MJ: CVData Output Questions

[MJ]

Three simple questions (click link below for screen shot):

Note: This is a WiWo simulation and assumes 2 hands upon entry and 1 hand when placing the waiting bets. 100 rounds/hr are set.

1) How does CVData compute 15.1% Hands played in the summary?

% Hands Played = [Hands Played/(Hands played + Hands skipped)] x 100

566 million hands/(566 million + 1.7 billion hands) x 100 ~ 25%

2) How does the software compute a DI of 6.69?

DI = [WR/Hand / SD/Hand] x 1000

[$0.3239/$124.68] x 1000 = 2.59

3) If 100 rounds are dealt per hour with 2 Billion rounds simulated, then why does the software indicate 19,999,410 hours instead of 20 million hours? I'm sure there must be a logical reason, but it eludes me.

Thanks in advance for any answers,
MJ

Screen Shot

[Norm Wattenberger]

> Three simple questions (click link below for screen
> shot):

> Note: This is a WiWo simulation and assumes 2 hands
> upon entry and 1 hand when placing the waiting bets.
> 100 rounds/hr are set.

> 1) How does CVData compute 15.1% Hands played in the
> summary?

> % Hands Played = [Hands Played/(Hands played + Hands
> skipped)] x 100

> 566 million hands/(566 million + 1.7 billion hands) x
> 100 ~ 25%

Rounds played - rounds skipped / Rounds played
2,000,000,000-1,698,000,000/2,000,000,000

You can't add 566 and 1.7 because the 566 includes two hand rounds. Also it includes splits.

> 2) How does the software compute a DI of 6.69?

> DI = [WR/Hand / SD/Hand] x 1000

> [$0.3239/$124.68] x 1000 = 2.59

It depends on the settings in Customize. You'd have to send me the sim as it does not use any of the numbers in the summary but total numbers.

> 3) If 100 rounds are dealt per hour with 2 Billion
> rounds simulated, then why does the software indicate
> 19,999,410 hours instead of 20 million hours? I'm sure
> there must be a logical reason, but it eludes me.

Rounding during calculation. I just saw no reason to keep that particular number to greater than five digits of precision.

[Don Schlesinger]

MJ: > 2) How does the software compute a DI of 6.69?

> DI = [WR/Hand / SD/Hand] x 1000

> [$0.3239/$124.68] x 1000 = 2.59

NW: It depends on the settings in Customize. You'd have to send me the sim as it does not use any of the numbers in the summary but total numbers.

DS: I can help here. The formula quoted above works only for hands SEEN, and it is stated that way on p. 203, in BJA3. Unfortunately, there has been confusion over the years, in this area, and I'm sorry.

If you take win per HOUR (normalizing for 100 hands SEEN, per hour) in your array, divide it by s.d. per HOUR, and multiply not by 1,000, in this case, but by 100 (because s.d. is a square root function), you should get Norm's 6.69.

Let me know.

Don

[MJ]

> MJ: > 2) How does the software compute a DI of
> 6.69?

> NW: It depends on the settings in Customize. You'd
> have to send me the sim as it does not use any of the
> numbers in the summary but total numbers.

> DS: I can help here. The formula quoted above works
> only for hands SEEN, and it is stated that way on p.
> 203, in BJA3. Unfortunately, there has been confusion
> over the years, in this area, and I'm sorry.

I thought Hands Seen is what we are dealing with here, because the last 2 Hands Played buttons are turned off under the Customize button. The Summary was set to 100 rounds dealt per hour. I guess since there is a lag time of 6 rounds, technically 100 rounds are not seen per hour, maybe that is why the DI formula did not work.

> If you take win per HOUR (normalizing for 100 hands
> SEEN, per hour) in your array, divide it by s.d. per
> HOUR, and multiply not by 1,000, in this case, but by
> 100 (because s.d. is a square root function), you
> should get Norm's 6.69.

