"That is not what happened at all. I made the books available, and canceled them when only a few people ordered."
I say, when the legend becomes fact, print the legend.
"That is not what happened at all. I made the books available, and canceled them when only a few people ordered."
I say, when the legend becomes fact, print the legend.
Hey Myooligan, I have something I think is quite interesting that you may not have thought about before and might get to work on. And if you have then please help me out on this one, I have been both confounded and horrified by my findings, See below.
> I think it's great that you're asking some of these
> questions. You know how I can tell if someone I meet
> wouldn't be a good blackjack player? They believe me
> when I tell them they can beat the casino. Only a
> sucker would believe something like that. A successful
> player figures out what kind of advantage he has for
> himself, rather than putting faith in convincing
> testimonies from the internet. Personally, I only
> became comfortable putting real money into blackjack
> once I started experimenting with simsimp (a
> bare-bones freeware blackjack simulator). It allowed
> me to confirm some of the key elements of
> "blackjack theory" for myself.
> Nevertheless, here's a simple explanation to why the
> "crudeness" of card counting doesn't make it
> invalid: For any given index, the majority of the time
> the TC will be other than that index. And for all
> index numbers greater than 1 or less than -1, the
> majority of the time the TC will be other than (index
> +- 2). (This is true for level two counts like the one
> used in the study). In general, the evs for the two
> play decisions, near the index number , are fairly
> close. So, less than half the time you encounter a
> given decision, you'll be playing in the (index +- 2)
> zone, where the ev differences are insignificant. No
> more than half of that time, you'll make the
> incorrect decision and sustain a negligible loss.
> Which is why it doesn't pay a whole lot to be precise
> when you count cards.
> But if you compare the two play decisions at a TC far
> away from the index number , there will generally be a
> significant difference in their EVs. So, more than
> half the time you encounter a given decision, you'll
> be playing outside the (index +- 2) zone, where the ev
> differences are significant. In 100% of those
> instances, you'll make the correct decision, and avoid
> a significant loss. Which is why it pays to count
> cards.
> Hope that helped.
Yes, after reading over the post for the fifth or sixth time now, I do think that I have a good and firm understanding of what you are saying about the value of index accuracy. I will not try to recreate your explanation here as I am sure that you understand it...You wrote it, but if you get this far down the page to read this response, then I think I have another, even more perplexing question about index values and strategy deviation that you will find quite interesting and hopefully not overly terrifying like I do.
I wonder about a negative efficiency for counting systems(all counting systems), meaning that counting cards can actually make your playing decisions worse than if you were just using Basic Strategy.
After reading Griffin's TOB, particularly Chapter 6 "Tables and Applications" and seeing how playing strategy calulations are made using the "virtually complete strategy tables" containing the effects of individual card removal for each play/decision of a hand versus a dealer's upcard, I have toyed around with the idea of creating an "absolute" counting system which uses the exact value of the card for determining playing decisions. As I have looked closely at these values found in the table, it became apparent that the individual card denomination takes on significantly different values for each specfic play(a player's hand vs. dealer's upcard). The problem with "traditional" counting systems is the fixed value(particularly the sign) associated with a card denomination. For example,(one that is pointed out clearly in the book) on the play of hard 14 vs. dealer 10. The seven has an "absolute" value of -4.21 and has an effect of -4.21% for the removal of each individual seven, making it by far the single most significant, valuable and important card for this play. In a standard 52 card deck, the advantage of drawing(over standing) to hard 14 vs. dealer 10 is 6.64%. If the deck is depleted by one seven, then the advantage becomes 2.43%(6.64%-4.21%). If the deck is depleted by one more seven, then the advantage becomes -1.78%(2.43%-4.21%) a negative, which is why it is better to stand on hard 14(7-7) vs. dealer 10 rather that hit in single deck, as I am sure you know.
Now if we were playing the Zen count then the seven would count as +1 unvaryingly, clearly far from the "correct" value of -4.21%, and it would be this information that we would use to make this play of (hard)14 vs. 10. It would seem to me that this is not very "good and reliable" information. And to make things worse, there are numerous many individual plays where this is the case, where the counting system "misvalues" the card as it is removed from the deck.
If my understanding is correct then counting cards could actually have a negative impact and a possibly detrimental effect on playing, making it again quite crude at best, if you don't mind my saying. :-)
Sorry I am late, but I have been busy and wish to be thorough as I am the Designated Driver after all.
Thanks,
Desi. D.
P.S. Maybe I am wrong, but it seems to me that a system's count values(tags or weights) are chosen based more on BC than on PE, and that the two are usually somewhat "contradictory" and "incompatible" metrics where a compromise needs to be made between the two when used in actual play.
Just want to say thanks Don. I just receieved my copy of BJA3 and it looks to make for some good reading.
And as for the posts, I will try not to turn into Francis(although I do believe he makes a good point), but then again, so do you. :-)
Desi. D.
> Depends on what you mean by "most." With the
> Catch 22 indices, in a shoe game, you get about 90% of
> all the gain by using all the indices. But, with no
> indices at all, you get probably only about 60-65% of
> all the gain.
> The point is -- whatever gain there is from indices,
> whether you use 1, 1.4, or 2, makes virtually no
> difference whatsoever.
> Yes.
