vpFREE2 Forums

Some questions on variance and EV

#1

Many months ago some of the experts here showed me there are other
strategies besides max ev. I did not think any of it would apply to
my play. Recently I have started playing some progressives and I
wonder if there is a way to evaluate increased EV versus increased
variance at strategy break points. Does anyone wait for a certain
amount of increased EV before accepting the increased variance.

The TomSki index can be used to compare the trade off in EV versus
variance for different games, but does it make sense to apply
something similar to a strategy choice.

In Bob Dancer's recent column I was shocked that an EV difference
of .000033 was too small for Bob to worry about. Now I wonder what
about the variance between the 2 choices. Is holding the QJ instead
of the A a better choice because it reduces the variance. I am
assuming that QJ will lower the variance, but can someone show me how
to do the calculation.

Thanks,

Chris

#2

Certainty Equivalence = EV - Variance/(Bankroll x 2)

http://www.google.com/search?q=Certainty+Equivalence

···

--- In vpFREE@yahoogroups.com, "kcace1024" <cy4873@...> wrote:

In Bob Dancer's recent column I was shocked that an EV difference
of .000033 was too small for Bob to worry about. Now I wonder what
about the variance between the 2 choices. Is holding the QJ instead
of the A a better choice because it reduces the variance. I am
assuming that QJ will lower the variance, but can someone show me how
to do the calculation.

#3

Iguana,

To use the equation below I must calculate the variance for holding the
ace alone and then the variance for holding just QJ. I would need to
add up all the winning contributions mnultiplied by the frequency of
each contribution for each hold. Is that correct? Does any of the VP
software have this capability?

Thanks

--- In vpFREE@yahoogroups.com, "nightoftheiguana2000"
<nightoftheiguana2000@...> wrote:

> In Bob Dancer's recent column I was shocked that an EV difference
> of .000033 was too small for Bob to worry about. Now I wonder what
> about the variance between the 2 choices. Is holding the QJ

instead

> of the A a better choice because it reduces the variance. I am
> assuming that QJ will lower the variance, but can someone show me

how

···

--- In vpFREE@yahoogroups.com, "kcace1024" <cy4873@> wrote:
> to do the calculation.

Certainty Equivalence = EV - Variance/(Bankroll x 2)

http://www.google.com/search?q=Certainty+Equivalence

#4

Kace wrote: In Bob Dancer's recent column I was shocked that an EV
difference

of .000033 was too small for Bob to worry about.

Shocked, huh? Perhaps you don't get out much.

In actuality, I've never played the game in question (9/5 White Hot
Aces). In the very similar 9/5 Triple Bonus Poker Plus, however, I DO
insist on making the correct play --- because there the EV difference is
"only" 0.0005 and twenty of them suckers would cost me a round Lincoln.

This happens to be an easy play for me to understand, so while I don't
worry about it, I make the correct play in these cases. The situations
where I guess on "close" situations are those where I'm unsuccessful at
simplifying the strategy into rules of thumb. Take a suited '89' with an
unsuited Q in NSU, for example. I do have the accurate strategy of

'89' with Q AK, A3, K3, 34 hold '89' when three remaining
cards are unsuited
      A4, K4 hold '89' when all unsuited, or
single high card suited with 4
      
In addition to this "rule", I have similar rules prepared for the other
2-card straight flushes in the range of '67'-'9T' along with all
possible straight penalties, which consists of almost 100 different
cases, some have to be all unsuited, some where the suit doesn't matter
(so long as there is no flush penalty to the 2-card straight flush), and
sometimes certain types of suitedness are allowed. But it doesn't all
flow for me. This is a case where I made up a rule of thumb (hold the
perfect 2-card straight flush when there is at most one at-the-extreme
straight penalty when the other three cards are unsuited with each
other.)

This rule covers most cases, but it misses some (both by including too
many and excluding some that shouldn't be excluded) and occasionally the
difference in EV on my omissions is as high as 0.006. This is HUGE
insofar as I'm concerned, but the price of actually looking up the
correct answers at the machine is even "more huge."

To counterbalance this, every month or so I practice the "special cases"
in NSU for about a half hour before I go play. Over time, I'm memorizing
all of the special cases and so don't always have to use the rule of
thumb, but that's the best I'm willing to do.

I hope this doesn't shock Kace too much. . .

