vpFREE2 Forums

Re; When Positive games don't really help, LONG

#1

Richard Boston wrote ..............................

Someone recently asked about how many hands or how long for the true
percentages to kick in. Well here's some interesting examples of how it's
not in the thousands. ...

Very true. But if you are playing 10/7 double bonus perfectly and have a
bad run where you lose 5% in 40,000 hands ( no aces or royals for example)
yes it will take you a long time to 'make that up'. What is most likely for
your next 40,000 hands is either a loss of $915 ( no royals) or a win of $85
( 1 royal) or a win of $1085 ( with 2 royals). There are other higher royals
possible as well. A losing session one day does not make the next session
more likely to be a winner.

The key point is that you don't 'make up' for a bad session with the next
session or next 2 sessions or next 20 sessions.

A few days ago I was playing various games because that were themselves not
positive but I had a promotion that made it so. I didn't see a quad for well
over 5000 hands. Not a single one. And when I finally did see some it came
out to only about 3 quads in nearly 7000 hands. Regardless of positive or
negative game, if you get on a roll like this you are going to lose big
time.

It's true that going 5000 hands without quads will generally be a big
losing session. But if you play 9/5 JOB instead of 9/6 JOB it would
be an even bigger losing session.

Another thing I love to laugh about is all the angst over a .2 or .5%
difference from one game to the next.
Across a two week span I'd say that I haven't had a session over 90% return.
This is playing both positive and negative games (w/ a promo) and it still
did not matter. In fact today I had 3 sessions. The first was 90%, the second
80% and the third was 50%. FIFTY PERCENT.

Without some more information about how many hands and how well you
play, it's hard to put an estimate on how 'unlucky' you were.
Short sessions can give some unusual results. I've gone 18 hands in
PE without a winner. If my session is 18 hands, I had a ZERO return.

The huge seeming paradox is that every hand is independent but if you
play long enough, you'll approach the EV of the game. They are both
true statements.

That's also happened to me more than once, in fact during the quad drought
it was looking like that.
Now based on my Winpoker rating I'm FAR FAR from a pro or an expert but I
don't play at 50% or even 90%, I believe it's more like 99% (but it's been a
while since I tested myself).

A few more sessions like these and it's going to take several sessions where
I come out w/ 2 royals per session just to get even close to 98% overall for
the month of April. And I don't see that happening playing quarter single
line VP which is what I've been playing.

You don't 'balance out' every month. In fact, the whole notion of a session
score loses its meaning if you believe in independent trials.
If you believe playing 100,000 hands in a row is different from playing
10,000 hands 10 times and comparing results, we have fundamentally
different beliefs in how video poker machines work.
With all the smart guys on this group, someone would have figured out an
optimal session length and we would put all the casinos out of business.
You can't order a set of independent trials in such a way as to make the
sum total different than any other ordering.

There's a document about this in the vpfree files section.

Plan your next roadtrip with MapQuest.com: America's #1 Mapping Site.

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

#2

>> The huge seeming paradox is that every hand is independent but if you
>> play long enough, you'll approach the EV of the game. They are both
>> true statements.

That's incorrect. If you "play long enough" your results will be
somewhere on a normal distribution with a mean equal to the EV and a
variance equal to the variance of one hand. What this means is that if
you "play long enough" somebody is gonna cry uncle, either you will
quit from losing too much or the casino will kick you out for winning
too much. For a breakeven game, the odds track the ratio, so if your
stop loss is equal to the casino's kick out point, the odds are 50-50.
Otherwise the key ratio is N0 (=variance/advantage^2). At N0 hands,
your odds of winning a positive gamble (or losing a negative gamble)
are about 84%. For FPDW N0=26/.0075^2 is about half a million hands.
For 9/6 DDB N0=42/-.01^2 is also about half a million hands.

#3

greeklandjohnny wrote:

> The huge seeming paradox is that every hand is independent but
> if you play long enough, you'll approach the EV of the game.

nightoftheiguana2000 wrote:

That's incorrect. If you "play long enough" your results will be
somewhere on a normal distribution with a mean equal to the EV and a
variance equal to the variance of one hand. What this means is that
if you "play long enough" somebody is gonna cry uncle, either you
will quit from losing too much or the casino will kick you out for
winning too much.

Good answer ... hope you don't mind if I come at gj's comments from a
different angle.

Had he made the comment about ER rather than EV, he'd be on target.
For the sake of clarity, let me step back and define both.

ER is total expected payout on your bets DIVIDED by the bets
themselves. If the expected payout is greater than the bets, the
value is greater than 100% ("positive").