> It's a free country. Don't accept it!
> Not at all.
> Sorry to disappoint you.
> Because we can. :-)
> The highest-winning teams in the history of the game
> all played Hi-Lo. Go figure. :-)
> For Francis! :-)
> Define "little." If you play for high
> stakes, then an extra, say, 10% annually can amount to
> a decent amount of money. Obviously, there is no
> penalty for playing "perfectly." So, people
> such as you will do just that, to be satisfied that
> they are doing everything they can to get all the
> advantage possible. There's nothing wrong with that.
> It only becomes silly when you delude yourself into
> thinking that indices with decimal points make a fig
> of difference, when they don't.
> That's your privilege. Why apologize?
> Enjoy it. It was written for people like you.
> Of course you can.
> Samuel Johnson's famous quote goes something like
> this: "I have furnished you a reason; I am not
> required to furnish an understanding." I did the
> best I can. The rest is up to you.
> Don
But, sims don't show a great improvement.
just kidding, dd.
Counting cards, when done correctly, will never have a negative effect on your ev. The closer the tags of your counting system to the effects listed in the Griffin tables, the more closely it will correlate with that play. Note that if the tags are close to exactly opposite of the effects (which I suppose is what you mean by "negative efficiency"), you can still make full use of the information; it's just reversed. So, above the index number for negatively correlated plays, we do worse for hitting than standing, doubling than hitting, or splitting rather than not splitting.
It's not these inversely correlated plays that are the problem, but rather, ones that do not correlate with the counting system at all. For a given play decision, certain counting systems just aren't good predictors. What we find is that you need some off the wall index number like 250 to deviate from BS on certain plays. That generally means the correlation on that play is very low.
Anyway, Playing Efficiencies are always positive, and negative hand correlations actually increase PE.
And yeah, you're right that more weight is generally given to bc. My prediction: You'll come to see that this is because it gives us the best bang for the buck, but not before you've completely exhausted yourself and thrown away your social life, hell-bent on successfully mastering the Gordon Count with 5 side counts.Been there, won't go back. . .
> Hey Myooligan, I have something I think is quite
> interesting that you may not have thought about before
> and might get to work on. And if you have then please
> help me out on this one, I have been both confounded
> and horrified by my findings, See below.
> Yes, after reading over the post for the fifth or
> sixth time now, I do think that I have a good and firm
> understanding of what you are saying about the value
> of index accuracy. I will not try to recreate your
> explanation here as I am sure that you understand
> it...You wrote it, but if you get this far down the
> page to read this response, then I think I have
> another, even more perplexing question about index
> values and strategy deviation that you will find quite
> interesting and hopefully not overly terrifying like I
> do.
> I wonder about a negative efficiency for counting
> systems(all counting systems), meaning that counting
> cards can actually make your playing decisions worse
> than if you were just using Basic Strategy.
> After reading Griffin's TOB, particularly Chapter 6
> "Tables and Applications" and seeing how
> playing strategy calulations are made using the
> "virtually complete strategy tables"
> containing the effects of individual card removal for
> each play/decision of a hand versus a dealer's upcard,
> I have toyed around with the idea of creating an
> "absolute" counting system which uses the
> exact value of the card for determining playing
> decisions. As I have looked closely at these values
> found in the table, it became apparent that the
> individual card denomination takes on significantly
> different values for each specfic play(a player's hand
> vs. dealer's upcard). The problem with
> "traditional" counting systems is the fixed
> value(particularly the sign) associated with a card
> denomination. For example,(one that is pointed out
> clearly in the book) on the play of hard 14 vs. dealer
> 10. The seven has an "absolute" value of
> -4.21 and has an effect of -4.21% for the removal of
> each individual seven, making it by far the single
> most significant, valuable and important card for this
> play. In a standard 52 card deck, the advantage of
> drawing(over standing) to hard 14 vs. dealer 10 is
> 6.64%. If the deck is depleted by one seven, then the
> advantage becomes 2.43%(6.64%-4.21%). If the deck is
> depleted by one more seven, then the advantage becomes
> -1.78%(2.43%-4.21%) a negative, which is why it is
> better to stand on hard 14(7-7) vs. dealer 10 rather
> that hit in single deck, as I am sure you know.
> Now if we were playing the Zen count then the seven
> would count as +1 unvaryingly, clearly far from the
> "correct" value of -4.21%, and it would be
> this information that we would use to make this play
> of (hard)14 vs. 10. It would seem to me that this is
> not very "good and reliable" information.
> And to make things worse, there are numerous many
> individual plays where this is the case, where the
> counting system "misvalues" the card as it
> is removed from the deck.
> If my understanding is correct then counting cards
> could actually have a negative impact and a possibly
> detrimental effect on playing, making it again quite
> crude at best, if you don't mind my saying. :-)
> Sorry I am late, but I have been busy and wish to be
> thorough as I am the Designated Driver after all.
> Thanks,
> Desi. D.
> P.S. Maybe I am wrong, but it seems to me that a
> system's count values(tags or weights) are chosen
> based more on BC than on PE, and that the two are
> usually somewhat "contradictory" and
> "incompatible" metrics where a compromise
> needs to be made between the two when used in actual
> play.
Posting Permissions
-*-
Bookmarks