Bob Dancer

For a 3-day free trial of Video Poker for Winners, the best video poker
computer trainer ever invented, go to //www.videopokerforwinners.com

#5

Iguana,

To use the equation below I must calculate the variance for holding the
ace alone and then the variance for holding just QJ. I would need to
add up all the winning contributions mnultiplied by the frequency of
each contribution for each hold. Is that correct?

Yes. If you can calculate EV you can calculate Variance. EV is the sum
of all possible wins times their probability, Variance is the sum of
the squares of all possible wins minus EV times their probability.
This works for any gamble, with one caveat, it is an approximation,
valid only if your bet size is much less than your bankroll, see:
http://www.bjmath.com/bjmath/kelly/break.htm

Does any of the VP
software have this capability?

I don't know. Sometimes you just have to do things by hand, or by
spreadsheet. FVP has the capability to calculate total EV and Variance
for a user generated strategy, which is useful for playing with
alternate strategies:
http://members.cox.net/vpfree/FAQ_S.htm

···

--- In vpFREE@yahoogroups.com, "kcace1024" <cy4873@...> wrote:

Thanks

--- In vpFREE@yahoogroups.com, "nightoftheiguana2000"
<nightoftheiguana2000@> wrote:
>
> --- In vpFREE@yahoogroups.com, "kcace1024" <cy4873@> wrote:
> > In Bob Dancer's recent column I was shocked that an EV difference
> > of .000033 was too small for Bob to worry about. Now I wonder what
> > about the variance between the 2 choices. Is holding the QJ
instead
> > of the A a better choice because it reduces the variance. I am
> > assuming that QJ will lower the variance, but can someone show me
how
> > to do the calculation.
>
> Certainty Equivalence = EV - Variance/(Bankroll x 2)
>
> http://www.google.com/search?q=Certainty+Equivalence
>

#6

Many months ago some of the experts here showed me there are other
strategies besides max ev. I did not think any of it would apply

to

my play. Recently I have started playing some progressives and I
wonder if there is a way to evaluate increased EV versus increased
variance at strategy break points. Does anyone wait for a certain
amount of increased EV before accepting the increased variance.

There is a similar concern in blackack when the (max ev) index
number for splitting tens has been reached. If you are already
making Kelly-sized bets, you don't want to assume a huge variance
jump for just a small increase in EV, which is what you get when you
split tens at counts just above the index number. A couple of
decades ago, Joel Friedman came up with the idea of risk-averse
indeces for card-counting. Since then, most serious players adopt
some variation of risk-averse strategy, unless their bet level is a
miniscule fraction of their bankroll.

When Stanford Wong first published his strategy for playing 8/5 JOB
progressives, there were indexed thresholds (for the size of the RF)
for tossing made hands. I'm pretty sure those were max-EV
thresholds. I don't know if anyone has published indeces that would
apply to someone playing near the max allowed by his/her bankroll.

--Dunbar

···

--- In vpFREE@yahoogroups.com, "kcace1024" <cy4873@...> wrote:

#7

When Stanford Wong first published his strategy for playing 8/5 JOB
progressives, there were indexed thresholds (for the size of the

RF)

for tossing made hands. I'm pretty sure those were max-EV
thresholds. I don't know if anyone has published indeces that

would

apply to someone playing near the max allowed by his/her bankroll.

--Dunbar

I use Win to find breakpoints. It's pretty easy on the brain. You
can go back and forth between "change payoff table" and "analyze any
hand" easily enough.

I'll use the Wendover Nugget 9/6 Jacks progressive play for an
example:

First of all I only play at $6000 or higher. There's some rare
exceptions, like I need some room comp or something, but my number
isn't gonna go much lower than $6000. The problem with playing too
marginal on progressives is twofold for me. 1. I want a decent
overlay to put my money into fluctuation--I have to have something to
win. 2. There's no worse feeling in the world than playing
marginal, driving the number up, then the hustler's flood onto the
bank snapping the royal off on me.

So since I play at $6000 or higher let's find some breakpoints for
that number in a way that makes it easy for the less mathematically
inclined.

THREE CARD ROYAL VS. HIGH PAIR

I start with the worst of the lot: As-Ks-Ts-7s-Kd. The AQT and AJT
are equivalent. It's the worst of the lot because it has only two
high cards, makes less straights, and has a flush penalty. The other
seven three card royals have higher EV's.

