Re: Thinking of taking the plunge into Professional Poker

On Jun 1 2008 3:59 PM, Preston wrote:

On Jun 1, 12:25 pm, A Man Beaten by Jacks <nob...@xxxxxxxx> wrote:
On Sun, 1 Jun 2008 01:58:37 -0700 (PDT), Preston

<prestonpoul...@xxxxxxxxxxx> wrote:
Hey Everyone,
Here's the May update. In May I played 145 hours and earned $5903.
That's roughly $40 per hour which is right in line with my longer term
average of 1.3 big bets per hours at $15-30 Hold 'em. In total, that's
424 big bets won over the last 335 hours of playing. I'm not sure how
this isn't fulfilling the goals that I set out to do as Irish Mike
suggests, given that they are my goals and I'm the one that set them
and I seem to have fulfilled them all, but I'm sure that Irish Mike
and all the other usenet haters are, of course, correct.
At any rate, I've now statistically proven that I'm better than break
even and can continue to post updates here about my progress as I've
moved up limits and won more money, but, as Irish Mike points out, who
on Earth would want to hear that on a forum like rec.gambling.poker ?
So this will probably be my last post here.

The fact that you think you've proven that shows that you have little
understanding of statistics.  The very phrase "statistically proven"
in conjunction with 335 hours shows a severe lack of understanding. At
most, you've shown it is more likely than not that you are better than
break even.

Anyway don't let the door hit you on the ass on your way out.

Opps, sorry, had to come back because "Man beaten by Jacks" has been
quit this whole time and he only responds now.
Just so you know "Man" I got an A in a college level course on
statistics. I've started breaking up my play into 100 hour groups and
comparing the variance between the hourly rates of a hundred hour
block. Right now, that's only there data points. But you can do
statistics on three data points, you just have to use the "Student t
Distribution" to generate a confidence interval. Because the data set
is so small the CI will be large, but it will come down over time as
more and more data is put into it.

Sigh.

I took more than one college level courses in statistical methods and more
than one college level course in math stat. In both cases at both
undergraduate and graduate levels. I've also taught that college level
course you're talking about.

So, slow down cowboy.

The t-distribution describes the sampling distribution for a normally
distributed random variable with unknown variance. You're estimating two
parameters from the data -- mean and variance. Making that estimate from
3 data points doesn't leave you with a lot of degrees of freedom. You
were probably taught that you have 2 degrees of freedom, that's for the
sampling distribution used to estimate the mean. You're trying to make a
whole bunch of inferences past that -- which burns up those 2 degrees of
freedom pretty fast.

But, then again, I don't know what the hell you're talking about with the
"blocks of 100". Is 3 blocks of 100 a sample of 300 or a sample of 3?



Here's the data so you can follow along at home.
Average Big Bet per hour win for 100 hours played
.94
1.68
1.12

The variance of this sample is .14894, the standard deviation is .
38593

This is awful. You got an A? What was it, a stat methods course in a
sociology department?

You're throwing away all the variance when you do those group means.


The Confidence interval can be figured using the Student t charts
choosing 95% confident with 2 degrees of freedom (degrees of freedom =
number of data points -1). According to the chart I have, the value is
4.303

So we are 95% confident that the true average for my big bet per hour
winnings should be somewhere between the range given by:

1.245 (my current average) plus or minus 4.303 x the standard
deviation of .38593 divided by the square root of the number of data
points

Crunch the numbers and you get 1.245 plus or minus .9587

You don't know what you're doing.



Therefore, as I said, I have statistically proven with 95% confidence
to be at least better than break even. So, Man Beaten by Jacks, you
are once again wrong. Furthermore, you have once again proven that you
haven't the slightest idea what you're talking about.

What you've demonstrated is that if you ignore most of the variance (i.e.,
use 3 group means instead of a single mean from 300 observations) that you
can make it look like there's not much variance in poker results.

You would not have gotten an A when I taught the course.

-------- 
* kill-files, watch-lists, favorites, and more.. www.recgroups.com

.

Relevant Pages