In the below experiments, it is assumed that the BS player knows someone is Wonging. The question is if the BS is different with this assumption. But, if it is known then the BS player should raise his bets when the Wonger sits in. If this is the case, then I would think it more likely that the BS would change.
> In the below experiments, it is assumed that the BS
> player knows someone is Wonging. The question is if
> the BS is different with this assumption. But, if it
> is known then the BS player should raise his bets when
> the Wonger sits in. If this is the case, then I would
> think it more likely that the BS would change.
That's a separate question, and you would now be talking about using CBS at the appropriate moments. Nothing you've said above is wrong, but it isn't the original question.
Frankly, I'm disappointed by the direction that the thread has taken. The original question was, actually, quite simple, and it has a simple answer: It is unquestionably true that certain plays will change. The problem with the discussion is, rather than simply answer the question, people have gone off on all sorts of tangents and are answering questions that they find interesting, rather than answering directly the original question.
Again, I'm not certain of every play, but it seems reasonably clear to me that A,2 v. 5 will change; 12 v. 4 will change; surrendering 15 v. 10 will change; and it may the case that A,4 v. 4 will change; surrendering 16 v. 9 will change; and splitting 3,3 v. 2 might change.
Now, what would be nice is if everyone would stop conducting his own private experiments and simply demonstrate, one by one, that if each of these plays is altered, the BS player will do better than following classic BS. A couple may not work. I'm confident that most, if not all, of them will.
So, one last time: Stop reinventing the wheel with extraneous experiments, and do the one I outlined originally, please.
Don
I have been following the whole issue, and I beleive (could be wrong) that if everything is so inconclusive, it has to do with the non-linearity of index plays.
Cacarulo's numbers show that for A2 v 5, the BS player would be way better off to just hit, unfortunately, that is a very rare occurence, so the overall effect is negligeable. But when it comes to hitting stiffs against low cards, perhaps, it does more damage in high counts; enough to counter the gains at low counts even tho it happens less frequently. Like I said, this is just a guess, could be wrong. I would also be curious in finding out the average TC for the BS player.
If my reasonning is correct, there would be more soft doubles that would become hits as well as 9 v 3 and maybe 1 or 2 borderline splits would be eliminated; I would test all the plays that involve putting more money on the table that are within 3pts. of 0. But I don't beleive that the modifications would exactly match a BS adapted for the exact average TC that this player would play.
How about an RA BS?
TAO
as u would have noticed from the results only the really marginal hands would outperform NBS.
and if that holds true. then there is no way logical that it'd cost more damage in positive counts than the gain from negative count the ABS player is spending more time playing negative hands without betting up in positive hands.
if in the case where the hand does more damage then that strategy is probably not marginal enough to start off with eg 13 vs 2.
remember that the BS player is not counting, and altering hilo's 0 index plays in some way be misleading since aces r counted -1. the play table of hilo is tailored to that, eg 12vs6
however A2vs5 and 12vs4 and 15vsT r the closest calls in most counts.
closet play following would probably be A4vs4, and the effect of a wonger is probably not so to the extend of hitting 9 vs 3 instead of double.
i would suggest on picking 0 index play from non ace minus reckoning count with high PE and test out which ones the wonger would actually effect, put all those outperforming the NBS together then we can find out the total difference which ABS would make (of cos choosing only the ones that r favourable) i think that's wat Don has been trying to do right from the very start.
not only can the ABS be used in wonger filled casinos. it can also be used in smaller casinos where the dealers like to shuffle up more often when the count rises.(even though its more hypothetical than realistic)
and about wat Norm has suggested. i think if a BS player is smart enough to know that he is being wonged and bets up when a wonger comes in, he/she have read a fair bit on bj and would be smart enough to learn a few most frequent play changes made in positive counts.
i mean the BS guy might be too lazy to count and calculate, but learning to stand 16vs10, double 9vs2, standing 12 vs 2 n 3 when he follows the wonger to bet up is not too hard to handle. most other play changes will require the knowledge of the count and just using the wongers's bet size as an indicator is of no help, since the person is not able to realise the count change after the hand is dealt. and if the player was to learn to realise those changes then he/she might as well just lean the whole thing.
I'm going to bow out of this discussion with this last post, because I truly am astonished by the twists, turns, and spins that virtually everyone in the discussion wants to interject. I just can't understand why everyone needs to "overthink" the original problem.
First, no one is asking for the OVERALL BS edge, playing two different ways, because we don't even know what one of those two ways is. The original question was as clear as a bell: Can the presence of a Wonger make it such that certain BS plays CHANGE? Period!!! No one asked for the overall, cumulative edge of the aggregate of all the changes, although we can ascertain that only at the end of the discovery phase.
