I estimate up to a 1/4deck precision near the end of the shoe where it really matters.
An RC of +6 is TC +6 if you estimate remaining cards as 1 deck but only 4.8 if you estimate them at 1.25 decks. However RC+21 with 5.25 decks remaining is TC+4 whereas with 5 decks its TC+4.2 - not much difference.
Francis Salmon
> First you have a team
> elaborate the index with utmost precision
> and then you make whole numbers out of
> it. This produces an inherent error of up to
> 0.5 because the indices won't please you and
> sit right on the whole numbers.
I think this is a misunderstanding. When a whole number index is generated - it is not generated as a precise index first and then rounded. It is generated as a whole number POOLING all the relevant data DURING the generation. The two are not equivalent and if it was done the way you said it would be it wouldn't make sense as you suggest. If the indices are generated the way they are used - then there is no "error."
> I think this is a misunderstanding. When a
> whole number index is generated - it is not
> generated as a precise index first and then
> rounded. It is generated as a whole number
> POOLING all the relevant data DURING the
> generation. The two are not equivalent and
> if it was done the way you said it would be
> it wouldn't make sense as you suggest. If
> the indices are generated the way they are
> used - then there is no "error."
Yours is an important comment for the better understanding of index generation. I think most readers don't know much about the concept of TC "binning," for index generation. As you mention, there is never an attempt to generate a precise (one decimal place) index and then to round, floor, or truncate it later. Rather, the entire interval is used, and it is noted where the change in strategy first becomes favorable, over the entire interval, rather than attempting to specify a precise point in that interval.
Don
> I have a feeling we're turning in circles so
> this will be my last post on the subject.I
> don't think the word "silly" is a
> fair qualification of what I'm doing. You
> cannot praise Cacarulo to heaven and then
> ignore what he said:"very accurate in
> most cases".One of the rare exceptions
> and the only significant one was the index
> for the play 15v10 which showed an error of
> 0.4.
> You accuse me of simplifying and then
> complicating but you're doing just the same
> in opposit direction. First you have a team
> elaborate the index with utmost precision
> and then you make whole numbers out of
> it.This produces an inherent error of up to
> 0.5 because the indices won't please you and
> sit right on the whole numbers.
> I don't say that this amounts to very much
> and it's probably wise not to scare off
> beginners with decimal. But don't ever say
> my indices are less accurate than yours.
> This is simply not true.
Francis,
The problem is that you don't know apriori which of these indices are wrong. I put one example (15vT) but maybe there are others. You can believe that all of them are correct but that simply is not true. So, what could you do in order to confirm the accuracy of your indices? The only and more accurate way is by simulation. Now, if you have to go by this method first, why would you need to use the "representative deck" methodology? See what the point is?
OTOH, you can't use the R-D methodology to generate RA indices but maybe you think these indices are a waste of time since you haven't made a single comment on what I've said about them.
BTW, did you read BJA2 or BJA3?
Sincerely,
Cacarulo
Actually, I wanted to end this thread but you asked me some questions, so I'll answer you.
> Francis,
> The problem is that you don't know apriori
> which of these indices are wrong. I put one
> example (15vT) but maybe there are others.
From your scientific post I got the impression that you had checked them all and you wouldn't have failed to point out that there was another outlayer.
> You can believe that all of them are correct
> but that simply is not true. So, what could
> you do in order to confirm the accuracy of
> your indices? The only and more accurate way
> is by simulation. Now, if you have to go by
> this method first, why would you need to use
> the "representative deck"
> methodology? See what the point is?
I remember you saying that the typical deck method was your first approach and that it was very accurate in most cases. Here you are contradicting yourself.
I never believed that my indices were 100% correct. What I still do believe is that even if some indices might be wrong at the decimal position they are still more accurate than just whole numbers.Checking them out by myself would be far too laborious when you think that it will bring me only one extra hamburger per year to speak in Don's words.
> OTOH, you can't use the R-D methodology to
> generate RA indices but maybe you think
> these indices are a waste of time since you
> haven't made a single comment on what I've
> said about them.
RA indices are certainly very useful for people with limited bankrolls. However this is not my case and I don't see how I could integrate this concept in my betting scheme.
