vpFREE2 Forums

Perfect Play

10a. Re: Perfect Play
Date: Thu Jan 29, 2009 4:50 pm ((PST))

> I've recently starting using Bob Dancer's video poker software, I find
> that I can play perfectly for stretches, but less then 1 percent of
> the time, I make errors. I know what to do but I still manage to miss
> things.
>
> Do you actually play perfectly or do you just assume you play
perfectly?
>
> Do you or should you factor some small percentage for errors into your
> profit calculations?
>

Knowing correct strategy and playing correct strategy are, as you observe, two different things. You need to SEE the hand correctly and then ACT correctly. I think it is pretty well established that our perceptive skills as human beings are not perfect, so I very much believe that no one plays perfectly.

It is possible for a player to play perfectly on a computer and verify it.

In my opinion...
It is impossible for a player to play perfectly in a casino setting AND TO VERIFY IT.

So if anyone tells you they play perfectly in the casino, ask them how they know.

While we certainly MAY be aware of them when they occur, we often ARE unaware of our errors, or at least some of them, and you can't report a mistake you didn't know you made.

It is certainly possible to make mistakes and admit to those, which would be a credible report of IMperfect play, I'd think (again, in a casino setting).

Only a properly programmeed computer can play perfectly in a casino setting and verify it - and that would require some non-standard technology to be applied.

That said...

You should KNOW that profit will be less in the long run if you play less than perfectly, but how to calculate it is tricky. It depends on what mistake you are making. If it's choosing one play over another one that is very marginally superior, not that big a deal, and you could even make that mistake routinely and calculate the small amount it costs you.

For example, if you NEVER consider penalty cards (many do not) it's possible to determine the expected return of your strategy vs. someone who uses penalty cards correctly.

If you don't notice dealt quads or flushes on occasion and sometimes break them up, that's more costly. If you make a small mistake routinely (or occasionally for that matter) on a very common hand, that's costly (e.g., JoB, not seeing the second pair you're dealt). If you don't know exactly which mistake you're making and exactly what frequency the mistake occurs (i.e., you don't always make it), there's not enough data to calculate the cost of the mistakes.

Only if you know exactly what mistake(s) you're making and you make it / them with a predictable frequency could you "factor in" an adjustment to your expected profit / loss.

A 1% error rate does not necessarily mean you'll make 1% less on your coin-in.

And earning more or less than the calculated return obviously does not necessarily mean you're playing worse (or even more obviously, better) than correct strategy!

VP is a game with an element of chance and an element of skill. In the short run, "luck" (the element of chance) will be a big factor in how you do. In the long run, "skill" will be the big factor, and mistakes reduce your skill level, just as not knowing the correct play reduces it.

--BG

···

===================

Look at the Math. For simplicity, let's say as you mention you only make errors 1% of the time. Assume you are at 100% game. Therefore you are at least at 99%, but that assumes that the error you make causes a 1% loss. Not possible unless you deliberately misplay a hand so bad that you turn a gtd winner in to a gtd loser. Usually the decision where you check the card might make a slight percentage difference at most, therefore it is like you get an EV of 75 cents vs 73 cents or similar. Thus u have 2% of 1% which is about 2 cents of 0.02%, in the overall scheme of things 2 cents out of $100 , not very much in reality.

···

--- On Fri, 1/30/09, b.glazer@att.net <b.glazer@att.net> wrote:
From: b.glazer@att.net <b.glazer@att.net>
Subject: [vpFREE] Re: Perfect Play
To: vpFREE@yahoogroups.com
Date: Friday, January 30, 2009, 7:51 AM

            > 10a. Re: Perfect Play

Date: Thu Jan 29, 2009 4:50 pm ((PST))

> I've recently starting using Bob Dancer's video poker software, I find

> that I can play perfectly for stretches, but less then 1 percent of

> the time, I make errors. I know what to do but I still manage to miss

> things.

>

> Do you actually play perfectly or do you just assume you play

perfectly?

>

> Do you or should you factor some small percentage for errors into your

> profit calculations?

>

Knowing correct strategy and playing correct strategy are, as you observe, two different things. You need to SEE the hand correctly and then ACT correctly. I think it is pretty well established that our perceptive skills as human beings are not perfect, so I very much believe that no one plays perfectly.