And, again, all I can do is offer an experiment; clearly, it has been uniformly ignored up until now, so I have little hope that it will ever be carried out "correctly," i.e., as I have suggested it be done.
Suppose that a lone BS player could play 200 hands per hour against the dealer at a 4.5/6, S17, DAS, LS game. Suppose 70% of the hands, or 140 hands, would be at true counts lower than +1. That would leave 60 hands at counts of +1 or higher. But, when +1 is attained, the Wonger enters. So, the BS player now plays only two-thirds of 60, or 40 hands, at the higher counts. (By the way, someone interjected that the Wonger leaves at -1, so as to imply that he plays through the 0 and slightly negative TCs. He doesn't!! He's a pure Wonger, playing +1 and higher or not at all.)
So, one of the questions posed above is, What is the average TC for our seated BS player? The answer is: (140 times the average TC from -99 to +1, plus 40 times the average TC from +1 to 99, divided by 180).
So, we need those missing values, above, and I don't have them. Norm probably has a chart somewhere that will provide them. But, one thing is clear -- at least to me (I've given up speaking for anyone else!) -- the final result is negative!!!
And, what does that imply? That the BS plays that are normally made at 0 or higher are probably not correct anymore. I've already suggested what those plays are. They need to be changed, separately, one by one, to see if BS improves, under the conditions outlined. My guess is that there may be as many as five or six such plays.
That's it. I'm done. I'd be interested in the results from the above scenario, and, frankly, I'm not terribly interested in anything esle. :-) But, that won't stop 10 people from proposing 10 new scenarios to explore. :-)
Don
> Frankly, I'm disappointed by the direction that the
> thread has taken. The original question was, actually,
> quite simple, and it has a simple answer: It is
> unquestionably true that certain plays will change.
> The problem with the discussion is, rather than simply
> answer the question, people have gone off on all sorts
> of tangents and are answering questions that they
> find interesting, rather than answering directly the
> original question.
Well, that's not quite true. I did run the experiments the way you asked to do it but you didn't like the answers and said they were wrong and flawed. However, I did run one last sim considering some of your new candidates.
> Again, I'm not certain of every play, but it seems
> reasonably clear to me that A,2 v. 5 will change; 12
> v. 4 will change; surrendering 15 v. 10 will change;
> and it may the case that A,4 v. 4 will change;
> surrendering 16 v. 9 will change; and splitting 3,3 v.
> 2 might change.
This time I ran 10000 million rounds per sim in order to get a lower SE.
1) NBS-PLAYER + WONGER at +1 (Hi-Lo)
Rounds = 10000000025
Shoes = 233419056
r/s = 42.841404
Ev = -0.005810
Var = 1.327425
Sd = 1.152139
Se = 0.000012
2) ABS-PLAYER + WONGER at +1 (Hi-Lo)
The Altered-BS uses the following 4 plays:
a) 12v4 ==> Hit (instead of Stand)
b) A2v5 ==> Hit (instead of Double)
c) A4v4 ==> Hit (instead of Double)
d) 33v2 ==> Hit (instead of Split)
Rounds = 10000000022
Shoes = 233207570
r/s = 42.880255
Ev = -0.005815
Var = 1.336033
Sd = 1.155869
Se = 0.000012
As you can see, there is no significative change in EV. In fact, I would say that the Normal-BS is a little bit better.
Hope this helps.
Sincerely,
Cac
1. The standard error is larger than the difference in EV, so we can't conclude anything at all.
2. You can't do all four plays at once. Three changes may be correct, and one may be so wrong so as to make the overall EV from the four worse than NBS.
3. If you don't do each play separately, no conclusion at all is possible.
4. Finally, it would be nice to add surrender 15 v. 10 and 16 v. 9, since I think each will change.
I appreciate your work, but, doing it your way, it is still impossible to decide if any plays change, and I still believe that they will.
Don
1. Are you able to tell what the global TC that the BS player faces is?
2. What percentage of total hands does the Wonger play? Table 10.51 of BJA3 would indicate 27+%, but I believe you may correspond to Table 10.44, which gives 29+%.
Don
> 1. The standard error is larger than the difference in
> EV, so we can't conclude anything at all.
Well, we can conclude that the first four decimals are correct in both cases (-0.0058). This means that there is no difference at all in using one strategy or the other. If instead the ABS were -0.0057 then I could affirm that those changes are meaningful.
> 2. You can't do all four plays at once. Three changes
> may be correct, and one may be so wrong so as to make
> the overall EV from the four worse than NBS.
I agree so I ran another sim with only two plays:
a) 12v4 ==> Hit (instead of stand)
b) A2v5 ==> Hit (instead of double)
Rounds = 10000000004
Shoes = 233458908
r/s = 42.834091
Ev = -0.005808
Var = 1.331867
Sd = 1.154065
Se = 0.000012
Once again we have -0.0058 so the changes are not a big deal for the basic strategist.
