > the parent population EV or first moment) haven't been defined (we
> don't really know what
> > they "mean" or even if they should "equate"), we're not done. So
> I'd say we haven't done a
> > complete analysis of the situation. Let's start with EV, as an
> example. Most of would
> > define EV as the first moment of the PDF. In other words, we
> assume that EV is the simple
> > arithmetric average (in any units we choose). We tend to stop our
> analysis at this level
> > (without going deeper) since, knowing how to compuet EV, we feel
> confident that we
> > can-- and should-- use "EV" all over the place. But what are we
> using it for? The EV we
> > all love is the first moment of the "parent popluation" ( that is
> the theoretical return) and
> > it is the simple aritmeric average of our actual results. Are they
> the same thing? DO the
> > quantities have the same meaning? For VP, I already know the parent
> popluation mean is a
> > poor proxy for our actual returns. And I think most playerw would
> agree with that
> > statement. But How poor? That's a VERY important question,
> perhaps the key question
> > that provides meaning to the term "EV". The answer to that
> question is given by the PDF
> > (if the PDF is normal, the answer can be equivelently given by
> just the variance). The PDF
> > gives the actual spread of EV's that we would be expected to
> acheive in real play. We
> > know how to compute the PDF for any number of hands played, for any
> bankroll, and for
> > any (properly defined) strategy. And given the PDF, we can answer
> all kinds of rigorous
> > statistical questions, technical ones like what is the likelihood
> of so and so's results being
> > from this VP game (vs another), and more useful ones, like I've
> lost X, what is the
> > theortical likihood of this happening, and are my actual results
> consistant with the
> > strategy I think I am playing? [There are a lot of other PDF's we
> could compute, like the
> > PDF of the variance, etc, and many question we can answer with
> them]. So the PDF can
> > help understand what EV means to us. Take JoB and FPDW as
> examples. FPDW is a
> > positive game, while JoB isn't. That means the EV of FPDW is
> always greater than for JoB--
> > for any number of hands. Does that mean the the we are more likely
> to win playing FPDW?
> > or less likely to lose? To answer these questions, you can look at
> the PDF's on Jazbo's
> > website (under volatility) or the CDFs. You might notice that the
> peaks in the PDFs don't
> > cluster around the theoretical EV (which isn't indicated, but is
> close to zero). the peak is
> > always to the "left" or the EV. That strongly suggests that EV (the
> first moment of the
> > parent population) is not a particualrly good indicatior of our
> real-play "EV". In other
> > words, the simple arithmetric average we compute for our results
> will tend not to cluster
> > about the theoretical EV. That is, what we mean by "EV for our own
> play" might is
> > something a bit different from what we mean when we
> say "theoritical EV". This issue is
> > nothing knew to statisticians: The statistical quantity (of the
> parent population) that
> > "best" (which needs to be rigoursly defined) approximates the
> average of our actual results
> > may not be the simple arithmetric mean. Indeed for a reasonable
> number of VP hands (in
> > the 10K's), most people would say the "mode" better approximates
> our actual results (look
> > at Jazbo's PDFs), not the "EV" or even the "median". Wow. EV is
> not king. But rather than
> > argue about what statistical quantity is best (if any!) , why not
> look at the entire PDF? It
> > contains everything we do know, so it must contain the answers we
> want if the answer is
> > knowable (strictly speaking this is not true, but it is true enough)
> >
> > So, what we need to do is compare the PDF for player 1 ( Bob), the
> max EV guy, to the PDF
> > for player 2 (the other stratgey) and look for the differences.
> The next step is to decide
> > which differences are important to us and statistically significant
> (This can be done
> > rigorously since we know the parent PDF). We can do this as a
> function of # hands played,
> > or better yet, time (given hands/hour). We can then see the real
> effect of different
> > strategies such as playing with and without penalty cards. Case
> closed.
> >
> > Conclusion: I beleive The PDF (or CDF) is the thing to look at. A
> change in strategy or
> > game may change the PDF. When choosing a game, I sometimes chose
> the game with the
> > lower EV becuase I like its PDF better: the "peak" of the PDF has a
> higher likely return (right
> > shifted is better) or the left "tail" is shifted to the right, or
> something else. Likewise, I
> > imagine that there are folks who choose a statgey in this same PDF
> based way, though the
> > differences in PDFs between strategies are often much smaller than
> those between games.
> > And of coarse, as others have pointed out, there are some out there
> who may (should?)
> > choose a strategy based upon its associated RoRBX. But IMHO, they
> ignore the PDF at
> > their own peril. We tend to throw around words like "superior"
> and "exact" and "better"
> > AND "my EV" as if WE all understand exactly (lol) what they mean
> and how they MUST be
> > used. There is nothing wrong with saying "this is the RoRBR of
> the lowest cost royal
> > strategy" or whatever (infact that may be valuable information)
> The problem apprears
> > when we say things like "playing strategy [x] is better than not
> playing [x]" without clearly
> > stating on what basis it is better and without demonstrating why
> (or how) that basis is
> > important (or statistically significant or meaningful, etc).
> >
> > --- In vpFREE@yahoogroups.com, "Bob Dancer" <bob.dancer@c...> wrote:
> > >
> > > I wrote: > The key to success at video poker (at least for those
> with a
> > > > profit-max sort of goal) is losing the minimum BETWEEN royals.
> > > > Increasing the rate at which royals occur is not a costless
> process.
> > >
> > > Harry responded: Now, here's where I'm going to disagree with
> you. (And
> > > it wouldn't be surprising if Steve Jacobs weighs in.)
