Systems Archive 1
[ Read Responses | Return to the Index ]
[ Previous | Previous in Thread | Next in Thread | Next ]
Re: correction
Posted by Rainman on Monday, 15 December 1997, at 8:05 a.m., in response to correction, posted by Danny Woolard on Thursday, 11 December 1997, at 4:16 p.m.
Danny,
Thank you for your response. Yes, the system I have in mind is where you track each individual rank using NO primary count whatsoever. I'm not looking for simulations to generate the indices, merely to test the effectiveness of the system: e.g., find win rates with various beting strategies (including depth charging), standard deviation, etc. As for the indices: I believe that Griffin's actual removal can be used, rounded to the nearest whole number.
Actually, I've used the system similar to which you are referring along the way to developing my skills for the full ten rank count. You're re-inventing the wheel. The "Heath" count you refer to was first (perhaps independantly) developed by Edward Gordon. Gordon counts the 2,3,4,5 vs the Tens (+1, -1), the 9,8,7,6 vs Tens, and a side Ace count. Note: that's TWO separate RUNNING counts, plus a side Ace count. In later versions of Gordon's count, he added the 9 vs 6, and 5 vs 2 (a "difference" count such as this, incidentally, is statistically independant of the primary count, see Griffin TOB, p. 65). He has generated a full set of indices for this count and has some simulation data, so if you're interested I suggest you contact him through the Gambler's Book Club as I did.
When first fooling with this multiparameter stuff, I worked up the count Griffin desribes on p. 60 of TOB (the "compleat fanatic" count). I used the Gordon primary count (2,3,4,5, vs Ts) with a side individual count of each 9,8,7,6 and Ace removed, adjusting for 1/4 deck density. One nice thing about Gordon's count (a.k.a., the D.H.M. ultimate, by the way--D.H. Mitchell is Gordon's alias) is that it does not require 1/4 deck adjustments and can be used for multiple deck play. The dificult thing, of course, is that we're now talking about keeping FOUR RUNNING COUNTS at once, plus a side Ace count! Personally, I find it much easier to track each rank with the method I described in my first post.
Although the count is supposed to be .922 efficient, Imming's UBE multiparameter version showed it to be inferior to the Hi-Opt II with side 9,8,7,6,5,A counts. The fully loaded Hi-Opt II (Imming calls it the Imp1) uses cards that are included in the count and which, by academic thinking, should hurt the correlation. But, surprisingly, this doesn't seem to matter. Apparrently, it makes no difference that the side counted card is included in the primary count. This being the case, it also seems that the side counts add about the same to the system's efficiency regardless of which count is used. Thus, if you want to use a series of side counts, you want to use as strong a playing primary count as you can manage. I'm really considerring going with the fully loaded Hi-Opt II unless I find that the full ten rank count works better.
I will be contacting you by e-mail.
Sincerely,
Rainman
Responses
The Systems Archive 1 is maintained with WebBBS 2.24.060398.