Let say you play 6D game with 2 or 3 other players for few hours a day at 3 or 4 casinos. You wong in/out and spread as count dictates, so you when wong out but still sitting at table does the rounds you see counted in your N0? on other hand, how about when you spread to 2 or 3 spots. For simplify the matter we disregard the heat.
Good cards to you.
So, if N0 is the number of played rounds accumulated EV to overcome one negative standard deviation, then how do the rounds with zero EV (wonged out rounds) contribute to EV?Originally Posted by therefinery
How then is EV calculated?
The reality is that I want N0 to tell me how many rounds I must PLAY to get past one standard deviation. Actually, I want to know N0 for 2 standard deviations (N0 X 4). I can stand and backcount forever and I will never get to N0!
Last edited by Stealth; 07-05-2018 at 09:07 AM.
Luck is nothing more than probability taken personally!
I think we’re talking semantics here. You can play 0 or 40 hands in an hour backcounting just as you could see 50 or 0 hands with a positive count over 4-5 shoes playing them. Saying that you don’t count the unplayed backcounting hands toward NO is the same as saying you accumulated negative EV In those 4-5 shoes all play where a count never happened. If you want to look at it that way is fine if you understand what you’re looking at , but simming wonging only on cvcx frequently comes up at what maybe 70% of the hands of a white rabbit/all play approaches NO? If the observed hands were not to count it would be more like 20% of the rounds.
Last edited by mcallister3200; 07-05-2018 at 09:37 AM.
OK, without going way overboard in the theoretical.
I want to know the number of rounds I need to play to overcome one negative standard deviation (N0).
In my scenario I use a spotter/BP protocol and the spotter is playing a break even game that I am not concerned about.
How may rounds (and call ins) do I need the BP to play to get to N0 for this game. I know by the sims and confirmed by actual sessions stats that the BP will play an average of 12 rounds (6 deck) when called in at TC+2.
I run a sim only playing at TC+2 and higher with the bet ramp we are using. I get a n0 that answers my question and I really do not care about rounds outside of my TC range. If I want to know the number of shoes required I can divide n0 by 12 rounds and get an reasonable estimate of how many call ins I need to generate enough EV to overcome 1SD. Including the rounds that were wonged out only confuses my question. Then I need to determine how many shoes a spotter must play to get that many call ins. Then how many spotters are needed to support 1,2 or 3 BP's. Guidance is available in BJAIII.
Three, what stats? If your bet = 0 then it generates no EV and is not mathmatically a factor in the calculation of N0. My objective is to know how many rounds I need to play to overcome the -1SD. It is a measure of the quantity of work to be done, not a measure of time to do the work.
Luck is nothing more than probability taken personally!
The sim stats. The sim counts rounds observed toward the sim's n0. If you use your own software then you can have the sims do what you want them to do. The sim data I use figures the percentage edge (IBA) you have for all the rounds observed or played. It divides your EV in dollars by your average bet per round observed to get IBA. It gets the variance by taking the total weighted variance and dividing by the square of you average bet per round observed. Then N0 is variance divided by the square of the IBA. Every step along the way uses average per rounds observed. The sim also gives the percentage of rounds you play so you could use the sim data to figure out at least an approximation of what you want to know. But I am sure you use your own software to get to your answers.
Yes, you are correct. I use my own software to do the math but it is based on CVCX sim data for win/loss, SD, frequency.
I believe your formula skews N0 to be smaller due to the use of a larger number of rounds. The formula I am using is the variance divided by the square of the win rate per round.
I will go back and verify my sources for the formula but it makes logical sense to me.
Luck is nothing more than probability taken personally!
Unless I'm missing something....wouldn't you just multiply the % of hands the BP plays against the N0 (N0 defined as where observing a hand counts). If a BP plays 5% of the time and the N0 is 10,000, then shouldn't the "playing N0" be 0.05*10,000 = 500 rounds played? Same logic for # of call-ins, which would be 500/12 = 41.66. The number of shoes required to spot to get that many call-ins (on average), wouldn't that be 1/Call_In_Rate, where Call_In_Rate = probability a spotter calls in a BP to a specific shoe?
I don't know about # of spotters for # of BPs, but I'd think that's a bit more complex. If it were me, I'd probably start with figuring out how long (minutes) a spotter is going to be playing (average time) before he makes a call in and how long the BP will be at the table. I'm not sure if variance (in time) would be something that needs to be considered. You could probably write a quick macro in Excel or a script and sim multiple spotters playing side by side (different tables obv). From there, you could determine how many spotters and BPs you want -- first, figure the number of BPs to maximize EV (eg: every table that hits the pivot point to make a call in, gets a BP). Then subtract 1 BP from the total # of necessary BPs, see what the EV looks like, then subtract another BP....
"Everyone wants to be rich, but nobody wants to work for it." -Ryan Howard [The Office]
This topic is covered in some detail in BJAIII. You are correct that it is a complex subject, especially when you apply even more criteria like:
No call in with less than 3 decks remaining which effects the percentage of shoes that will become BP playable.
My rule of thumb is that a spotter will find a positive shoe about a third of the time.
Luck is nothing more than probability taken personally!
I'm stat challenged and still struggling with CVData. But IIRC from stats 101, Bayesian stats might be a useful adjunct to NO and CVData. Bayesian stats, I think, gives the probability of empirical data (e.g., BJ results) matching theoretical results. Thus, one could apparently match one's actual results to the theoretical CVDATA results, and see if one's practical experience is anywhere close to predicted results. Maybe Don S and Norm have already done this, but I'm too unsophisticated to use it.
BTW, for the realists amongst us, the randomness of flipped coins (fair and biased) and Bayesian Stats (using colored ping-pong balls), was EMPIRICALLY proven by John Kerrich when he was imprisoned for 2 years by the Nazis. He actually got some publications out of his experience after the War. I'm hopeful that Don S. will take a similar approach with physical cards in BJA 13 or 14, lol.
https://en.wikipedia.org/wiki/John_Edmund_Kerrich
bj results
Posting Permissions
-*-
Reply With Quote
Bookmarks