> > >
> > >
> > > And well you should disagree. I wasn't speaking precisely. I was
> > > speaking in general terms comparing max-EV strategy (which I
> support)
> > > and a "get the royals quickest" strategy (which I don't support).
> I
> > > meant to say the first strategy has a lower cost between royals
> than the
> > > second. I did not mean to say that a strategy devised to reduce
> loss
> > > between royals would be superior to a max-EV strategy.
> > >
> > >
> > > Bob Dancer
> > >
> >
>
Pardon me but, I didn't see anything in your reply about "penalty
cards" and it doesn't show that the wizardofodds.com has people there
who used to work and program for IGT's video poker games !? :>
--- In vpFREE@yahoogroups.com, "nightoftheiguana2000"
<nightoftheiguana2000@y...> wrote:
for example, perfect 9/6 JOB strategy, found here:
http://wizardofodds.com/jacksorbetter
has er=0.9954, var=19.78
spreadsheet wizard's numbers:
er=0.995439044, var=19.51467643
this simple strategy:
5SF>4RF>PAT>4SF>HP>3RF>4FL>KQJT>LP>4STo>3SF0>AKQJ>2RF2H>4STi3H>3SF1>KQ
2H>JT>QT>1H>3SF2
has er=0.9954, var=19.40--- In vpFREE@yahoogroups.com, "nightoftheiguana2000"
<nightoftheiguana2000@y...> wrote:
>
> frugal video poker will answer this question, and more
>
> --- In vpFREE@yahoogroups.com, "gilbert_616" <gilbert_616@y...>
wrote:
> >
> > Each time I see/read the topic on "penalty cards", I get this
feeling
> > that I should write to Santa (Ho ho ho) and present my wish as
> > follows:
> >
> > Dear Santa,
> >
> > Could you please knock on IGT's door and ask the
programmers/software
> > engineers to "Explain penalty cards in video poker games in
details
> > and how much more a player can win if s/he fully understand the
> > correct HOLDS when they show up on the screen?".
> >
> > Merry X'mas!
> >
> >
> > --- In vpFREE@yahoogroups.com, "cdfsrule" <groups.yahoo@v...>
wrote:
> > >
> > > I guess I have to disagree a wee bit with all of you (or at
least
> > it seems like a disagreement
> > > at this time). While Steve considers non-MAX EV strategies (
and
> > may go so far as to
> > > propose that such stratgeis are appropriate for some
players), I
> > want to go farther: I don't
> > > think "EV" per se is the only quantity-- or even the most
important
> > quanntity for us ( and
> > > especially Bob's mythical pro-player ) to worry about (that
is
> > optimize).
> > >
> > > I don't have a problem with the logic or math that folks are
using--
> > it all seems quite
> > > kosher-- my disagreement lies with what it all "means", how
we tend
> > to mis-intrepret the
> > > statistical quantities when we don't understand their meaning
(or
> > haven't defined their
> > > meaning), how we continue to argue about the numerical
quantites
> > rather than what the
> > > quantites mean or say. Once we take a look at what these
> > quantities mean, I think folks
> > > may come to see the EV is not exactly what they think it is.
> > Admittedly this is a difficult
> > > subject to discuss in a tidy manner. So let me play a little
fast
> > and loose and use a story:
> > >
> > > Assume Player 1 follows a strict max EV stratgey. And I mean
> > strict, every penalty card.
> > > Player 1 plays perfectly, just like a computer. (I'll call
this
> > player "Bob")
> > >
> > > Now Player 2 comes along, sits down next to Bob, and plays a
> > different Stratgey. Bob
> > > knows its a different stratgey becuase he sees hands where he
would
> > have made a
> > > different play. When Bob notices the different play, he finds
that
> > the EV of his play was
> > > sometimes higher and sometimes the same as player 2's play.
Since
> > Bob plays (by
> > > definition) the true max EV strategy, this other player must
be
> > playing a stratgey with a
> > > lower EV.
> > >
> > > So Bob says to Player 2, "you know you can do better if you
use my
> > strategy, since your
> > > strategy isn't max-EV". Player 2 response, "Nope. I'm
trying to
> > get as many Royals as
> > > possible at the lowest cost to me. My buddy says that if I
get
> > more Royals them him, he'll
> > > give me a phat check. That's a lot of money to me. I think
it is
> > worth it for me to not play
> > > max-EV. I am playing a max-RF strategy because with the
potential
> > of the phat prize, my
> > > overall EV is higher with it than with the Max EV
strategy"
> > Player #2's logic seems to
> > > make sense to Bob so he smiles, and each continues to play
their
> > stratgey.
> > >
> > > But we are not done yet. Player 2's choice seems to make
sense,
> > but since the the terms
> > > "my EV" (that is my EV from play) and the stratgies EV (that
is,
> > the EV from the stratgey,
> > > the parent population EV or first moment) haven't been
defined (we
> > don't really know what
> > > they "mean" or even if they should "equate"), we're not done.
