BlackjackInstitute.com

[MJ]

A poster on another board claims that SE WR= SD WR/sqrt(N), where N = (#rounds simulated / #rounds per hour). Thus N = 2Billion/100 = 20M. So, the 100 round experiment was repeated 20M times.

The good news is that the programmer recently added SE for three different categories. This means that we can easily put the proposed formula for SE to the test and see how it holds up!

Ok, now let us examine the charts. One 2 billion round (10^9) simulation was run. Bet spreads for the charts were changed post sim. The first chart shows a 20/1 spread and the second chart shows a 40/1 spread; other than that the rest of the parameters are the same. Note that the SE is provided in the upper right columns of each chart.

In the first chart, using the proposed formula, SE for WR = $314/(sqrt 20M) ~ $0.07. But wait a minute, the software reports SE WR as $0.19!!

OUCH!! The formula understates SE WR by 63%!!

I wish I could say that the accuracy of the formula improved looking at the chart that used a 40 to 1 spread, but I can't. The formula gives SE WR ~ $0.07 whereas the software has it at $0.28. The formula grossly understates SE WR once again, this time by around 75%!!

Notice that SE given by the software increases as bet spread increases. The formula given by this poster does not even take bet spread into account!

There are only 3 possibilities to explain the discrepancies for SE:
1) The software gives incorrect SE
2) The poster gave a bogus formula
3) Both the software and this poster's formula are incorrect

On a side note, I would like to thank this poster for pointing out some flaws in the earlier part of this experiment, which I erased and will redo at a later time.

MJ



_________________
"Thou shalt NOT cheat. Thou shalt, however, take FULL advantage of dealer and casino errors. Thou shalt NOT tip dealers" -JP