So I go to 9/6 Jacks on Win, hit "change payoff table" then
hit "default" to make sure I got the payscale right. Then I go
to "analyze any hand." I punch in the above hand and get a picture
of how much lower the RF3 is than the high pair. About 30%. So then
I go to "change payoff table" and change the royal to $5000. Then
back to "analyze hand" and punch in the same hand. The RF3 is
getting closer but still not there. Then I change the royal to
$5500. The RF3 is still a tad under. At $5600 it's still a wee bit
under. It's over at $5700. So I back off to $5650. It's still
over. $5625 and it's still over. $5615 it's still over. $5610 it's
still over. $5609, $5608, $5607, $5606. $5605 and Bingo!!! We're
sitting right on top of the breakpoint.

I just found ten breakpoints in one. As I stated earlier I don't
play at under $6000. I don't have to sweat the other RF3's because
their breakpoints are at or lower than $5605. I'm gonna play all
RF3's over high pairs.

How important are the breakpoints above $6000? Not as important as
the ones below. If you know all the breakpoits for $6000 you are
90% there.

Royal Flush odds at $4000....40,391
Royal Flush odds at $6000....33,690
Royal Flush odds at $10,000..32,456

The RF3 over the high pair has the most dramatic effect on royal
odds. Good luck.

···

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

#8

Bob,

I understand your professional approach to get the most out of every
hand. I also appreciate your analysis of various games, penalty
card, situations, etc. I actually play 9/5 TBP+ once in awhile.
My "shock" comment was just a little sarcasm I can't seem to control.

I also enjoyed your response. I think I got another entire column
out of my comment. I probably should get out more.

Chris

Kace wrote: In Bob Dancer's recent column I was shocked that an EV
difference
> of .000033 was too small for Bob to worry about.

Shocked, huh? Perhaps you don't get out much.

In actuality, I've never played the game in question (9/5 White Hot
Aces). In the very similar 9/5 Triple Bonus Poker Plus, however, I

DO

insist on making the correct play --- because there the EV

difference is

"only" 0.0005 and twenty of them suckers would cost me a round

Lincoln.

This happens to be an easy play for me to understand, so while I

don't

worry about it, I make the correct play in these cases. The

situations

where I guess on "close" situations are those where I'm

unsuccessful at

simplifying the strategy into rules of thumb. Take a suited '89'

with an

unsuited Q in NSU, for example. I do have the accurate strategy of

'89' with Q AK, A3, K3, 34 hold '89' when three remaining
cards are unsuited
      A4, K4 hold '89' when all unsuited,

or

single high card suited with 4
      
In addition to this "rule", I have similar rules prepared for the

other

2-card straight flushes in the range of '67'-'9T' along with all
possible straight penalties, which consists of almost 100 different
cases, some have to be all unsuited, some where the suit doesn't

matter

(so long as there is no flush penalty to the 2-card straight

flush), and

sometimes certain types of suitedness are allowed. But it doesn't

all

flow for me. This is a case where I made up a rule of thumb (hold

the

perfect 2-card straight flush when there is at most one at-the-

extreme

straight penalty when the other three cards are unsuited with each
other.)

This rule covers most cases, but it misses some (both by including

too

many and excluding some that shouldn't be excluded) and

occasionally the

difference in EV on my omissions is as high as 0.006. This is HUGE
insofar as I'm concerned, but the price of actually looking up the
correct answers at the machine is even "more huge."

To counterbalance this, every month or so I practice the "special

cases"

in NSU for about a half hour before I go play. Over time, I'm

memorizing

all of the special cases and so don't always have to use the rule of
thumb, but that's the best I'm willing to do.

I hope this doesn't shock Kace too much. . .

Bob Dancer

For a 3-day free trial of Video Poker for Winners, the best video

poker

···

--- In vpFREE@yahoogroups.com, "Bob Dancer" <bdancer@...> wrote:

computer trainer ever invented, go to //www.videopokerforwinners.com

#9

Iguana,

I have used FVP to tweak strategies and find break points for
progressives, but it only gives me the new EV not the variance of the
new strategy. I will go back and check it again. Thanks for your
help.