So
> > I'd say we haven't done a
> > > complete analysis of the situation. Let's start with EV, as
an
> > example. Most of would
> > > define EV as the first moment of the PDF. In other words, we
> > assume that EV is the simple
> > > arithmetric average (in any units we choose). We tend to
stop our
> > analysis at this level
> > > (without going deeper) since, knowing how to compuet EV, we
feel
> > confident that we
> > > can-- and should-- use "EV" all over the place. But what
are we
> > using it for? The EV we
> > > all love is the first moment of the "parent popluation" (
that is
> > the theoretical return) and
> > > it is the simple aritmeric average of our actual results. Are
they
> > the same thing? DO the
> > > quantities have the same meaning? For VP, I already know the
parent
> > popluation mean is a
> > > poor proxy for our actual returns. And I think most playerw
would
> > agree with that
> > > statement. But How poor? That's a VERY important question,
> > perhaps the key question
> > > that provides meaning to the term "EV". The answer to that
> > question is given by the PDF
> > > (if the PDF is normal, the answer can be equivelently given
by
> > just the variance). The PDF
> > > gives the actual spread of EV's that we would be expected to
> > acheive in real play. We
> > > know how to compute the PDF for any number of hands played,
for any
> > bankroll, and for
> > > any (properly defined) strategy. And given the PDF, we can
answer
> > all kinds of rigorous
> > > statistical questions, technical ones like what is the
likelihood
> > of so and so's results being
> > > from this VP game (vs another), and more useful ones, like
I've
> > lost X, what is the
> > > theortical likihood of this happening, and are my actual
results
> > consistant with the
> > > strategy I think I am playing? [There are a lot of other
PDF's we
> > could compute, like the
> > > PDF of the variance, etc, and many question we can answer
with
> > them]. So the PDF can
> > > help understand what EV means to us. Take JoB and FPDW as
> > examples. FPDW is a
> > > positive game, while JoB isn't. That means the EV of FPDW is
> > always greater than for JoB--
> > > for any number of hands. Does that mean the the we are more
likely
> > to win playing FPDW?
> > > or less likely to lose? To answer these questions, you can
look at
> > the PDF's on Jazbo's
> > > website (under volatility) or the CDFs. You might notice
that the
> > peaks in the PDFs don't
> > > cluster around the theoretical EV (which isn't indicated, but
is
> > close to zero). the peak is
> > > always to the "left" or the EV. That strongly suggests that
EV (the
> > first moment of the
> > > parent population) is not a particualrly good indicatior of
our
> > real-play "EV". In other
> > > words, the simple arithmetric average we compute for our
results
> > will tend not to cluster
> > > about the theoretical EV. That is, what we mean by "EV for
our own
> > play" might is
> > > something a bit different from what we mean when we
> > say "theoritical EV". This issue is
> > > nothing knew to statisticians: The statistical quantity
(of the
> > parent population) that
> > > "best" (which needs to be rigoursly defined) approximates the
> > average of our actual results
> > > may not be the simple arithmetric mean. Indeed for a
reasonable
> > number of VP hands (in
> > > the 10K's), most people would say the "mode" better
approximates
> > our actual results (look
> > > at Jazbo's PDFs), not the "EV" or even the "median". Wow.
EV is
> > not king. But rather than
> > > argue about what statistical quantity is best (if any!) , why
not
> > look at the entire PDF? It
> > > contains everything we do know, so it must contain the
answers we
> > want if the answer is
> > > knowable (strictly speaking this is not true, but it is true
enough)
> > >
> > > So, what we need to do is compare the PDF for player 1 (
Bob), the
> > max EV guy, to the PDF
> > > for player 2 (the other stratgey) and look for the
differences.
> > The next step is to decide
> > > which differences are important to us and statistically
significant
> > (This can be done
> > > rigorously since we know the parent PDF). We can do this as
a
> > function of # hands played,
> > > or better yet, time (given hands/hour). We can then see the
real
> > effect of different
> > > strategies such as playing with and without penalty cards.
Case
> > closed.
> > >
> > > Conclusion: I beleive The PDF (or CDF) is the thing to look
at. A
> > change in strategy or
> > > game may change the PDF. When choosing a game, I sometimes
chose
> > the game with the
> > > lower EV becuase I like its PDF better: the "peak" of the PDF
has a
> > higher likely return (right
> > > shifted is better) or the left "tail" is shifted to the
right, or
> > something else. Likewise, I
> > > imagine that there are folks who choose a statgey in this
same PDF
> > based way, though the
> > > differences in PDFs between strategies are often much smaller
than
> > those between games.
> > > And of coarse, as others have pointed out, there are some out
there
> > who may (should?)
> > > choose a strategy based upon its associated RoRBX. But IMHO,
they
> > ignore the PDF at
> > > their own peril. We tend to throw around words
like "superior"
> > and "exact" and "better"
> > > AND "my EV" as if WE all understand exactly (lol) what they
mean
> > and how they MUST be
> > > used. There is nothing wrong with saying "this is the RoRBR
of
> > the lowest cost royal
> > > strategy" or whatever (infact that may be valuable
information)
> > The problem apprears
> > > when we say things like "playing strategy [x] is better than
not
> > playing [x]" without clearly
> > > stating on what basis it is better and without demonstrating
why
> > (or how) that basis is
> > > important (or statistically significant or meaningful, etc).
> > >
> > > --- In vpFREE@yahoogroups.com, "Bob Dancer" <bob.dancer@c...>
wrote:
> > > >
> > > > I wrote: > The key to success at video poker (at least for
those
> > with a
> > > > > profit-max sort of goal) is losing the minimum BETWEEN
royals.
> > > > > Increasing the rate at which royals occur is not a
costless
> > process.
> > > >
> > > > Harry responded: Now, here's where I'm going to disagree
with
> > you. (And
> > > > it wouldn't be surprising if Steve Jacobs weighs in.)
> > > >
> > > >
> > > > And well you should disagree. I wasn't speaking precisely.
I was
> > > > speaking in general terms comparing max-EV strategy (which
I
> > support)
> > > > and a "get the royals quickest" strategy (which I don't
support).
> > I
> > > > meant to say the first strategy has a lower cost between
royals
> > than the
> > > > second. I did not mean to say that a strategy devised to
reduce
···
> > loss
> > > > between royals would be superior to a max-EV strategy.
