Dog Hand: Don's Experiment (from below and WAY to the right!)

[Dog Hand]

Don,

Using CVData's Multitracking feature, I ran a quarter-billion hand sim for 6D, S17, DA2, DAS with 5/6 pen. Three players were involved: Mr. B, who played Complete Basic Strategy; Mr. A, who played Altered Complete Basic Strategy (he always hit 12 vs. 4); and Mr. W, who wonged in whenever the TC >= +1, and wonged back out whenever the TC < +1, and so as a consequence played only 30.4% of the hands.

I checked the advantage for Mr. B and Mr. A for the two-card hand 12 vs. 4: for Mr. B, this was -21.20%; for Mr. A, it was -21.22%. Thus, when dealt 12 vs. 4, hitting or standing was a coin flip for the B.S players.

However, I should point out that this standing EV, -21.20%, is lower than the off-the-top EV for a straight B.S. player, which (according to Eric Farmer's program... I confess... I left the house this morning without my pocket edition of BJA3!) is -21.1115% for 10,2 vs. 4; -20.8039% for 9,3 vs. 4; -20.6788% for 8,4 vs. 4; and -20.2485% for 7,5 vs. 4. I took a weighted average of these, 4*(-21.1115)-20.8039-20.6788-20.2485)/7 = -20.8825%, for the OTT EV for standing on 12 vs. 4. Thus, we see that the presence of Mr. W has LOWERED the EV of standing on 12 vs. 4 for the Mr. B from -20.88% to -21.20%.

Furthermore, according to CVData's output, Mr. B and Mr. A each faced this situation on 0.646% of the hands each played. A rough calculation (ignoring the change in probability due to the dealer's upcard of 4) shows that, in the absence of Mr. W, Mr. B (and Mr. A) should face this situation 100%*2*(4/169+1/169+1/169+1/169)/13= 0.63723...%. Thus, we see that, due to Mr. W, Mr. B is faced with this crappy hand more frequently than "normal".

To summarize: the presence of the wonger does have an effect on the B.S. EV of the B.S. players, but in this situation not enough of an effect to make hitting 12 vs. 4 better than standing.

I ran just this one strategy variation because, as far as I can recall, it is the "closest call" for the 6D, S17, DA2, DAS game. If I've missed another one, just let me know!

Dog Hand

[Don Schlesinger]

> I ran just this one strategy variation because, as far
> as I can recall, it is the "closest call"
> for the 6D, S17, DA2, DAS game. If I've missed another
> one, just let me know!

Nice work, but it's not the closest call. A,2 v. 5 is. In any event, the other problem is having three players at the table, max, instead of two, which was my experiment.

The negative effect of the Wonger is strongest when he Wongs in on a lone player. With two players at the table, before he arrives, the effect is diluted.

If you try it again, my way, I think that maybe 12 v. 4 will flip-flop, and I'm certain it will be right to hit A,2 v. 5. I also think 15 v. 10 will be hit, rather than surrender.

Don

Don

[Dog Hand]

> Nice work, but it's not the closest call. A,2 v. 5 is.
> In any event, the other problem is having three
> players at the table, max, instead of two, which was
> my experiment.

> The negative effect of the Wonger is strongest when he
> Wongs in on a lone player. With two players at the
> table, before he arrives, the effect is diluted.

> If you try it again, my way, I think that maybe 12 v.
> 4 will flip-flop, and I'm certain it will be right to
> hit A,2 v. 5. I also think 15 v. 10 will be hit,
> rather than surrender.

> Don

Don,

Ok, I ran two separate CVData quarter-billion hand sims for a 6D, S17, DA2, DAS, 5/6 pen game. In each a B.S. player plays all, and a wonger plays all TC>=+1 (HiLo).

In sim #1, the B.S. player plays Complete Basic Strategy. In sim #2, the B.S. player plays a Complete B.S. slightly altered by having him hit (not double) A-2 vs. 5, and hit (not stand) 13 vs. 2, 12 vs. 4, and 12 vs. 6.

Here are the results for the two B.S. players for these hands:

Percent Advantage
...Hand ...B.S. Altered B.S.
A2 vs. 5 +6.72 +13.56
13 vs. 2 -29.53 -30.64
12 vs. 4 -21.59 -21.12
12 vs. 6 -15.79 -16.96

Thus, the Altered B.S. outperforms normal B.S. for A2 vs. 5 (since 2*6.72% = 13.44% < 13.56%) and for 12 vs. 4, but not for 13 vs. 2 nor for 12 vs. 6.

I didn't run with LS, so I didn't check whether 15 vs. 10 becomes hit.

In these sims, the wonger now plays only 28.2%. Also, the B.S. player is dealt 12 vs. 4 on 0.649% of his hands (0.648% for the A.B.S. player), which is significantly worse that the "expected" 0.637% (from my earlier post).

Dog Hand

[HALVESX2]

as the sim results has shown, wongers does in fact affect the standard BS player by increasing the BS player's percentage of hands played in negative count.
but only enough to effect those very marginal BS decisions.
but wat if the wonger comes in n play 2 hands in positive counts??
would that further affect the BS player??
is it then enough to further alter BS decisions other than 12vs4, A2vs5, and hitting 15vs10???

[Don Schlesinger]

> Don,

> Ok, I ran two separate CVData quarter-billion hand
> sims for a 6D, S17, DA2, DAS, 5/6 pen game. In each a
> B.S. player plays all, and a wonger plays all
> TC>=+1 (HiLo).

> In sim #1, the B.S. player plays Complete Basic
> Strategy. In sim #2, the B.S. player plays a Complete
> B.S. slightly altered by having him hit (not double)
> A-2 vs. 5, and hit (not stand) 13 vs. 2, 12 vs. 4, and
> 12 vs. 6.

Good work. But, I understand that, for 13 v .2 and 12 v. 6, where the index is -1 and not 0, they may not work.

> Here are the results for the two B.S. players for
> these hands:

> Percent Advantage
> ...Hand ...B.S. Altered B.S.
> A2 vs. 5 +6.72 +13.56
> 13 vs. 2 -29.53 -30.64
> 12 vs. 4 -21.59 -21.12
> 12 vs. 6 -15.79 -16.96

> Thus, the Altered B.S. outperforms normal B.S. for A2
> vs. 5 (since 2*6.72% = 13.44%)and for 12 vs. 4, but not for 13 vs. 2 nor for 12 vs. 6.

I'm not surprised.

>I didn't run with LS,
> so I didn't check whether 15 vs. 10 becomes hit.

It will; I'm certain.

> In these sims, the wonger now plays only 28.2%. Also,
> the B.S. player is dealt 12 vs. 4 on 0.649% of his
> hands (0.648% for the A.B.S. player), which is
> significantly worse that the "expected"
> 0.637% (from my earlier post).

Right. Nice job! Results as expected ... finally! :-)

Don

[Don Schlesinger]

> as the sim results has shown, wongers does in fact
> affect the standard BS player by increasing the BS
> player's percentage of hands played in negative count.

It couldn't have been otherwise. :-)

> but only enough to affect those very marginal BS
> decisions.

Right. More an academic exercise than anything else.

> but wat if the wonger comes in n play 2 hands in
> positive counts??
> would that further affect the BS player??

See my above post. We might just influence a couple of the -1 plays along with the 0 ones.

> is it then enough to further alter BS decisions other
> than 12vs4, A2vs5, and hitting 15vs10???

Maybe (he said, straddling the fence! :-)).

Don

[Don Schlesinger]

Two more nominations: A,4 v. 4 (a likely change, even when the Wonger plays one hand), and 3,3 v. 2 (need the Wonger to play two hands, and still not so sure).

Don

[Cacarulo]

> Don,

> Ok, I ran two separate CVData quarter-billion hand
> sims for a 6D, S17, DA2, DAS, 5/6 pen game. In each a
> B.S. player plays all, and a wonger plays all
> TC>=+1 (HiLo).

> In sim #1, the B.S. player plays Complete Basic
> Strategy. In sim #2, the B.S. player plays a Complete
> B.S. slightly altered by having him hit (not double)
> A-2 vs. 5, and hit (not stand) 13 vs. 2, 12 vs. 4, and
> 12 vs. 6.

> Here are the results for the two B.S. players for
> these hands:

> Percent Advantage
> ...Hand ...B.S. Altered B.S.
> A2 vs. 5 +6.72 +13.56
> 13 vs. 2 -29.53 -30.64
> 12 vs. 4 -21.59 -21.12
> 12 vs. 6 -15.79 -16.96

> Thus, the Altered B.S. outperforms normal B.S. for A2
> vs. 5 (since 2*6.72% = 13.44% I didn't run with LS,
> so I didn't check whether 15 vs. 10 becomes hit.

> In these sims, the wonger now plays only 28.2%. Also,
> the B.S. player is dealt 12 vs. 4 on 0.649% of his
> hands (0.648% for the A.B.S. player), which is
> significantly worse that the "expected"
> 0.637% (from my earlier post).

But there is something you are not posting: What is the OVERALL EV for each player (not the wonger)?

Sincerely,
Cac

[OldCootFromVA]

Did you separate out the hands?

One of the "fine points" of basic strategy is to hit X2 vs 4, but stand all other 12's vs 4.

[kc]

> But there is something you are not posting: What is
> the OVERALL EV for each player (not the wonger)?

How about considering the cases where the wonger never enters the game? In these rounds, an argument can be made that more high cards would appear since it is the appearance of high cards that keeps the wonger out of the game. The posted sims indicate that this occurs about 70% of the time for a wong-in TC of +1.

In my estimation, the best OVERALL basic strategy would still be the basic we have all come to know and love, wonger or not.

kc

[Magician]

Well, I certainly didn't expect such a theoretical question to generate so many posts, or so much controversy.

I think Dog Hand's sim methodology is sound, although I'd like to see a lot more than a 250 million rounds when we're analysing just a few types of hands. The problem with this approach is you need to guess the answer first and then check to see if you're right. We don't all have Don's decades of experience at blackjack.

I thought I'd outline a combined simulation/combinatorial analysis approach that should give you the optimum strategy for a given situation without having to guess it first.

1. Simulate a basic strategist (player A) and a wonger (player B) for your chosen set of conditions, keeping track of all the cards that are dealt to B.
2. Removing the cards that B uses, calculate the optimal strategy based on the "shorted" deck that A is playing.
3. Run the simulation again with A using the new strategy and again keep track of all the cards dealt to B.
4. Removing those cards, calculate the optimal strategy again. If it has changed, go back to the previous step.

When the iteration stops, you should have the optimal basic strategy for the conditions. Now, I don't have any software that could do this. Does anyone think they can?

[Cacarulo]

> How about considering the cases where the wonger never
> enters the game? In these rounds, an argument can be
> made that more high cards would appear since it is
> the appearance of high cards that keeps the wonger out
> of the game. The posted sims indicate that this occurs
> about 70% of the time for a wong-in TC of +1.

> In my estimation, the best OVERALL basic strategy
> would still be the basic we have all come to know and
> love, wonger or not.

That's what I'm saying and have proven through several sims

Sincerely,
Cac

[Dog Hand]

> But there is something you are not posting: What is
> the OVERALL EV for each player (not the wonger)?

> Sincerely,
> Cac
Cac,

For Altered B.S., TBA = -0.540%, IBA = -0.610%
For normal B.S., TBA = -0.520%, IBA = -0.589%

The only reason I focused on the "altered" hands is that I expect ABS and nBS will have the same EV for all other hands.

Wonger (playing with ABS guy): TBA = 1.290%, IBA = 1.521%
Wonger (playing with nBS guy): TBA = 1.272%, IBA = 1.501%

By the way, the Wonger was using a simple 1:4 spread that I found in the software. I included his results because I would expect, in the long run, that his results should be nearly the same for playing with either player. I say "nearly", because the ABS guy's strategy tends to eat more cards. Thus, if any "true" difference exists for the wonger, I'd expect he would have a HIGHER advantage when playing with the nBS guy than when playing with the ABS guy. The results show just the opposite: perhaps as Don suggested I need to sim more than a quarter-billion rounds.

Dog Hand