Chris

In vpFREE@yahoogroups.com, "nightoftheiguana2000"
<nightoftheiguana2000@...> wrote:

> Iguana,
>
> To use the equation below I must calculate the variance for

holding the

> ace alone and then the variance for holding just QJ. I would

need to

> add up all the winning contributions mnultiplied by the frequency

of

> each contribution for each hold. Is that correct?

Yes. If you can calculate EV you can calculate Variance. EV is the

sum

of all possible wins times their probability, Variance is the sum of
the squares of all possible wins minus EV times their probability.
This works for any gamble, with one caveat, it is an approximation,
valid only if your bet size is much less than your bankroll, see:
http://www.bjmath.com/bjmath/kelly/break.htm

> Does any of the VP
> software have this capability?

I don't know. Sometimes you just have to do things by hand, or by
spreadsheet. FVP has the capability to calculate total EV and

Variance

for a user generated strategy, which is useful for playing with
alternate strategies:
http://members.cox.net/vpfree/FAQ_S.htm

>
> Thanks
>
> --- In vpFREE@yahoogroups.com, "nightoftheiguana2000"
> <nightoftheiguana2000@> wrote:
> >
> > > In Bob Dancer's recent column I was shocked that an EV

difference

> > > of .000033 was too small for Bob to worry about. Now I

wonder what

> > > about the variance between the 2 choices. Is holding the QJ
> instead
> > > of the A a better choice because it reduces the variance. I

am

> > > assuming that QJ will lower the variance, but can someone

show me

···

--- In vpFREE@yahoogroups.com, "kcace1024" <cy4873@> wrote:
> > --- In vpFREE@yahoogroups.com, "kcace1024" <cy4873@> wrote:
> how
> > > to do the calculation.
> >
> > Certainty Equivalence = EV - Variance/(Bankroll x 2)
> >
> > http://www.google.com/search?q=Certainty+Equivalence
> >
>

#10

Mickey,

I now use FVP to find the breakpoints, but as usual you bring a sharp
focus to the critical parts of a problem. The 3 Card Royal vs the
High Pair is a great example. I have to rate you as one of my top
teachers. You have the knack for making things easy to understand.
It probably helps that I am interested in the the subject matter.

Chris

>
I use Win to find breakpoints. It's pretty easy on the brain. You
can go back and forth between "change payoff table" and "analyze

any

hand" easily enough.

I'll use the Wendover Nugget 9/6 Jacks progressive play for an
example:

First of all I only play at $6000 or higher. There's some rare
exceptions, like I need some room comp or something, but my number
isn't gonna go much lower than $6000. The problem with playing too
marginal on progressives is twofold for me. 1. I want a decent
overlay to put my money into fluctuation--I have to have something

to

win. 2. There's no worse feeling in the world than playing
marginal, driving the number up, then the hustler's flood onto the
bank snapping the royal off on me.

So since I play at $6000 or higher let's find some breakpoints for
that number in a way that makes it easy for the less mathematically
inclined.

THREE CARD ROYAL VS. HIGH PAIR

I start with the worst of the lot: As-Ks-Ts-7s-Kd. The AQT and

AJT

are equivalent. It's the worst of the lot because it has only two
high cards, makes less straights, and has a flush penalty. The

other

seven three card royals have higher EV's.

So I go to 9/6 Jacks on Win, hit "change payoff table" then
hit "default" to make sure I got the payscale right. Then I go
to "analyze any hand." I punch in the above hand and get a picture
of how much lower the RF3 is than the high pair. About 30%. So

then

I go to "change payoff table" and change the royal to $5000. Then
back to "analyze hand" and punch in the same hand. The RF3 is
getting closer but still not there. Then I change the royal to
$5500. The RF3 is still a tad under. At $5600 it's still a wee

bit

under. It's over at $5700. So I back off to $5650. It's still
over. $5625 and it's still over. $5615 it's still over. $5610

it's

···

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

still over. $5609, $5608, $5607, $5606. $5605 and Bingo!!! We're
sitting right on top of the breakpoint.

I just found ten breakpoints in one. As I stated earlier I don't
play at under $6000. I don't have to sweat the other RF3's because
their breakpoints are at or lower than $5605. I'm gonna play all
RF3's over high pairs.

How important are the breakpoints above $6000? Not as important as
the ones below. If you know all the breakpoits for $6000 you are
90% there.