"EV" is sometimes used a little ambiguously, but it's best thought of
as the expected payout MINUS the bets -- i.e., your expected profit
or loss. This is the true sense of a "positive" game.

As you play, you look for the difference between ER and the actual %
return realized to grow smaller. However, most players don't give a
damn about a statistic ... they care about the absolute dollars in
their pocket. It's the unfortunate and inescapable fact that the
longer you play the GREATER you look for actual results to deviate
from EV.

But there's a saving grace that NOTI introduced a few years ago --
the math concept of "NO" (a reference to a math term, not something
like "just say NO" ;). The fallout of "NO" is that the longer you
play a "positive" game, the greater the probability that you will end
up in positive turf. And the more positive and lower variance that
play, the faster you can look to have strong expectation of a
profit.

It's the rather modest magnitude of the NO numbers for a strong game
that shifts the vp "advantage play" concept from
entailing "lifetimes" of play to something that can involve just a
few years.

- Harry

#4

Not at all, you're much better with the words. My point is that it's a
common fallacy that if you play a breakeven gamble long enough, you
will eventually breakeven. In the real world (excluding some special
cases), either you or your opponent will bust out. You can see the
effect graphically with this calculator:
http://members.aol.com/lotsie/GamblersRuin.html
Pick 4700 Jacks or better (a breakeven game) and busted and retired at
$200 and hit run and watch the middle disappear. Visual proof of the
above statement. Go ahead and vary the busted and retired amounts, it
just takes longer to run larger amounts, but the result (disappearance
of the middle, i.e. "breakeven") is the same.

···

--- In vpFREE@yahoogroups.com, "Harry Porter" <harry.porter@...> wrote:

Good answer ... hope you don't mind if I come at gj's comments from a
different angle.

#5

Minor point: N0 works for negative games as well, it's just that the
situation is reversed, you can look to have a strong expectation of
losing. This is how casinos "grind down" their bread and butter clientele.

http://members.cox.net/vpfree/Bank_NO.htm

···

--- In vpFREE@yahoogroups.com, "Harry Porter" <harry.porter@...> wrote:

The fallout of "NO" is that the longer you
play a "positive" game, the greater the probability that you will end
up in positive turf. And the more positive and lower variance that
play, the faster you can look to have strong expectation of a
profit.

#6

<nightoftheiguana2000@...> wrote:

You can see the effect graphically with this calculator:
http://members.aol.com/lotsie/GamblersRuin.html
Pick 4700 Jacks or better (a breakeven game) and busted and
retired at $200 and hit run and watch the middle disappear.
Visual proof of the above statement. Go ahead and vary the
busted and retired amounts, it just takes longer to run larger
amounts, but the result (disappearance of the middle,
i.e. "breakeven") is the same.

The "disappearance" of the middle is due to the very small bankroll
of $200 for a quarter VP game. To illustrate a well-bankrolled
player, use the same calculator, with the busted and retired set to
$10,000. After a couple hundred hours, you'll see a classic bell-
curve, showing that most likely, you'll indeed be near the middle,
aka breakeven.

Minor point: N0 works for negative games as well, it's just that
the situation is reversed, you can look to have a strong
expectation of losing. This is how casinos "grind down" their
bread and butter clientele.

Casinos make money on negative games (e.g., 9/6/4000 JOB, with no
progressive or cashback or promos). If you play these games for a
long time, you have a strong expectation of losing. It makes no
difference whether or not you're a bread and butter client. Playing
100 hours of video poker in a single week, or over a couple of years,
leads to the same expectation.

The machine and the math don't know if you're a local or a visitor,
and math makes no distinction about who's getting "ground down." It's
just the simple matter of a small casino edge repeated over and over,
tending towards a mean expectation as time goes by.

The casino only needs to understand the mathematical odds. It's
irrelevent to them whether any inidividual gambler is under-
bankrolled or is getting ground down.

Stuart
http://stuart-randomthoughts.blogspot.com/

#7

A classic bell-curve is "near the middle"? Wouldn't it be more correct
to think in terms of a "standard deviation"? And doesn't the "standard
deviation" always increase in dollar value the more hands you play? Or
do you assume the dollar standard deviation actually decreases or is
stable at some fixed point?

···

--- In vpFREE@yahoogroups.com, "Stuart" <sresnick2@...> wrote:

The "disappearance" of the middle is due to the very small bankroll
of $200 for a quarter VP game. To illustrate a well-bankrolled
player, use the same calculator, with the busted and retired set to
$10,000. After a couple hundred hours, you'll see a classic bell-
curve, showing that most likely, you'll indeed be near the middle,
aka breakeven.