> > > >
> > > >
> > > > Bob Dancer
> > > >
> > >
> >
>
Pardon me but, I didn't see anything in your reply about "penalty
cards"
http://wizardofodds.com/jacksorbetter
and it doesn't show that the wizardofodds.com has people there
who used to work and program for IGT's video poker games !? :>
http://wizardofodds.com/general/
The Wizard of Odds is Michael Shackleford, A.S.A., a professional
actuary who has made a career of analyzing casino games. He runs the
numbers on new games for casinos and game developers and has helped
design many of the popular slot machines on the Internet. He is
currently an Adjunct Professor of Casino Math at the University of
Nevada, Las Vegas, a former contributing editor to Casino Player
magazine, and the author of the book Gambling 102, recently published
by Huntington Press. The Wizard's landmark research into the actual
returns of slot machines on the Las Vegas strip garnered international
attention in 2002, and he has appeared numerous times on national
television as a recognized expert on gambling strategy.
About our sit
···
--- In vpFREE@yahoogroups.com, "gilbert_616" <gilbert_616@y...> wrote:
--- In vpFREE@yahoogroups.com, "nightoftheiguana2000"
<nightoftheiguana2000@y...> wrote:
>
> for example, perfect 9/6 JOB strategy, found here:
> http://wizardofodds.com/jacksorbetter
> has er=0.9954, var=19.78
> spreadsheet wizard's numbers:
> er=0.995439044, var=19.51467643
> this simple strategy:
>
5SF>4RF>PAT>4SF>HP>3RF>4FL>KQJT>LP>4STo>3SF0>AKQJ>2RF2H>4STi3H>3SF1>KQ
>2H>JT>QT>1H>3SF2
> has er=0.9954, var=19.40
>
> --- In vpFREE@yahoogroups.com, "nightoftheiguana2000"
> <nightoftheiguana2000@y...> wrote:
> >
> > frugal video poker will answer this question, and more
> >
> > --- In vpFREE@yahoogroups.com, "gilbert_616" <gilbert_616@y...>
wrote:
> > >
> > > Each time I see/read the topic on "penalty cards", I get this
feeling
> > > that I should write to Santa (Ho ho ho) and present my wish as
> > > follows:
> > >
> > > Dear Santa,
> > >
> > > Could you please knock on IGT's door and ask the
programmers/software
> > > engineers to "Explain penalty cards in video poker games in
details
> > > and how much more a player can win if s/he fully understand the
> > > correct HOLDS when they show up on the screen?".
> > >
> > > Merry X'mas!
> > >
> > >
> > > --- In vpFREE@yahoogroups.com, "cdfsrule" <groups.yahoo@v...>
wrote:
> > > >
> > > > I guess I have to disagree a wee bit with all of you (or at
least
> > > it seems like a disagreement
> > > > at this time). While Steve considers non-MAX EV strategies (
and
> > > may go so far as to
> > > > propose that such stratgeis are appropriate for some
players), I
> > > want to go farther: I don't
> > > > think "EV" per se is the only quantity-- or even the most
important
> > > quanntity for us ( and
> > > > especially Bob's mythical pro-player ) to worry about (that
is
> > > optimize).
> > > >
> > > > I don't have a problem with the logic or math that folks are
using--
> > > it all seems quite
> > > > kosher-- my disagreement lies with what it all "means", how
we tend
> > > to mis-intrepret the
> > > > statistical quantities when we don't understand their meaning
(or
> > > haven't defined their
> > > > meaning), how we continue to argue about the numerical
quantites
> > > rather than what the
> > > > quantites mean or say. Once we take a look at what these
> > > quantities mean, I think folks
> > > > may come to see the EV is not exactly what they think it is.
> > > Admittedly this is a difficult
> > > > subject to discuss in a tidy manner. So let me play a little
fast
> > > and loose and use a story:
> > > >
> > > > Assume Player 1 follows a strict max EV stratgey. And I mean
> > > strict, every penalty card.
> > > > Player 1 plays perfectly, just like a computer. (I'll call
this
> > > player "Bob")
> > > >
> > > > Now Player 2 comes along, sits down next to Bob, and plays a
> > > different Stratgey. Bob
> > > > knows its a different stratgey becuase he sees hands where he
would
> > > have made a
> > > > different play. When Bob notices the different play, he finds
that
> > > the EV of his play was
> > > > sometimes higher and sometimes the same as player 2's play.
Since
> > > Bob plays (by
> > > > definition) the true max EV strategy, this other player must
be
> > > playing a stratgey with a
> > > > lower EV.
> > > >
> > > > So Bob says to Player 2, "you know you can do better if you
use my
> > > strategy, since your
> > > > strategy isn't max-EV". Player 2 response, "Nope. I'm
trying to
> > > get as many Royals as
> > > > possible at the lowest cost to me. My buddy says that if I
get
> > > more Royals them him, he'll
> > > > give me a phat check. That's a lot of money to me. I think
it is
> > > worth it for me to not play
> > > > max-EV. I am playing a max-RF strategy because with the
potential
> > > of the phat prize, my
> > > > overall EV is higher with it than with the Max EV
strategy"
> > > Player #2's logic seems to
> > > > make sense to Bob so he smiles, and each continues to play
their
> > > stratgey.
> > > >
> > > > But we are not done yet. Player 2's choice seems to make
sense,
> > > but since the the terms
> > > > "my EV" (that is my EV from play) and the stratgies EV (that
is,
> > > the EV from the stratgey,
> > > > the parent population EV or first moment) haven't been
defined (we
> > > don't really know what
> > > > they "mean" or even if they should "equate"), we're not done.
