I apologize for breaking into this thread so late in its developmekktmdI say that because:
1.) I have no naked babe screen shots.
2.) I'm responding to two posts made by kktmd way back on July 24 and July 25. (Who would have expected to find statistical data analysis in a thread about a Playboy slot?)
The First Post contained chart data, one column of which was labeled "Variance".
The Second Post contained the explanation that the chart data was from a MySQL database.
It also responded to Balthazar's question: with the answer:
@kktmd:
I
Started a Thread awhile ago to discuss, among other things, slot variance as a number rather than a word (high, medium, low).
The thread begins to pick up steam concerning this specific topic on page 6. Jufo's contributions to this thread were inspirationally extraordinary. (Check out
The Following Post.) It's a bit of a long thread, and there is a lot of pretty meaty stuff in there. But I can tell you that the whole thing was a tremendous learning experience for me.
The values for some of the various slot performance variables that we published (follow the link on the thread's first post) came from an analysis of the PAR sheets for each of the slots. The variance and standard deviation values came from a database analysis, using samples of 20 million rounds per slot.
The equation for calculating the variance, which is at the bottom of the thread's first post, came from Mike Shackleford. I'll quote it here for reference:
Example data for one of our single-line slots:
WRC = 20000000
WRB = 3
S1 = 58547877
S2 = 4411176273
If you put these numbers into Excel, and use the formula above, then you should see:
Variance = 23.55
Standard Deviation (= Square Root of Variance) = 4.85
One of the major points of discussion in this older thread was - OK, we have a number. It's an absolute number for this slot, like TRTP. But what, if anything, does it mean?
(As to its being an "absolute number" for a slot - Our original sample run for the analysis was 20 million rounds, yielding a variance of 23.18. We have since run another 20 million round sample, which yielded a variance of 23.55.)
As I mentioned above, the really tricky bit in all of this is dealing with the Free Spin data for the S2 calculation.
Example: Standard spin, 20 lines, $1 per line, spin returns $6 and 10 free spins.
- Free Spin 1 returns 0
- Free Spin 2 returns 30
- Free Spin 3 returns 10
- Free Spin 4 returns 0
- Free Spin 5 returns 250
- Free Spin 6 returns 0
- Free Spin 7 returns 0
- Free Spin 8 returns 60
- Free Spin 9 returns 10
- Free Spin 10 returns 0
Sum of return amounts (including the original spin that won the free spins) = 366. The square of this is 133956. This 133956 value is what gets added to the S2 value. (In this example, WRB = 20.)
You can do this in SQL. You'll need a scrollable cursor, an IF statement (to catch the free spin win round) and a WHILE statement (to iterate through the subsequent free spin rounds).
Or you can pull the data using a SELECT statement, load it into a multi-dimensional array (a rather huge array) and do your analysis using the array's data.
I can tell you that I, and I would presume to say several other CM members, would be
extremely interested in seeing your results.
Chris