I set it for KO to study parts of shoe.
ScreenShot291.jpg
Result is
ScreenShot294.jpg
Look at last line. Frequency is weird.
I set it for KO to study parts of shoe.
ScreenShot291.jpg
Result is
ScreenShot294.jpg
Look at last line. Frequency is weird.
I found "optimal ramps by depth" for the KO strategy by running MRI's for each deck in a 6 deck shoe and then looking at the TBA's at each count in each deck depth to figure out optimal ramps. Then I checked my work with CVCX by using the depth-betting feature. Generally speaking, I found I got a little boost in SCORE and win rate by raising my bets earlier deeper into the shoe. I'm happy to share any of my work if you would like. Organizing all the TBA data so you can study it takes a lot of time.
Yes, that should provide some improvement. What I'd like to do is improve upon this by calculating separate ramps by shoe section all at once to optimize an overall SCORE. This is rather messy as any adjustment in any section affects overall SCORE and can affect adjustments in other sections. The additional gain is minimal. I just want to do it when I can find some time -- and I do have a design --but it has to be done without slowing the simulations.
"I don't think outside the box; I think of what I can do with the box." - Henri Matisse
Norm, great.
Don, as you know, with unbalanced counts a single ramp is a compromise.
It is optimal only at the pivot.
For counts below the pivot, it underbet at the begining of the shoe and overbet at the end.
The reverse for counts after the pivot.
MercySakesAlive, thanks. It's actually for a bastard version of spanish 21 that I encounter in cruise.
Hi,Originally Posted by Phoebe
I remember having done something similar with KO several years ago until I ran into a small problem. Let me explain: In scenario A, I calculate the SCORE for a complete shoe from the beginning until the cut card (5/6).
The SCORE calculation defines an optimal betting ramp for a given spread. Let's call this SCORE, S1 (with a ramp 1).
In scenario B, I calculate a SCORE S2 (with a ramp 2) for the first half (2.5 decks) and another SCORE S3 (with a ramp 3) for the next 2.5 decks. Clearly, S3 is greater than S2 due to the "floating advantage".
Now comes the problem. To determine which of these two scenarios is better, we would have to compare S1 against the SCORE produced between S2 and S3, correct?
But the problem is that it cannot be affirmed that the SCORE of scenario B is simply S2+S3. So, what is the optimal way to calculate it? It's obvious that scenario B is better, but how much better is it?
In all cases, I use the same spread (1:16).
Sincerely,
Cac
Luck is what happens when preparation meets opportunity.
No, it would make no sense to think that the SCORE for scenario B is S2 + S3. It's fine to use the metric of SCORE for each half of the shoe, but that value is gained for only a half at a time. So, while you might be tempted to say that the overall SCORE for scenario B is (more properly, I think) (S2 + S3)/2, this has the "feel" to me of a root-mean-square calculation. Try sqrt(S2^2 + S3^2) for the SCORE of scenario B.
Don
Hi Don,
I've also considered that root-mean-square, but my problem is that I'm not sure which "mean" is the most accurate. If it were EVs, it would be simpler since the problem is solved by adding EV1 + EV2.
I thought about adding the SCOREs as if it were a sum of Win Rates, that's why I thought of S2 + S3. Or perhaps SCORE isn't the most suitable for this calculation. Let's consider that S1 plays 100
hands per hour while S2 and S3 each play half of that. I don't know, the result doesn't give me confidence.
Sincerely,
Cac
Luck is what happens when preparation meets opportunity.
Posting Permissions
-*-
Reply With Quote
Bookmarks