So
> > > I'd say we haven't done a
> > > > complete analysis of the situation. Let's start with EV, as
an
> > > example. Most of would
> > > > define EV as the first moment of the PDF. In other words, we
> > > assume that EV is the simple
> > > > arithmetric average (in any units we choose). We tend to
stop our
> > > analysis at this level
> > > > (without going deeper) since, knowing how to compuet EV, we
feel
> > > confident that we
> > > > can-- and should-- use "EV" all over the place. But what
are we
> > > using it for? The EV we
> > > > all love is the first moment of the "parent popluation" (
that is
> > > the theoretical return) and
> > > > it is the simple aritmeric average of our actual results. Are
they
> > > the same thing? DO the
> > > > quantities have the same meaning? For VP, I already know the
parent
> > > popluation mean is a
> > > > poor proxy for our actual returns. And I think most playerw
would
> > > agree with that
> > > > statement. But How poor? That's a VERY important question,
> > > perhaps the key question
> > > > that provides meaning to the term "EV". The answer to that
> > > question is given by the PDF
> > > > (if the PDF is normal, the answer can be equivelently given
by
> > > just the variance). The PDF
> > > > gives the actual spread of EV's that we would be expected to
> > > acheive in real play. We
> > > > know how to compute the PDF for any number of hands played,
for any
> > > bankroll, and for
> > > > any (properly defined) strategy. And given the PDF, we can
answer
> > > all kinds of rigorous
> > > > statistical questions, technical ones like what is the
likelihood
> > > of so and so's results being
> > > > from this VP game (vs another), and more useful ones, like
I've
> > > lost X, what is the
> > > > theortical likihood of this happening, and are my actual
results
> > > consistant with the
> > > > strategy I think I am playing? [There are a lot of other
PDF's we
> > > could compute, like the
> > > > PDF of the variance, etc, and many question we can answer
with
> > > them]. So the PDF can
> > > > help understand what EV means to us. Take JoB and FPDW as
> > > examples. FPDW is a
> > > > positive game, while JoB isn't. That means the EV of FPDW is
> > > always greater than for JoB--
> > > > for any number of hands. Does that mean the the we are more
likely
> > > to win playing FPDW?
> > > > or less likely to lose? To answer these questions, you can
look at
> > > the PDF's on Jazbo's
> > > > website (under volatility) or the CDFs. You might notice
that the
> > > peaks in the PDFs don't
> > > > cluster around the theoretical EV (which isn't indicated, but
is
> > > close to zero). the peak is
> > > > always to the "left" or the EV. That strongly suggests that
EV (the
> > > first moment of the
> > > > parent population) is not a particualrly good indicatior of
our
> > > real-play "EV". In other
> > > > words, the simple arithmetric average we compute for our
results
> > > will tend not to cluster
> > > > about the theoretical EV. That is, what we mean by "EV for
our own
> > > play" might is
> > > > something a bit different from what we mean when we
> > > say "theoritical EV". This issue is
> > > > nothing knew to statisticians: The statistical quantity
(of the
> > > parent population) that
> > > > "best" (which needs to be rigoursly defined) approximates the
> > > average of our actual results
> > > > may not be the simple arithmetric mean. Indeed for a
reasonable
> > > number of VP hands (in
> > > > the 10K's), most people would say the "mode" better
approximates
> > > our actual results (look
> > > > at Jazbo's PDFs), not the "EV" or even the "median". Wow.
EV is
> > > not king. But rather than
> > > > argue about what statistical quantity is best (if any!) , why
not
> > > look at the entire PDF? It
> > > > contains everything we do know, so it must contain the
answers we
> > > want if the answer is
> > > > knowable (strictly speaking this is not true, but it is true
enough)
> > > >
> > > > So, what we need to do is compare the PDF for player 1 (
Bob), the
> > > max EV guy, to the PDF
> > > > for player 2 (the other stratgey) and look for the
differences.
> > > The next step is to decide
> > > > which differences are important to us and statistically
significant
> > > (This can be done
> > > > rigorously since we know the parent PDF). We can do this as
a
> > > function of # hands played,
> > > > or better yet, time (given hands/hour). We can then see the
real
> > > effect of different
> > > > strategies such as playing with and without penalty cards.
Case
> > > closed.
> > > >
> > > > Conclusion: I beleive The PDF (or CDF) is the thing to look
at. A
> > > change in strategy or
> > > > game may change the PDF. When choosing a game, I sometimes
chose
> > > the game with the
> > > > lower EV becuase I like its PDF better: the "peak" of the PDF
has a
> > > higher likely return (right
> > > > shifted is better) or the left "tail" is shifted to the
right, or
> > > something else. Likewise, I
> > > > imagine that there are folks who choose a statgey in this
same PDF
> > > based way, though the
> > > > differences in PDFs between strategies are often much smaller
than
> > > those between games.
> > > > And of coarse, as others have pointed out, there are some out
there
> > > who may (should?)
> > > > choose a strategy based upon its associated RoRBX. But IMHO,
they
> > > ignore the PDF at
> > > > their own peril. We tend to throw around words
like "superior"
> > > and "exact" and "better"
> > > > AND "my EV" as if WE all understand exactly (lol) what they
mean
> > > and how they MUST be
> > > > used. There is nothing wrong with saying "this is the RoRBR
of
> > > the lowest cost royal
> > > > strategy" or whatever (infact that may be valuable
information)
> > > The problem apprears
> > > > when we say things like "playing strategy [x] is better than
not
> > > playing [x]" without clearly
> > > > stating on what basis it is better and without demonstrating
why
> > > (or how) that basis is
> > > > important (or statistically significant or meaningful, etc).