Royal Flush odds at $4000....40,391
Royal Flush odds at $6000....33,690
Royal Flush odds at $10,000..32,456

The RF3 over the high pair has the most dramatic effect on royal
odds. Good luck.

#11

Mickey,
  You're using the hunt and peck long hand method here. Why not take the
easy route. Keying your hand of As-Ks-Ts-7s-Kd into the analyze hand screen,
you will note the value of the best draw, the KK, is 7.6827 and five slots
down, your 3 RF draw of AKT is valued at 6.1980. Subtract the lower value
from the higher value and your remainder is 1.4847. On the same hand details
screen right next to AKT hold, the 1,081 represents the number of possible
draws left in the remaining cards when rank, suit, and order are considered.
I know that you know this number, but I am just pointing it out for the
novices that might not be aware. Multiply your 1.4847 difference in hand
values by 1,081, and you get 1604.9607. This is he exact number of coins
that you would have to add to the 4000 royal to make these two holds
identical in value, or 5,604.9607, or rounded up to 5,605. Using this method
and a calculator, you can eliminate all the guess until you hit it amounts
approach and save quite a bit of time.
                                                        Nudge

Then I go

···

From: "mickeycrimm"
Subject: [vpFREE] Re: Some questions on variance and EV

to "analyze any hand." I punch in the above hand and get a picture
of how much lower the RF3 is than the high pair. About 30%. So then
I go to "change payoff table" and change the royal to $5000. Then
back to "analyze hand" and punch in the same hand. The RF3 is
getting closer but still not there. Then I change the royal to
$5500. The RF3 is still a tad under. At $5600 it's still a wee bit
under. It's over at $5700. So I back off to $5650. It's still
over. $5625 and it's still over. $5615 it's still over. $5610 it's
still over. $5609, $5608, $5607, $5606. $5605 and Bingo!!!

#12

sure. as it happens i was just looking at this the other day.

suppose you're playing $1 9/6 jacks with a $5500 prog meter, and your
bankroll is $40000. you're dealt A(AQT)7. hold the aces, or draw to
the royal?

someone who's blindly maximizing EV will do the following computation,
or more likely, have their computer do it for them. they'll see that
there are 16215 possible draws off the AA, of which 11559 don't
improve and return $5, 2592 make two pair and return $10, etc:

AA: EV = 1/16215 * (11559*5 + 2592*10 + 1854*15 + 165*45 + 45*125) = $7.68
AQT: EV = 1/1081 * (786*0 + 207*5 + 21*10 + 7*15 + 15*20 + 44*30 +
1*5500) = $7.84

easy, draw to the royal!

but suppose you're a kelly bettor. you want to optimize your rate of
bankroll growth, and you know that the way to do that is by maximizing
log-utility, not flat EV. now you're doing a different computation. if
you hold the AA, 11559 times you're ending up with $40005, 2592 times
you're ending up with $40010, etc:

AA: UTIL = 1/16215 * (11559*log(40005) + 2592*log(40010) + ... +
45*log(40125)) ~= log($40007.68)
AQT: UTIL = 1/1081 * (786*log(40000) + 207*log(40005) + ... +
1*log(45500)) ~= log($40007.51)

easy, hold the aces!

you don't have to go to these lengths to apply this idea in practice.
log(X) is pretty damn close to linear for 40000 <= X <= 40250 ... in
other words, you can ignore the difference between dollars and
log-utility for anything below the royal or other big jackpots.
instead, just replace the face value of the royal (R) with
S*(log(B+R)-log(B))/(log(B+S)-log(B)), where B is your bankroll and S
is your bet size.

in this case, R=5500, B=40000, S=5, and you get a utility-adjusted
royal value of $5154. if you want to play nice by kelly, use that in
your strategy computations instead of $5500.

as expected, you'd want to play more conservatively than this if your
bankroll is smaller (B=20000 leads to a $4860 value), or more
aggressively if your bankroll is larger (B=80000 leads to a $5319
value). but even a very well bankrolled player should be playing a
little more conservatively than straight EV considerations would
suggest.

cheers,

five

···

On Thu, Apr 3, 2008 at 9:47 AM, kcace1024 <cy4873@hotmail.com> wrote:

Many months ago some of the experts here showed me there are other
strategies besides max ev. I did not think any of it would apply to
my play. Recently I have started playing some progressives and I
wonder if there is a way to evaluate increased EV versus increased
variance at strategy break points. Does anyone wait for a certain
amount of increased EV before accepting the increased variance.

