I think I understand what variance is - the lower it is the
less "streaky" the game is. However, I dont understand the numbers. For
example, the variance of 9/6 JoB is 19.51 as oppossed to 25.78 for
NSUD. How are those numbers arrived at? What do they mean?
Thanks for any help.
martinvegas wrote:
I think I understand what variance is - the lower it is the
less "streaky" the game is. However, I dont understand the numbers.
For example, the variance of 9/6 JoB is 19.51 as oppossed to 25.78
for NSUD. How are those numbers arrived at? What do they mean?
Variance is one of the less useful of all discussed vp concepts. In
fact, beyond your stated understanding, there isn't much of great
value to be said.
The mechanics of the variance calculation are straightforward. Given
a data distribution (such as the ages of people in a room), for each
data point you subtract the mean value from the value for that data
point. Square this value, do the same for all other data points, sum
the squares and then take the square root of the sum. You now have
the variance for that population.
When it comes to video poker, the data points are the possible wins
from a hand. You express the values in "bet" units, where a payout of
10 = 2 bets when you wager 5 credits. The variance calculation varies
slightly in that you weight the squared differences from the mean by
their probabilities and sum the result.
This gist of this calculation is that where data is widely dispersed
from the data mean, the variance is high. Where data falls close to
the mean, variance is low. When talking about data that is "normally"
distributed (such as the number of 6's that come up on a die when you
roll it 1,000 times), the variance statistic by itself will tell you
the shape of the distribution curve of "6" frequency and also gives
you a meaningful comparison of dispersion against other normal
distributions.
However, video poker probabilities involve numerous payout outcomes
from playing a hand, each with disparate probabilities. The
consequence is that the vp payout distribution isn't normally
distributed. In this case, variance is far less telling about
probabilities and how one game compares against another.
It's sometimes observed that pick'em has the lowest of vp variety
variances, yet over the course of a few hours of play a player is
subject to greater loss risk than if they were playing 9/6 Jacks.
All this said, what's important to take away is exactly what you
stated up front: the lower the game variance, the less "streaky" the
game is.
- Harry
Harry Porter wrote:
The mechanics of the variance calculation are straightforward. Given
a data distribution (such as the ages of people in a room), for each
data point you subtract the mean value from the value for that data
point. Square this value, do the same for all other data points, sum
the squares and then take the square root of the sum. You now have
the variance for that population.
I slipped. Stop at the sum. (FWIW, taking the square root yields the
"standard deviation".)
Thanks for the reply Harry. Appreciated.
Regards
Woe be unto the VP player who ignores variance!
An EV > 100% is not enough. You must know the game's variance (or better yet the entire
PDF)-- to choose which game to play (and how much to bet, etc.).
Extreme Example:
Double Bonus Double Jackpot (DBDJ) has an EV of 100.09%
But don't rush out and cash in your IRA to play it just yet.
It's variance is 92.7. Sure, it's RF cycle is a modest 33K hands.
But with a variance of 92.7, you are going to go broke playing it.
OK-- I am being presumptive. I guess I should put it this way:
you need alot more money to play DBDJ for a couple of hours than you do for many
negative EV games. Yes, negative EV games. Ignore the variance here, and you could be
making a big mistake. EV
isn't everything (I hope I don't start something with that statement. LOL).
Moral: Don't ignore the variance. Understanding it and its role in gambling (and VP) is of
tremendous value. If it was zero, and the EV <100, we could never win. On the other
hand, if it was too big, and the EV>100, we might never win-- but we could win.
Somewhere in there lies the balence of EV and variance that makes gambling gambling.
--- In vpFREE@yahoogroups.com, "Harry Porter" <harry.porter@...> wrote:
Variance is one of the less useful of all discussed vp concepts. In
fact, beyond your stated understanding, there isn't much of great
value to be said.
cdfsrule wrote:
Moral: Don't ignore the variance.
"Less Useful" doesn't equal "Ignore".
My point (expressed in "15 words or less"): Don't stay up all night
trying to figure out what to do with it.
The poster had the concept nailed in one sentence.
- H.