> > > >
> > > > --- In vpFREE@yahoogroups.com, "Bob Dancer" <bob.dancer@c...>
wrote:
> > > > >
> > > > > I wrote: > The key to success at video poker (at least for
those
> > > with a
> > > > > > profit-max sort of goal) is losing the minimum BETWEEN
royals.
> > > > > > Increasing the rate at which royals occur is not a
costless
> > > process.
> > > > >
> > > > > Harry responded: Now, here's where I'm going to disagree
with
> > > you. (And
> > > > > it wouldn't be surprising if Steve Jacobs weighs in.)
> > > > >
> > > > >
> > > > > And well you should disagree. I wasn't speaking precisely.
I was
> > > > > speaking in general terms comparing max-EV strategy (which
I
> > > support)
> > > > > and a "get the royals quickest" strategy (which I don't
support).
> > > I
> > > > > meant to say the first strategy has a lower cost between
royals
> > > than the
> > > > > second. I did not mean to say that a strategy devised to
reduce
> > > loss
> > > > > between royals would be superior to a max-EV strategy.
> > > > >
> > > > >
> > > > > Bob Dancer
> > > > >
> > > >
> > >
> >
>
That's good info but, it still doesn't say that Michael Shackleford
works or used to work for IGT as a designer or programmer of the
Video Poker games that we play. So, I'll just ask Santa if he can get
the real stuffs (the truth about penalty cards) from IGT. :>
--- In vpFREE@yahoogroups.com, "nightoftheiguana2000"
<nightoftheiguana2000@y...> wrote:
--- In vpFREE@yahoogroups.com, "gilbert_616" <gilbert_616@y...>
wrote:
> Pardon me but, I didn't see anything in your reply about "penalty
> cards"http://wizardofodds.com/jacksorbetter
> and it doesn't show that the wizardofodds.com has people there
> who used to work and program for IGT's video poker games !? :>http://wizardofodds.com/general/
The Wizard of Odds is Michael Shackleford, A.S.A., a professional
actuary who has made a career of analyzing casino games. He runs the
numbers on new games for casinos and game developers and has helped
design many of the popular slot machines on the Internet. He is
currently an Adjunct Professor of Casino Math at the University of
Nevada, Las Vegas, a former contributing editor to Casino Player
magazine, and the author of the book Gambling 102, recently
published
by Huntington Press. The Wizard's landmark research into the actual
returns of slot machines on the Las Vegas strip garnered
international
attention in 2002, and he has appeared numerous times on national
television as a recognized expert on gambling strategy.About our sit
>
>
> --- In vpFREE@yahoogroups.com, "nightoftheiguana2000"
> <nightoftheiguana2000@y...> wrote:
> >
> > for example, perfect 9/6 JOB strategy, found here:
> > http://wizardofodds.com/jacksorbetter
> > has er=0.9954, var=19.78
> > spreadsheet wizard's numbers:
> > er=0.995439044, var=19.51467643
> > this simple strategy:
> >
>
5SF>4RF>PAT>4SF>HP>3RF>4FL>KQJT>LP>4STo>3SF0>AKQJ>2RF2H>4STi3H>3SF1>KQ
> >2H>JT>QT>1H>3SF2
> > has er=0.9954, var=19.40
> >
> > --- In vpFREE@yahoogroups.com, "nightoftheiguana2000"
> > <nightoftheiguana2000@y...> wrote:
> > >
> > > frugal video poker will answer this question, and more
> > >
> > > --- In vpFREE@yahoogroups.com, "gilbert_616"
<gilbert_616@y...>
> wrote:
> > > >
> > > > Each time I see/read the topic on "penalty cards", I get
this
> feeling
> > > > that I should write to Santa (Ho ho ho) and present my wish
as
> > > > follows:
> > > >
> > > > Dear Santa,
> > > >
> > > > Could you please knock on IGT's door and ask the
> programmers/software
> > > > engineers to "Explain penalty cards in video poker games in
> details
> > > > and how much more a player can win if s/he fully understand
the
> > > > correct HOLDS when they show up on the screen?".
> > > >
> > > > Merry X'mas!
> > > >
> > > >
> > > > --- In vpFREE@yahoogroups.com, "cdfsrule"
<groups.yahoo@v...>
> wrote:
> > > > >
> > > > > I guess I have to disagree a wee bit with all of you (or
at
> least
> > > > it seems like a disagreement
> > > > > at this time). While Steve considers non-MAX EV
strategies (
> and
> > > > may go so far as to
> > > > > propose that such stratgeis are appropriate for some
> players), I
> > > > want to go farther: I don't
> > > > > think "EV" per se is the only quantity-- or even the most
> important
> > > > quanntity for us ( and
> > > > > especially Bob's mythical pro-player ) to worry about
(that
> is
> > > > optimize).
> > > > >
> > > > > I don't have a problem with the logic or math that folks
are
> using--
> > > > it all seems quite
> > > > > kosher-- my disagreement lies with what it all "means",
how
> we tend
> > > > to mis-intrepret the
> > > > > statistical quantities when we don't understand their
meaning
> (or
> > > > haven't defined their
> > > > > meaning), how we continue to argue about the numerical
> quantites
> > > > rather than what the
> > > > > quantites mean or say. Once we take a look at what
these
> > > > quantities mean, I think folks
> > > > > may come to see the EV is not exactly what they think it
is.
> > > > Admittedly this is a difficult
> > > > > subject to discuss in a tidy manner. So let me play a
little
> fast
> > > > and loose and use a story:
> > > > >
> > > > > Assume Player 1 follows a strict max EV stratgey. And I
mean
> > > > strict, every penalty card.