#13

Seems like it wouldn't be that hard to generate a table of Kelly
breakpoints (instead of the more common EV breakpoints). Due to the
fixed bet nature of video poker, you need a bankroll range that you
play in, I would suggest 1x to 4x Kelly (if your bankroll falls below
Kelly you have to find another game of half the current denomination,
if your bankroll grows beyond quad Kelly, you should try to find
another game of double the current denomination). So, the Kelly
breakpoint table would have two values, one at Kelly bankroll and one
at quad Kelly bankroll.

···

--- In vpFREE@yahoogroups.com, fivespot <fivespot55@...> wrote:

you don't have to go to these lengths to apply this idea in practice.
log(X) is pretty damn close to linear for 40000 <= X <= 40250 ... in
other words, you can ignore the difference between dollars and
log-utility for anything below the royal or other big jackpots.
instead, just replace the face value of the royal (R) with
S*(log(B+R)-log(B))/(log(B+S)-log(B)), where B is your bankroll and S
is your bet size.

#14

Fivespot and Iguana,

Thanks for your replies. My play on progressives has been minimal so
far. However, one of my local casinos no longer has any playable VP
except for progressives. If I do start down this path I want to do
it in a conservative manner.

How do you factor the base game reset value into an evaluation. Is
7/5 bonus poker too low even at 101%? Does anyone play progressives
between 100% and 101% to catch bounceback coupons, promotions, and
comps? The opportunities for play on the various progressives may be
marginal, but I am trying to be patient and wait for the best
situations. Thanks again to all the advantage players willing to
share their expertise.

Chris

--- In vpFREE@yahoogroups.com, "nightoftheiguana2000"
<nightoftheiguana2000@...> wrote:

> you don't have to go to these lengths to apply this idea in

practice.

> log(X) is pretty damn close to linear for 40000 <= X <= 40250 ...

in

> other words, you can ignore the difference between dollars and
> log-utility for anything below the royal or other big jackpots.
> instead, just replace the face value of the royal (R) with
> S*(log(B+R)-log(B))/(log(B+S)-log(B)), where B is your bankroll

and S

> is your bet size.

Seems like it wouldn't be that hard to generate a table of Kelly
breakpoints (instead of the more common EV breakpoints). Due to the
fixed bet nature of video poker, you need a bankroll range that you
play in, I would suggest 1x to 4x Kelly (if your bankroll falls

below

Kelly you have to find another game of half the current

denomination,

if your bankroll grows beyond quad Kelly, you should try to find
another game of double the current denomination). So, the Kelly
breakpoint table would have two values, one at Kelly bankroll and

one

···

--- In vpFREE@yahoogroups.com, fivespot <fivespot55@> wrote:
at quad Kelly bankroll.

#15

Mickey,
  You're using the hunt and peck long hand method here. Why not

take the

easy route. Keying your hand of As-Ks-Ts-7s-Kd into the analyze

hand screen,

you will note the value of the best draw, the KK, is 7.6827 and

five slots

down, your 3 RF draw of AKT is valued at 6.1980. Subtract the lower

value

from the higher value and your remainder is 1.4847. On the same

hand details

screen right next to AKT hold, the 1,081 represents the number of

possible

draws left in the remaining cards when rank, suit, and order are

considered.

I know that you know this number, but I am just pointing it out for

the

novices that might not be aware. Multiply your 1.4847 difference in

hand

values by 1,081, and you get 1604.9607. This is he exact number of

coins

that you would have to add to the 4000 royal to make these two holds
identical in value, or 5,604.9607, or rounded up to 5,605. Using

this method

and a calculator, you can eliminate all the guess until you hit it

amounts

approach and save quite a bit of time.
                                                        Nudge

This is a very nice trick. Thanks for the clinic. I used the hunt
and peck to demonstrate to those that won't or can't do the math that
they are not left out in the cold. They can find breakpoints without
using math via the hunt and peck.

···

--- In vpFREE@yahoogroups.com, "nudge51" <nudge51@...> wrote:

#16

If I do start down this path I want to do
it in a conservative manner.

