Small error and the wonders of technology..
by
, 02-28-2013 at 12:04 AM (46547 Views)
In my re-derivation of the Patrick Sileo result ROR = e^(-2Bank/EKB) in my version of the Optimal Betting Spread proof found here
http://www.blackjackforumonline.com/...ng_Spreads.pdf , some tinkering with my WolframAlpha App on my Ipad I discovered a pleasing error.
The offending expression was that equation (75) was that the quantity Y = e^2 + O(1/sqrt(n0)). I was playing with the expression (75) using Mathematica online, and it actually gave me the expansion
Y = e^2 { 1 + 2/(3 n0) + O(n0^-2) }
It always slightly worried me that I assumed the next term was of order 1/sqrt(n0) and that n0 of the order of a 1000 may actually deviate from the simplified expression. Now I don't have to worry anymore as it is of order 1/n0 - must have come from the cancellation in eq(75)
The mistake will be corrected in the Brh Systems Book immediately.
Brett.