> BTW, did you read BJA2 or BJA3?
Only BJA1. From discussions on this and other sites, I have a pretty good idea about the content of the new editions.May be I miss some interesting tables. The book focuses on American conditions. In Europe, things are somewhat different.
Francis Salmon
> From your scientific post I got the
> impression that you had checked them all and
> you wouldn't have failed to point out that
> there was another outlayer.
Yes, I did that a long time ago and found some more. Don't remember exactly what were those discrepancies but you can use a good simulator and verify your indices by yourself. Maybe you don't mind playing with inaccurate indices but if I were you I would check them all.
> I remember you saying that the typical deck
> method was your first approach and that it
> was very accurate in most cases. Here you
> are contradicting yourself.
No, I'm not contradicting myself. I said most cases not ALL cases.
> I never believed that my indices were 100%
> correct. What I still do believe is that
> even if some indices might be wrong at the
> decimal position they are still more
> accurate than just whole numbers.
That depends on the accuracy of your TC calculation so it's arguable.
> Checking
> them out by myself would be far too
> laborious when you think that it will bring
> me only one extra hamburger per year to
> speak in Don's words.
The point is that you'll get that extra hamburger only if you use the precise indices.
> RA indices are certainly very useful for
> people with limited bankrolls.
This is not correct but I won't discuss RA indices in a public forum.
> However this
> is not my case and I don't see how I could
> integrate this concept in my betting scheme.
Well, one way would be to calculate the variance of those representative decks. Another way is to use a good simulator. Question: Which software do you have?
> Only BJA1. From discussions on this and
> other sites, I have a pretty good idea about
> the content of the new editions.May be I
> miss some interesting tables. The book
> focuses on American conditions. In Europe,
> things are somewhat different.
You're missing a lot more that some interesting tables. Now, if bankroll is not your problem you could invest some few bucks and buy the new edition.
Sincerely,
Cacarulo
>> RA indices are certainly very useful for
>> people with limited bankrolls.
> This is not correct but I won't discuss RA
> indices in a public forum.
I'm confused, if you could clarify. You had mentioned in a previous post that "...RA-indices get almost 3% over EM-indices." That would be true regardless of the size of the bankroll, right?
If it is, shouldn't that statement be the end of the debate, hands-down?
(I see you said you weren't going to discuss RA indices, but I figured I'd bring up something that you already said).
> I'm confused, if you could clarify. You had
> mentioned in a previous post that
> "...RA-indices get almost 3% over
> EM-indices." That would be true
> regardless of the size of the bankroll,
> right?
Not exactly. Suppose your bankroll is $1 million and the table max is, say, $10,000. As your max bet is 1% of your bank, r-a indices aren't going to do you any good. But, of course, this is an extreme situation. For most people, whose bankrolls are not ridiculously larger than the max bets they make, r-a indices will add value. They would surely add value to Francis's play, as well, but he just doesn't know it! :-)
Don
Of course if you have an infinite bankroll, all points are moot since you can neither increase nor decrease infinity
Sorry for the delay in the response.
> I'm confused, if you could clarify. You had
> mentioned in a previous post that
> "...RA-indices get almost 3% over
> EM-indices." That would be true
> regardless of the size of the bankroll,
> right?
That's correct if you can always bet optimally and where spread is not an issue. If you can bet whatever you want then you will have that 3% with any bankroll. The exception would be an infinite bankroll which is something unreal.
> If it is, shouldn't that statement be the
> end of the debate, hands-down?
Yes, that would be the end of the debate but the problem is that I don't know whether Salmon bets optimally or not and also don't know what his betting spread is.
Sincerely,
Cac
> That's correct if you can always bet
> optimally and where spread is not an issue.
> If you can bet whatever you want then you
> will have that 3% with any bankroll. The
> exception would be an infinite bankroll
> which is something unreal.
Does this mean that that (small) 3% gain would only show up if you made fully kelly wagers every time? How could anyone do that, though?
> Does this mean that that (small) 3% gain
> would only show up if you made fully kelly
> wagers every time? How could anyone do that,
> though?
It's not that difficult and don't need to be perfect bets. CVCX can help you on that task.
Sincerely,
Cac
Posting Permissions
-*-
Bookmarks