A Gent: Rain Man?

[A Gent]

1) What exactly is the increase in EV to the player in a single-deck downtown game for using only Hi-Lo and a 1-6 spread vs. flat betting and playing BS?

2) What exactly is the additional advantage Rain Man can extract in that same game with the same 1-6 spread if he always knows the exact composition of the remaining deck and can instantly calculate the EV for each playing decision not only on the first two cards but after subsequent hits?

Thanks!

[Dancer]

> 1) What exactly is the increase in EV to the
> player in a single-deck downtown game for
> using only Hi-Lo and a 1-6 spread vs. flat
> betting and playing BS?

The EV for a flat betting BS player is pretty straight forward, but you're not providing nearly enough info to calculate the EV for a Hi-Lo player. What's the penetration level, how many players at the table, what's your bet ramp?

> 2) What exactly is the additional advantage
> Rain Man can extract in that same game with
> the same 1-6 spread if he always knows the
> exact composition of the remaining deck and
> can instantly calculate the EV for each
> playing decision not only on the first two
> cards but after subsequent hits?

Same problem as above. You might want to refer to the "Theory of Blackjack" by Peter Griffin about the advantage gained from computer-perfect play, but you'll be disappointed how little there is to be gained.

> Thanks!

[Bearer]

Have actually been reading ToB.

Let me ask a slightly different question:

Is saying that a count system has a .66 playing efficiency exactly the same as saying that the system captures 66% of the effect of computer-perfect play?

If not, please clarify. If so, then does it follow that computer-perfect play should yield 50% more gain than this count system?

To provide the other info you requested, let's assume 50% pen in a heads-up game and, for simplicity, 1 unit on TC of zero or less and 6 on all positive counts.

Thanks.

[ET Fan]

Is saying that a count system has a .66 playing efficiency exactly the same as saying that the system captures 66% of the effect of computer-perfect play?

66% of the strategy portion effect on your bottom line, yes.

If not, please clarify. If so, then does it follow that computer-perfect play should yield 50% more gain than this count system?

Except that most of your yeild comes from betting -- not strategy. For example, suppose the ratio in some game is 4 to 1. That means 80% of your profits come from betting, 20% from strategy departures. The 50% increase would cause a 10% overall gain to your bottom line. So if you get 1% with the count system, you'd be at 1.1% with a computer.

ETF

[Dancer]

> Have actually been reading ToB.

> Let me ask a slightly different question:

> Is saying that a count system has a .66
> playing efficiency exactly the same as
> saying that the system captures 66% of the
> effect of computer-perfect play?

> If not, please clarify. If so, then does it
> follow that computer-perfect play should
> yield 50% more gain than this count system?

> To provide the other info you requested,
> let's assume 50% pen in a heads-up game and,
> for simplicity, 1 unit on TC of zero or less
> and 6 on all positive counts.

> Thanks.

ET Fan is right. PE and IC contribute far less to your overall winrate than BE -- especially as the number of decks increases. If I remember correctly, ToB talks about that as well.

As for the second question, here are the numbers I get for a single deck, H17 game, $5 units, 100 hands/hr, 250 million hands:

BS EV: -$0.82/hr
Hi-Lo EV (per above): $17.77/hr

If only we could get away with betting like that...

[Bearer]

Thanks for the responses.

OK, so for Hi-Lo PE of .51, computer-perfect play should be almost twice as profitable in terms only of PE comparison.

But as you've noted, BC dominates PE, and given Hi-Lo's .97 BC, computer-perfect play yields little.

My question then is: How/where do I determine, for a given game, the relative importance of BC to PE? In general I'd assume that the more decks (and also the greater the betting spread), the more BC matters. How big is the difference in effect as you move from game to game? Are there any rule variations that have a significant effect on the relative importance of PE and BC?

And what about IC? Is this not a significant component of profitability, especially for shoe games?

Again, thanks.

[Dancer]

> OK, so for Hi-Lo PE of .51, computer-perfect
> play should be almost twice as profitable in
> terms only of PE comparison.

Yes. The best single-parameter systems approach .70.

> But as you've noted, BC dominates PE, and
> given Hi-Lo's .97 BC, computer-perfect play
> yields little.

That's true, although PE and IC do play a more significant role in single deck than in shoe games.

> My question then is: How/where do I
> determine, for a given game, the relative
> importance of BC to PE? In general I'd
> assume that the more decks (and also the
> greater the betting spread), the more BC
> matters. How big is the difference in effect
> as you move from game to game? Are there any
> rule variations that have a significant
> effect on the relative importance of PE and
> BC?

That's a great question. Perhaps Don or someone more qualified than me can take a shot at it. For my purposes, I just use simulations. They don't give me the relative importance of each component, but I do get the overall answer. With CVData, I can run a 250 million hand simulation in about a minute, so I can test all sorts of combinations with very little effort.

> And what about IC? Is this not a significant
> component of profitability, especially for
> shoe games?

IC is important, but you've got it backwards with respect to shoe games. Insurance is far more profitable in single deck than in shoe games. Hi-Lo has an IC of around 72%, so computer-perfect play would help significantly here.

Several months ago, I posted an insurance study where I compared perfect insurance against the system I play and threw in Hi-Lo for comparison. I didn't run it for single deck (since I rarely play it), but I did run it for 2 deck and 6 deck. Here are the partial results:

Perfect Insurance (2D, 75% pen, avg insurance bet = 1.5 units)
Frequency: 1.915%
Advantage: 8.609%
Units/Hour: (1.915% * 8.609% * 1.5 X 100) = 0.247

Hi-Lo (TC = +2, 2D, 75% pen, avg insurance bet = 1.5 units)
Frequency: 1.592%
Advantage: 6.436%
Units/Hour: (1.592% * 6.436% * 1.5 X 100) = 0.154

-----

Perfect Insurance (6D, 5/6, avg insurance bet = 2.5 units)
Frequency: 1.123%
Advantage: 6.232%
Units/Hour: (1.123% * 6.232% * 2.5 X 100) = 0.175

Hi-Lo (TC = +3, 6D, 5/6, avg insurance bet = 2.5 units)
Frequency: 0.744%
Advantage: 4.973%
Units/Hour: (0.744% * 4.973% * 2.5 X 100) = 0.092

As you can see, the frequency of insurance bets and the advantage from making them both drop significantly as more decks are added.