[$32.39 / $124.68] x 100 = 6.688 ~ 6.69

It works!! But why are we concerning ourselves with Win/hr and SD/hr?

Thanks,
MJ

[Don Schlesinger]

> [$32.39 / $124.68] x 100 = 6.688 ~ 6.69

> It works!!

It works, if you put in the right numbers, which you haven't done. The $124.68 isn't per hour; it's per hand.. You need to multiply by sqrt(15.1), which is 3.886. So, the denominator should be $484.49, and then it will work.

> But why are we concerning ourselves with
> Win/hr and SD/hr?

Thought I explained that, above. SCORE is normalized so that it values an hour of your time at the BJ table, and DI is just the square root of SCORE. The formula on p. 203 of BJA3 works for play-all, but needs the adjustment I've described for back-counting. I'm sorry I never explained that in the book. It has confused people over the years. That's my fault.

Don

[MJ]

> It works, if you put in the right numbers, which you
> haven't done. The $124.68 isn't per hour; it's per
> hand.. You need to multiply by sqrt(15.1), which is
> 3.886. So, the denominator should be $484.49, and then
> it will work.

Opppsss....sorry I meant to write $484.49 in the denominator. That is what I used in my calculation but for some reason I wrote the SD/Hand.

> Thought I explained that, above. SCORE is normalized
> so that it values an hour of your time at the BJ
> table, and DI is just the square root of SCORE. The
> formula on p. 203 of BJA3 works for play-all, but
> needs the adjustment I've described for back-counting.

SCORE Backcounting = [(WR/Hr / SD/Hr) x 100]^2

> I'm sorry I never explained that in the book. It has
> confused people over the years. That's my fault.

No problem, perhaps it will be included in a subsequent edition. In the meantime, maybe Norm can state the formula given above in the help section of the software.

BTW, someone wrote on another board:

N0 = SD^2/EV^2

I understand SCORE Play-all = [(EV/Hand / SD/Hand) x 1000]^2 = (EV^2 / SD^2) x 1000^2 = (EV^2/Variance) x 1000^2

I know N0 = 1/SCORE x 1000^2. This is equivalent to SD^2/EV^2 x 1000^2. Why was 1000^2 omitted when the person wrote N0 = SD^2/EV^2?

Thanks,
MJ

[MJ]

Thanks for the clarification.

> Rounds played - rounds skipped / Rounds played
> 2,000,000,000-1,698,000,000/2,000,000,000

Clever.

> You can't add 566 and 1.7 because the 566 includes two
> hand rounds. Also it includes splits.

Gotcha.

> It depends on the settings in Customize. You'd have to
> send me the sim as it does not use any of the numbers
> in the summary but total numbers.

No need for that, Don figured it out, see below.

> Rounding during calculation. I just saw no reason to
> keep that particular number to greater than five
> digits of precision.

I guess this must be one of those weird computer science things where an exact answer cannot be given for something so obvious.

MJ

[Don Schlesinger]

> SCORE Backcounting = [(WR/Hr / SD/Hr) x 100]^2

Right.

> BTW, someone wrote on another board:

> N0 = SD^2/EV^2

Right. It's the reciprocal of SCORE, without factoring in the various multipliers I use to keep the numbers manageable.

> I understand SCORE Play-all = [(EV/Hand / SD/Hand) x
> 1000]^2 = (EV^2 / SD^2) x 1000^2 = (EV^2/Variance) x
> 1000^2

Right. See above. The multiplier, as you've written it (1,000^2), would actually be one million.

> I know N0 = 1/SCORE x 1000^2.

Right again. You're on a roll! :-)

> This is equivalent to
> SD^2/EV^2 x 1000^2. Why was 1000^2 omitted when the
> person wrote N0 = SD^2/EV^2?

Why are you asking me? :-) Ask HIM! Sloppiness comes to mind. :-)

Actually, it just depends on whether you use per-hand data, or per-hour data, and whether you care about expressing the multipliers or just stating the concepts, without worrying about the decimals.

Don