> > > > > Player 1 plays perfectly, just like a computer. (I'll
call
> this
> > > > player "Bob")
> > > > >
> > > > > Now Player 2 comes along, sits down next to Bob, and
plays a
> > > > different Stratgey. Bob
> > > > > knows its a different stratgey becuase he sees hands
where he
> would
> > > > have made a
> > > > > different play. When Bob notices the different play, he
finds
> that
> > > > the EV of his play was
> > > > > sometimes higher and sometimes the same as player 2's
play.
> Since
> > > > Bob plays (by
> > > > > definition) the true max EV strategy, this other player
must
> be
> > > > playing a stratgey with a
> > > > > lower EV.
> > > > >
> > > > > So Bob says to Player 2, "you know you can do better if
you
> use my
> > > > strategy, since your
> > > > > strategy isn't max-EV". Player 2 response, "Nope. I'm
> trying to
> > > > get as many Royals as
> > > > > possible at the lowest cost to me. My buddy says that if
I
> get
> > > > more Royals them him, he'll
> > > > > give me a phat check. That's a lot of money to me. I
think
> it is
> > > > worth it for me to not play
> > > > > max-EV. I am playing a max-RF strategy because with the
> potential
> > > > of the phat prize, my
> > > > > overall EV is higher with it than with the Max EV
> strategy"
> > > > Player #2's logic seems to
> > > > > make sense to Bob so he smiles, and each continues to
play
> their
> > > > stratgey.
> > > > >
> > > > > But we are not done yet. Player 2's choice seems to make
> sense,
> > > > but since the the terms
> > > > > "my EV" (that is my EV from play) and the stratgies EV
(that
> is,
> > > > the EV from the stratgey,
> > > > > the parent population EV or first moment) haven't been
> defined (we
> > > > don't really know what
> > > > > they "mean" or even if they should "equate"), we're not
done.
> So
> > > > I'd say we haven't done a
> > > > > complete analysis of the situation. Let's start with EV,
as
> an
> > > > example. Most of would
> > > > > define EV as the first moment of the PDF. In other
words, we
> > > > assume that EV is the simple
> > > > > arithmetric average (in any units we choose). We tend to
> stop our
> > > > analysis at this level
> > > > > (without going deeper) since, knowing how to compuet
EV, we
> feel
> > > > confident that we
> > > > > can-- and should-- use "EV" all over the place. But
what
> are we
> > > > using it for? The EV we
> > > > > all love is the first moment of the "parent popluation"
(
> that is
> > > > the theoretical return) and
> > > > > it is the simple aritmeric average of our actual results.
Are
> they
> > > > the same thing? DO the
> > > > > quantities have the same meaning? For VP, I already know
the
> parent
> > > > popluation mean is a
> > > > > poor proxy for our actual returns. And I think most
playerw
> would
> > > > agree with that
> > > > > statement. But How poor? That's a VERY important
question,
> > > > perhaps the key question
> > > > > that provides meaning to the term "EV". The answer to
that
> > > > question is given by the PDF
> > > > > (if the PDF is normal, the answer can be equivelently
given
> by
> > > > just the variance). The PDF
> > > > > gives the actual spread of EV's that we would be expected
to
> > > > acheive in real play. We
> > > > > know how to compute the PDF for any number of hands
played,
> for any
> > > > bankroll, and for
> > > > > any (properly defined) strategy. And given the PDF, we
can
> answer
> > > > all kinds of rigorous
> > > > > statistical questions, technical ones like what is the
> likelihood
> > > > of so and so's results being
> > > > > from this VP game (vs another), and more useful ones,
like
> I've
> > > > lost X, what is the
> > > > > theortical likihood of this happening, and are my actual
> results
> > > > consistant with the
> > > > > strategy I think I am playing? [There are a lot of other
> PDF's we
> > > > could compute, like the
> > > > > PDF of the variance, etc, and many question we can answer
> with
> > > > them]. So the PDF can
> > > > > help understand what EV means to us. Take JoB and FPDW
as
> > > > examples. FPDW is a
> > > > > positive game, while JoB isn't. That means the EV of
FPDW is
> > > > always greater than for JoB--
> > > > > for any number of hands. Does that mean the the we are
more
> likely
> > > > to win playing FPDW?
> > > > > or less likely to lose? To answer these questions, you
can
> look at
> > > > the PDF's on Jazbo's
> > > > > website (under volatility) or the CDFs. You might notice
> that the
> > > > peaks in the PDFs don't
> > > > > cluster around the theoretical EV (which isn't indicated,
but
> is
> > > > close to zero). the peak is
> > > > > always to the "left" or the EV. That strongly suggests
that
> EV (the
> > > > first moment of the
> > > > > parent population) is not a particualrly good indicatior
of
> our
> > > > real-play "EV". In other
> > > > > words, the simple arithmetric average we compute for our
> results
> > > > will tend not to cluster
> > > > > about the theoretical EV. That is, what we mean by "EV
for
> our own
> > > > play" might is
> > > > > something a bit different from what we mean when we
> > > > say "theoritical EV". This issue is
> > > > > nothing knew to statisticians: The statistical
quantity
> (of the
> > > > parent population) that
> > > > > "best" (which needs to be rigoursly defined) approximates
the
> > > > average of our actual results
> > > > > may not be the simple arithmetric mean. Indeed for a
> reasonable
> > > > number of VP hands (in
> > > > > the 10K's), most people would say the "mode" better
> approximates
> > > > our actual results (look
> > > > > at Jazbo's PDFs), not the "EV" or even the "median".
Wow.