It is possible for a player to play perfectly on a computer and verify it.

In my opinion...

It is impossible for a player to play perfectly in a casino setting AND TO VERIFY IT.

So if anyone tells you they play perfectly in the casino, ask them how they know.

While we certainly MAY be aware of them when they occur, we often ARE unaware of our errors, or at least some of them, and you can't report a mistake you didn't know you made.

It is certainly possible to make mistakes and admit to those, which would be a credible report of IMperfect play, I'd think (again, in a casino setting).

Only a properly programmeed computer can play perfectly in a casino setting and verify it - and that would require some non-standard technology to be applied.

That said...

You should KNOW that profit will be less in the long run if you play less than perfectly, but how to calculate it is tricky. It depends on what mistake you are making. If it's choosing one play over another one that is very marginally superior, not that big a deal, and you could even make that mistake routinely and calculate the small amount it costs you.

For example, if you NEVER consider penalty cards (many do not) it's possible to determine the expected return of your strategy vs. someone who uses penalty cards correctly.

If you don't notice dealt quads or flushes on occasion and sometimes break them up, that's more costly. If you make a small mistake routinely (or occasionally for that matter) on a very common hand, that's costly (e.g., JoB, not seeing the second pair you're dealt). If you don't know exactly which mistake you're making and exactly what frequency the mistake occurs (i.e., you don't always make it), there's not enough data to calculate the cost of the mistakes.

Only if you know exactly what mistake(s) you're making and you make it / them with a predictable frequency could you "factor in" an adjustment to your expected profit / loss.

A 1% error rate does not necessarily mean you'll make 1% less on your coin-in.

And earning more or less than the calculated return obviously does not necessarily mean you're playing worse (or even more obviously, better) than correct strategy!

VP is a game with an element of chance and an element of skill. In the short run, "luck" (the element of chance) will be a big factor in how you do. In the long run, "skill" will be the big factor, and mistakes reduce your skill level, just as not knowing the correct play reduces it.

--BG

============ =======

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

This is an excellent point. The type of errors you make is a much
more significant factor than the rate at which you make them. Anyone
who has training software can easily assess the impact of their errors
on the theoretical return of the game they are playing.

For example in WinPoker, select "Analyze" then "Overall play to see
the details. Then choose the appropriate game from the drop down
menu. I've played 65,685 hands of NSUD on this game with 98.75% of
the hands correct. So I miss about 1 in 100 hands. So does this mean
I'm loosing an additional 1% of edge. Not at all, my return
percentage is 99.96%. So for every $100 I wager I give up an extra 4
cents to errors.

Some may find this unacceptable, but I'm quite content with this error
rate as I don't bother with penalty cards and more intricate plays
that are required to completely master NSUD. Every now and then I may
make a bonehead play too, but they are infrequent enough as to not
cause me an unacceptable loss of return in the game.

When I analyze playing opportunities I calculate my NSUD return as
99.688%(0.9996*.99728) not the normal 99.73%. I do this for every
type of game I want to pay for any serious stakes to see if it is a
play that I'm willing to entertain. Rarely do I play any game with
less than 1% mistakes, but I also do not play anything where my
effective return rate is less than 99.9% of perfect play. Each person
can adjust the limits according to their individual standards.

By the way the same analysis can be done with VPFW, but one really
annoying limitation of VPFW is that overall play is not segregated by
game type. You have to change users to get different game statistics.
I wish VPFW would have a similar analyze play interface as WinPoker
as I really enjoy seeing hand statistics by game type as well as the
ability to segregate playing results by game type. Also no session
only record is available for VPFW.

SB

Look at the Math. For simplicity, let's say as you mention you only

make errors 1% of the time. Assume you are at 100% game. Therefore
you are at least at 99%, but that assumes that the error you make
causes a 1% loss. Not possible unless you deliberately misplay a hand
so bad that you turn a gtd winner in to a gtd loser. Usually the
decision where you check the card might make a slight percentage
difference at most, therefore it is like you get an EV of 75 cents vs
73 cents or similar. Thus u have 2% of 1% which is about 2 cents of
0.02%, in the overall scheme of things 2 cents out of $100 , not very
much in reality.

···

--- In vpFREE@yahoogroups.com, pesach kremen <royalflush2222@...> wrote: