Gramazeka: Split 2,2 and 3,3 vs 8

[Gramazeka]

Hello colleagues. I live in Russia. Has noticed, that some indexes do not coincide in various programs. For example under Korovin's program (the Russian program) Split 2,2 against 8 we do at 3,5 count(hi lo). And at Wong with 5. Accordingly split 3 vs 8 with 5 (on Korovin), and at Wong with 4. Why there are such discrepancies? Share ideas?...Rules 6 deks and DAS.ENHC

cgm.ru

[Norm Wattenberger]

I don't know what Korovin's program is; but it's wrong.

> Hello colleagues. I live in Russia. Has noticed, that
> some indexes do not coincide in various programs. For
> example under Korovin's program (the Russian program)
> Split 2,2 against 8 we do at 3,5 count(hi lo). And at
> Wong with 5. Accordingly split 3 vs 8 with 5 (on
> Korovin), and at Wong with 4. Why there are such
> discrepancies? Share ideas?...Rules 6 deks and
> DAS.ENHC

> cgm.ru

[Francis Salmon]

It's common knowledge that the two is the stronger card than the three.Therefore the index for 3/3 v 8 must be higher than the one for 2/2 v 8 and Wong for once is wrong. My CA-program gives an index of 3.5 for 2/2 v 8 and 5.1 for 3/3v8. So this Korovin-program seems quite trustworthy.

Francis Salmon

[cooper(jr)]

> I don't know what Korovin's program is; but it's
> wrong.

why?

[GeoC]

is TC +3 for HiLo assumes shoe game with DDAS. This number comes from David Smith's PBA.

[David Spence]

> It's common knowledge that the two is the stronger
> card than the three.

You're absolutely correct--it's better for a player to have a hand of 2 than 3.

> Therefore the index for 3/3 v 8
> must be higher than the one for 2/2 v 8

Not necessarily. Though the end result of splitting 2-2 (having two hands of 2) is better than the end result of splitting 3-3 (having two hands of 3), the GAIN from splitting 3-3 might be greater. This is because an initial hand of 6 is worse than an initial hand of 4.

Using Casino V?rit?, I get an index of 5 for splitting 2-2 vs. 8 and an index of 5 for splitting 3-3 vs. 8.

The main point, however, is that splitting 2-2 does NOT produce more of a gain than splitting 3-3, since 2-2 is a better starting hand than 3-3. Splitting 3-3 allows you to "escape" the horrible starting hand of 6. Though a starting hand of 4 is not great either, it's not quite as bad as 6.

David Spence

[Norm Wattenberger]

> You're absolutely correct--it's better for a player to
> have a hand of 2 than 3.

> Not necessarily. Though the end result of splitting
> 2-2 (having two hands of 2) is better than the end
> result of splitting 3-3 (having two hands of 3), the
> GAIN from splitting 3-3 might be greater. This is
> because an initial hand of 6 is worse than an initial
> hand of 4.

Quite right. Whether the Split of 2's or 3's is at the higher index varies with both the dealer upcard and the strategy. This can be seen by looking at different counts.

> Using Casino V?rit?, I get an index of 5 for splitting
> 2-2 vs. 8 and an index of 5 for splitting 3-3 vs. 8.

Depends on the exact settings. I get +5 and +5 for rounding and +5 and +4 for flooring.

> The main point, however, is that splitting 2-2 does
> NOT produce more of a gain than splitting 3-3, since
> 2-2 is a better starting hand than 3-3. Splitting 3-3
> allows you to "escape" the horrible starting
> hand of 6. Though a starting hand of 4 is not great
> either, it's not quite as bad as 6.

Yes, defensive splits are odd. It is very difficult to come up with any 'rules of thumb' about defensive splits.

[Norm Wattenberger]

Interesting. This means BCA may disagree with PBA - which was, at least originally, based on BCA. In any case, I get the same numbers as Wong's original numbers. These are unusual indexes - not much about them in the literature. They aren't even in Braun's Hi-Opt II tables.

> is TC +3 for HiLo assumes shoe game with DDAS. This
> number comes from David Smith's PBA.

[Don Schlesinger]

> Quite right. Whether the Split of 2's or 3's is at the
> higher index varies with both the dealer upcard and
> the strategy. This can be seen by looking at different
> counts.

> Depends on the exact settings. I get +5 and +5 for
> rounding and +5 and +4 for flooring.

> Yes, defensive splits are odd. It is very difficult to
> come up with any 'rules of thumb' about defensive
> splits.

No one is discussing the correlation of these plays. 3,3 v. 8 is a very odd index play. Naturally, the most important card to receive on a split 3 is an 8, and Hi-Lo (and virtually all counts) is incapable of telling us when we have an excess or dearth of 8s, since the 8 is not counted in all systems.

It isn't sufficient to simply generate an index number to determine the value of a departure; in addition, you have to determine whether the counted is suited to make the play in the first place.

I can't prove it vigorously, but I would argue strongly that "knowing" an index for splitting 3,3, v. 8 in Hi-Lo, or any other count, is next to meaningless and worthless.

Note, also, that, although Wong, strangely, furnishes an index ofr this play for Hi-Lo, he does no such thing for Halves, the "superior" count. What sense does that make?

Don

[Titan5]

I believe Split 2,2 and 3,3 vs 8 are unique in that Hi-Lo indices are relatively weak and not very useful. If you use 678 side count, then this would be much more important than HiLo indices. If the remaining deck is full of 7 or 8, the split hands will become very strong hands.
Hi, Francis. Nice to see your post again.
Titan5 in LA.

[Titan5]

I meant 789 side count.

[Francis Salmon]

It's true that 2/2v8 is the better starting hand than 3/3v8 .EV-wise, this is in a neutral deck -15.9% for 2/2v8 and -21.7% for 3/3v8. Splitting in a neutral deck is a slightly bigger error for 2/2v8 (-1.5%) than for 3/3v8 (-1.3%.
But what matters more is the gain achieved by TC-unit and here comes in the power of the two since splitting 2/2 catches up about 0.45% per TC and breaks even with hitting at TC 3.5. Splitting 3/3 catches up only about 0.25% per TC and breaks even with splitting at TC 5.1.
I quite agree with Don that these two indices are practically worthless but if I had to choose between the two I would certainly pick 2/2v8.

Francis Salmon

[Norm Wattenberger]

I had briefly consider running sims to see the difference in the indexes. But it would take tens of billions of hands, the difference is so tiny.