> 3. If you don't do each play separately, no conclusion
> at all is possible.
> 4. Finally, it would be nice to add surrender 15 v. 10
> and 16 v. 9, since I think each will change.
I'll owe you those since my surrender sub needs to be fixed.
> I appreciate your work, but, doing it your way, it is
> still impossible to decide if any plays change, and I
> still believe that they will.
Doing it my way can tell you whether a candidate is worth it or not. If the OEV doesn't change in the long run, why should you worry about?. Besides, I don't have other way. It must be a brute-force experiment.
I know you will probably disagree with what I'm going to say but I'm starting to believe that the "optimal strategy" is the one that appears in BJA3 (Normal BS).
Sincerely,
Cac
> 1. Are you able to tell what the global TC that the BS
> player faces is?
What do you mean by global TC? I could post the TC distribution before and after the wonger.
> 2. What percentage of total hands does the Wonger
> play? Table 10.51 of BJA3 would indicate 27+%, but I
> believe you may correspond to Table 10.44, which gives
> 29+%.
This is an interesting question for which I get 21.47% with the BS-Player AND the WONGER. Note that when the BS-player is alone he will get a 27.95% of TCs >= +1.
Sincerely,
Cac
> What do you mean by global TC? I could post the TC
> distribution before and after the wonger.
No, I want the average TC faced by the BS player, who never leaves the table, whether the Wonger is there or not.
I asked: 2. What percentage of total hands does the Wonger
> play? Table 10.51 of BJA3 would indicate 27+%, but I
> believe you may correspond to Table 10.44, which gives
> 29+%.
> This is an interesting question for which I get 21.47%
> with the BS-Player AND the WONGER.
Oh, yes, right. In my example, it would be 40/180 = 22.2%, so your number is fine.
> Note that when the
> BS-player is alone he will get 27.95% of TCs >=
> +1.
IF he were alone then; but he isn't. The Wonger enters. :-)
Don
> Well, we can conclude that the first four decimals are
> correct in both cases ( -0.0058 ). This means that
> there is no difference at all in using one strategy or
> the other. If instead the ABS were -0.0057 then I
> could affirm that those changes are meaningful.
We're not trying to find out if anything is meaningful. It won't be. In your work, you often give BS to ten decimal places, so you aren't indifferent to being precise. :-)
We're trying to find out if any changes are CORRECT. Period. To me, it's a forgone conclusion that they WON'T be meaningful and that this is purely an academic exercise.
> I agree so I ran another sim with only two plays:
Why two???? Why can't you do them one at a time????
> a) 12v4 ==> Hit (instead of stand)
> b) A2v5 ==> Hit (instead of double)
> Rounds = 10000000004
> Shoes = 233458908
> r/s = 42.834091
> Ev = -0.005808 Var = 1.331867
> Sd = 1.154065
> Se = 0.000012
> Once again we have -0.0058 so the changes are not a
> big deal for the basic strategist.
But, together, they are now an improvement, no? Again the SE is too large for any conclusions. But, again, doing them two at a time makes it impossible to tell if one is "good" and one is "bad." You do see that, don't you??
> I'll owe you those since my surrender sub needs to be
> fixed.
I'm not going anywhere.
> Doing it my way can tell you whether a candidate is
> worth it or not. If the OEV doesn't change in the long
> run, why should you worry about?
See above. I've never "worried about it." This exercise has ZERO practical use.
> Besides, I don't
> have other way. It must be a brute-force experiment.
> I know you will probably disagree with what I'm going
> to say but I'm starting to believe that the
> "optimal strategy" is the one that appears
> in BJA3 (Normal BS).
I respect your opinion. Mine is different.
It would help if we knew how negative the global TC turns out to be for the BS player.
Don
> No, I want the average TC faced by the BS player, who
> never leaves the table, whether the Wonger is there or
> not.
So, I wrote to Norm:
Norm,
Suppose a 4.5/6, DAS, S17, LS game.
Consider the expression (3.5x + y)/4.5, where x is the average (weighted for frequency) TC from
-99 to +0.999 and y is the average TC from +1.0 to +99.
Can you tell me the (slightly negative) value of
(3.5x + y)/4.5?
Thanks.
Don
Norm's response:
x=-1.449
y=2.345
(3.5x + y)/4.5 = -2.73/4.5 = -0.61.
So, I now ask one last time: Given the above, which has been clear to me since the initial post, how can we conclude anything other than that switching A,2 v. 5; 12 v. 4;
A,4 v. 4; A,7 v. 2 (a new one); 15 v. T; 16 v. 9; and 3,3 v. 2 (all zero Hi-Lo indices) is not the right thing to do???
Don
Posting Permissions
-*-
Bookmarks