> EV is
> > > > not king. But rather than
> > > > > argue about what statistical quantity is best (if any!) ,
why
> not
> > > > look at the entire PDF? It
> > > > > contains everything we do know, so it must contain the
> answers we
> > > > want if the answer is
> > > > > knowable (strictly speaking this is not true, but it is
true
> enough)
> > > > >
> > > > > So, what we need to do is compare the PDF for player 1 (
> Bob), the
> > > > max EV guy, to the PDF
> > > > > for player 2 (the other stratgey) and look for the
> differences.
> > > > The next step is to decide
> > > > > which differences are important to us and statistically
> significant
> > > > (This can be done
> > > > > rigorously since we know the parent PDF). We can do this
as
> a
> > > > function of # hands played,
> > > > > or better yet, time (given hands/hour). We can then see
the
> real
> > > > effect of different
> > > > > strategies such as playing with and without penalty
cards.
> Case
> > > > closed.
> > > > >
> > > > > Conclusion: I beleive The PDF (or CDF) is the thing to
look
> at. A
> > > > change in strategy or
> > > > > game may change the PDF. When choosing a game, I
sometimes
> chose
> > > > the game with the
> > > > > lower EV becuase I like its PDF better: the "peak" of the
> has a
> > > > higher likely return (right
> > > > > shifted is better) or the left "tail" is shifted to the
> right, or
> > > > something else. Likewise, I
> > > > > imagine that there are folks who choose a statgey in this
> same PDF
> > > > based way, though the
> > > > > differences in PDFs between strategies are often much
smaller
> than
> > > > those between games.
> > > > > And of coarse, as others have pointed out, there are some
out
> there
> > > > who may (should?)
> > > > > choose a strategy based upon its associated RoRBX. But
IMHO,
> they
> > > > ignore the PDF at
> > > > > their own peril. We tend to throw around words
> like "superior"
> > > > and "exact" and "better"
> > > > > AND "my EV" as if WE all understand exactly (lol) what
they
> mean
> > > > and how they MUST be
> > > > > used. There is nothing wrong with saying "this is the
RoRBR
> of
> > > > the lowest cost royal
> > > > > strategy" or whatever (infact that may be valuable
> information)
> > > > The problem apprears
> > > > > when we say things like "playing strategy [x] is better
than
> not
> > > > playing [x]" without clearly
> > > > > stating on what basis it is better and without
demonstrating
> why
> > > > (or how) that basis is
> > > > > important (or statistically significant or meaningful,
etc).
> > > > >
> > > > > --- In vpFREE@yahoogroups.com, "Bob Dancer"
<bob.dancer@c...>
> wrote:
> > > > > >
> > > > > > I wrote: > The key to success at video poker (at least
for
> those
> > > > with a
> > > > > > > profit-max sort of goal) is losing the minimum
BETWEEN
> royals.
> > > > > > > Increasing the rate at which royals occur is not a
> costless
> > > > process.
> > > > > >
> > > > > > Harry responded: Now, here's where I'm going to
disagree
> with
> > > > you. (And
> > > > > > it wouldn't be surprising if Steve Jacobs weighs in.)
> > > > > >
> > > > > >
> > > > > > And well you should disagree. I wasn't speaking
precisely.
> I was
> > > > > > speaking in general terms comparing max-EV strategy
(which
> I
> > > > support)
> > > > > > and a "get the royals quickest" strategy (which I don't
> support).
> > > > I
> > > > > > meant to say the first strategy has a lower cost
between
> royals
> > > > than the
> > > > > > second. I did not mean to say that a strategy devised
to
···
> reduce
> > > > loss
> > > > > > between royals would be superior to a max-EV strategy.
> > > > > >
> > > > > >
> > > > > > Bob Dancer
> > > > > >
> > > > >
> > > >
> > >
> >
>
I'm sorry, I missed much of this thread. Could you please clarify
what the programming of VP games has to do with penalty cards? It
sounds like there is a really interesting point here.
Thanks.
Bill
···
At 03:20 PM 12/7/2005, you wrote:
That's good info but, it still doesn't say that Michael Shackleford
works or used to work for IGT as a designer or programmer of the
Video Poker games that we play. So, I'll just ask Santa if he can get
the real stuffs (the truth about penalty cards) from IGT. :>
[Non-text portions of this message have been removed]
why does an IGT employee have to be the one to explain?
anyone can analyze the EV of a specific play on a specific hand can't they?
···
On 12/7/05, gilbert_616 <gilbert_616@yahoo.com> wrote:
That's good info but, it still doesn't say that Michael Shackleford
works or used to work for IGT as a designer or programmer of the
Video Poker games that we play. So, I'll just ask Santa if he can get
the real stuffs (the truth about penalty cards) from IGT. :>---
Not all IGT employee can explain the truth on penalty cards. The ones
who designed/developed/programmed/wrote the codes would know. Or the
ones who have seen and understand the source codes.
You're right! Anyone can analyze but, without seeing the actual codes
(line by line) and knowing exactly what the program does when penalty
cards exist, s/he can still be just guessing.
why does an IGT employee have to be the one to explain?
anyone can analyze the EV of a specific play on a specific hand
can't they?
> That's good info but, it still doesn't say that Michael
Shackleford
> works or used to work for IGT as a designer or programmer of the
> Video Poker games that we play. So, I'll just ask Santa if he can
get
···
--- In vpFREE@yahoogroups.com, LHOOQ <fieldcommand@g...> wrote:
On 12/7/05, gilbert_616 <gilbert_616@y...> wrote:
> the real stuffs (the truth about penalty cards) from IGT. :>
>
>
>
> ---