*Variance,* not expected value, of BS play
Thread Rating:
So the question is, are there *variance* tables for basic strategy plays? another way to answer my question would be to come up with simple probability densities for different outcomes (ie, what percentage of the time do you win 2, win 1, push, lose 1, lose 2 units when splitting 8s vs 10?)
The reason this information is useful is because it can help the more conservative bettor by avoiding volatile bets (like, *ahem* doubling 11 on 10, which I hate doing, or splitting 9s on 2 or 3 instead of standing, which I actually avoid altogether...)
bjstandarddeviationinfo
keep in mind that variance is the square of standard deviation.
Quote: beachbumbabsYou might find some of your answers here, and then explore some of the links within that post.
bjstandarddeviationinfo
keep in mind that variance is the square of standard deviation.
This link is a really good start and shows for basic strategy overall (aggregate probabilities) how likely different outcomes are given the player's action. I'm wondering if someone has performed some kind of Monte Carlo simulation on each play in the BS table simply to highlight the most volatile ones. Any ideas?
Quote: IbeatyouracesIt's actually possible to lose or win up to 8 units on a split.
More at one of my casinos which allows unlimited re splitting except with aces. I've seen some interesting ones.
Quote: arcticfunEveryone knows splitting 88 vs dealer 10 is the "right" way to play, in the sense that the expected value of that outcome is -0.48 vs standing, which has an EV of -0.54 -- but anyone who's ever been in that situation always hates to make that play because the likely prospect of losing two chips hurts the ego.
So the question is, are there *variance* tables for basic strategy plays? another way to answer my question would be to come up with simple probability densities for different outcomes (ie, what percentage of the time do you win 2, win 1, push, lose 1, lose 2 units when splitting 8s vs 10?)
The reason this information is useful is because it can help the more conservative bettor by avoiding volatile bets (like, *ahem* doubling 11 on 10, which I hate doing, or splitting 9s on 2 or 3 instead of standing, which I actually avoid altogether...)
this is an extremely interesting concept to me. does anyone know where to find this data, or how to generate it?
Quote: arcticfunThe reason this information is useful is because it can help the more conservative bettor by avoiding volatile bets
For me the problem with BJ is that the variance is too low already. So, as tempting as it is to deviate for the reasons you give [and, yes, I can relate!] the fact is you need to take your chances IMO. Unless it's AP play, in which case the player needs to guard against giving up even tiny edge.
Quote: arcticfunEveryone knows splitting 88 vs dealer 10 is the "right" way to play...but anyone who's ever been in that situation always hates to make that play because the likely prospect of losing two chips hurts the ego.
do you hit a pair of 8s vs 10? when in that particular situation I have no problem keeping to BS.
just surrender ;)Quote: odiousgambitFor me the problem with BJ is that the variance is too low already. So, as tempting as it is to deviate for the reasons you give [and, yes, I can relate!] the fact is you need to take your chances IMO. Unless it's AP play, in which case the player needs to guard against giving up even tiny edge.
do you hit a pair of 8s vs 10? when in that particular situation I have no problem keeping to BS.
Quote: AnzaExothis is an extremely interesting concept to me. does anyone know where to find this data, or how to generate it?
Try Blackjack Appendix 9 on the WOV site.
Quote: 1BBTry Blackjack Appendix 9 on the WOV site.
Blackjack Appendix 9 on the WOV Wizord of Odds site.[?]
Quote: odiousgambitBlackjack Appendix 9 on the WOV Wizord of Odds site.[?]
Only EVs were found through that link
Those calculated variances will be really close to simulated variances for shoe games.
Quote: Boney526Blackjack is one of the least risky games you can play....
...lol. If you don't have the proper bankroll (which the good MAJORITY of people don't) then it is extremely risky. It's also quite risky in general, over the short run. The only time it becomes stable is over the long run.
Quote: Romes...lol. If you don't have the proper bankroll (which the good MAJORITY of people don't) then it is extremely risky. It's also quite risky in general, over the short run. The only time it becomes stable is over the long run.
But the fact remains that it has extremely low variance.
So, yes, the swings can be big, but they are bigger in almost every other game. Exceptions would be games where all bets pay 1:1 and you never increase your bet mid-hand (craps with no odds, baccarat, PGP, even-money roulette bets). But these games all have much higher house edges than a decent blackjack game.
nothing to laugh about. How do you define "risky" ?
Quote: odiousgambitRomes
nothing to laugh about. How do you define "risky" ?
I've personally had months and months of bad variance and taken swings up and down. Kewlj was nice enough to share that one year he experienced 6 months of losing before closing out the year on the positive side. I also referenced how if you don't have the proper bankroll, which a lot of people appear to underestimate the variance/swings, then it is very risky because your ROR goes skyrocketing upwards.
It's quite possible we were comparing apples to oranges. When I saw his quote my response is the first thing I thought (counting, spreading, BR management, etc). However, if you're just looking at just the game as is, flat betting, one night, then compared to other games doing the same thing yes, it's less 'risky.' My apologies if this was the orange you were referring to.
Quote: Romes...lol. If you don't have the proper bankroll (which the good MAJORITY of people don't) then it is extremely risky. It's also quite risky in general, over the short run. The only time it becomes stable is over the long run.
Betting anything underbankrolled or -EV is by definition risky, but comparing it to most casino games, it really doesn't have high variance per hand. Per hour, yeah I guess it does since it is dealt way faster than a lot of those other games.
I don't really like to make -EV plays in general, but my point was that if you don't like to split hands like 9,9 against 2 and 3 in Blackjack, it seems like you're not much of a gambler to me.
EDIT, but if you were talking about counting, then yes the variance is way higher obviously because you're pushing a marginal edge with big bets.
Quote: RomesMy apologies if this was the orange you were referring to.
OK, I get it.
Pretty ironic the RoR is so real for the AP, clearly can be maddening, yet it's hard to find a game with less variance.
Quote: dwheatleyYou can calculate the variance of every play under an infinite deck assumption. Just create (or borrow) a spreadsheet like the one the wizard made a video of recently.
Those calculated variances will be really close to simulated variances for shoe games.
I was able to find the data I need, thank you .
Quote: arcticfun
The reason this information is useful is because it can help the more conservative bettor by avoiding volatile bets (like, *ahem* doubling 11 on 10, which I hate doing, or splitting 9s on 2 or 3 instead of standing, which I actually avoid altogether...)
If your goal is to minimize Variance then the Strategies that minimizes most varaiance (in order) are
1. Never play
2. Always surrender
3. Never Double , Never split
4. Always Hit Stiffs
But I suppose that is not what you want. You just want to know for some borderline plays where the EV is similar for 2 different plays but the decision with the slighly better EV has a lot more Variance, so as to chose the one with the less EV but a lot less variance.
If you google 'Risk Averse Basic Strategy for Blackjack' or 'Risk Averse Indices' for counters you will find some information on the subject.
Such changes from Basic will be very few and would relate to Splits and Doubles
And someone should use them Only when he is betting a lot of money (compared to his bankroll).