The most conservative strategy is minimum risk strategy. Steve Jacobs
gave it for 9/6 JOB, it's on the FAQ, for typical progressive values
there are no breakpoints (strategy does not change).

How do you factor the base game reset value into an evaluation. Is
7/5 bonus poker too low even at 101%?

Well, depends. The variance is high, which means the bankroll is high.
Basically doing the bankroll calculation is how you factor in
different games. The math is the math. If you're unhappy with the
bankroll number, don't play the game. If a game has a high bankroll,
it has a high bankroll for a reason.

Does anyone play progressives
between 100% and 101% to catch bounceback coupons, promotions, and
comps?

Sure, why not? Well, assuming there's nothing better, and the
promotions make it worthwhile. As Mickey pointed out, there is the
issue of protecting your meter input. If you're only playing enough to
put say 5 bucks on the meter, it's not worth losing sleep over, but it
sure would be a shame if you banged away all day only to have the
locals come in over night and snap off your contribution to the
progressive. Not only would it be a shame, it would mean you are not
getting your share of the meter, somebody else is getting your share.

···

--- In vpFREE@yahoogroups.com, "kcace1024" <cy4873@...> wrote:

#17

kcace1024 wrote:

Many months ago some of the experts here showed me there are other
strategies besides max ev. I did not think any of it would apply to
my play. Recently I have started playing some progressives and I
wonder if there is a way to evaluate increased EV versus increased
variance at strategy break points. Does anyone wait for a certain
amount of increased EV before accepting the increased variance.

A few general comments, for the sake of added perspective.

Progressive variance is just another dimension on the variance
challenge posed by any vp play. For example, for a given bankroll,
when is 10/7 DB (or a strong STP game) a good play? When is it best
left alone?

It's all a question of bankroll strength and personal taste for risk.
Other opportunities for play are to be factored.

Confronted with a play that presents a challenging ROR, the question
of adjusting strategy to reduce bankroll risk often arises. If you're
looking at a play where the jackpot represents 8%+ of the game return,
there can be some merit into looking into a strategy shift that
reduces bankroll risk (or, for example, the average expected loss
between jackpots).

However, I suggest that where the jackpot represents 5% or less ER
such concerns signal a good likelihood that the game simply is beyond
your risk tolerance. Certainly you'll find that if you run the
numbers and look at the ROR bankroll requirements (for max-EV strategy
vs any other risk-reduced strategy) it'll likely be the case that the
difference in respective bankrolls is of a magnitude that in comparing
any two non-progressive plays you wouldn't blink twice at the difference.

Max-EV strategy isn't the holy grail, but I'm not convinced outside of
extraordinary circumstances that there are any appreciable ways (via
strategy) to vent the heat of a hot play.

- Harry

#18

The point multipliers of 5x are still a constant for Wednesdays and Sundays. The marquee
has it advertised and my play last Wednesday, the 2nd, verified same.

My mailers have not been reduced from the $30 once their system of twice a week bounceback started.
I did not assume that 3K would be the level but mixed it up a bit by playing more (4-6K) on some days especially the 10x point days. I also dipped
to maybe 2 - 2.5K but with an overall average of 5K.

I feel that a obvious 3K, 3K each session is now undesirable and tags one definitely as "those advantage players". At 5x, the 3x5xBP bank at the center
bar was available. The off limit exclusion sign right now is only for 10x point days.

Most of us received an erratum special mailer clarifying the many mistakes of the original April calendar. I can't speculate at this time what the true
level is to retain the $30 bb level. I personally think it is unwise to query of the marketing or
slot club personnel what are their tier divisions, levels and such. It alerts and verifies to management the growing number(s) of players that are
systematically looking for the minimum handle level for maximum rewards.

Anteroz

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

#19

>
This is a very nice trick. Thanks for the clinic. I used the hunt
and peck to demonstrate to those that won't or can't do the math that
they are not left out in the cold. They can find breakpoints without
using math via the hunt and peck.

Mickey, I think the way you descibe things is often much easier for
those of us that aren't math experts to get...... you obviously have a
very firm grip on the math of VP but maybe more in a common folk or
reasonable deduction kinda way.
when the math experts on here get going I have to admit sometimes my
eyes glaze over.....I'd like to be so good a math to always "get" what
they are saying but that just isn't the case :o)
so thanks for explaining things in an easier to understand way!
trish