--- In vpFREE@yahoogroups.com, Barry Glazer <b.glazer@...> wrote:
> I wonder how much the likelihood of hitting a royal would have to be reduced by the machine in order to change the expected return to the player by 1% -- it is likely that a slight reduction in frequency of the "big event" (by slight, I certainly don't mean a very small reduction in frequency, but one that would not be easily noticed, given that the event is infrequent to begin with) would easily reduce the return significantly, while not getting too much suspicion (as is the case if something NEVER hits and is more frequently expected, as was the concern expressed in the post below).
If the average royal cycle is 40,000, the return from the royal is 2% (800/40000), so you'd need to adjust the cycle to 80,000 to remove 1% from the total return. At the point where the 2sd's touch; x/2+2sqrt(x/2)=x-2sqrt(x); x=47, =47x80,000 = 3.8 million hands with 2.5% overlap or error rate. So, at 3.8 million hands you could tell the difference with 97.5% confidence.
If I remember correctly, the gaff in the Vegas American Coin Scandal prevented max coin royals from ever happening on bartop progressives, which was eventually noticed by the bartenders and players.
http://www.google.com/search?q=Vegas+American+Coin+Scandal
http://www.google.com/search?q=Venetian+rigged+drawing
http://en.wikipedia.org/wiki/Ronald_Dale_Harris
