No Even Money!

[Renzey]

Unlike Assume_R, I am guessing you do not understand risk averse play. I will try to explain.
Sims show that the bet resizing required as risk is increased for a minimal gain in that hand's EV actually hurts your long term EV due to excessive variance.
Which is more important -- maximizing the return on your bets that are placed, or maximizing your overall return by making decisions that have you placing larger bets sooner giving you a snowball effect for every hand match up in the future?

My understanding of risk averse strategy is that aggressive hand decisions which eek out a smidge of extra gain while incurring a proportionally larger blip in ROR would be better passed up. Then you could increase your base unit size (while keeping the same spread ratio) by an amount that will bring you back up to the original ROR. This latter (risk averse) strategy will net an overall higher net dollar earn via the larger bets for the ROR and B/R you are using -- abeit, the difference between the two is small.

My question was: Is it possible to widen the difference between the bottom line yields of optimal EV hand decisions vs. risk averse decisions by passing up many splits and doubles that most of us routinely make, then increase our unit size up even further, back to the same ROR?

The only risk averse indices I'm currently aware of are the seven plays listed in BJAIII, and the only significant one of those is 10 vs. 10. Might it not take many more than that to make risk averse feasability more than an academic argument -- perhaps similar to the floating advantage debate.

 

[assume_R]

My question was: Is it possible to widen the difference between the bottom line yields of optimal EV hand decisions vs. risk averse decisions by passing up many splits and doubles that most of us routinely make, then increase our unit size up even further, back to the same ROR?

The only risk averse indices I'm currently aware of are the seven plays listed in BJAIII, and the only significant one of those is 10 vs. 10. Might it not take many more than that to make risk averse feasability more than an academic argument -- perhaps similar to the floating advantage debate.

Well, whenever a software calculates the Risk Averse index, it is maximizing Certainty Equivalent (CE), which will grow your bankroll fastest for your desired level of risk and for your desired bankroll. So mathematically, there may be a few double and split indices where EV-maximizing would be about 0 or -1 (making it a basic strategy play), while the Risk Averse (RA) index is actually positive (making it not a basic strategy play). So there could be indeed a separate "risk averse" basic strategy chart. But RA indices maximize CE, which essentially does what you asked.

Now an interesting aside is that the CE maximizes log(Bankroll + CE), and hence the CE play is dependent on your bankroll. If your bankroll is extremely large compared to the bets you are making, then the RA index is going to be about the same as the EV-maximizing index. But when you are making bets close to the Kelly fraction of your bankroll, those 2 indices diverge. When your bankroll is very small proportional to your bet sizes (early in your career, or betting a large proportion of your bankroll with a large EV), then maximizing the CE becomes much more important than maximizing your EV, since you really need to avoid going broke, and your basic strategy chart could be wildly different.

So essentially 2 people at the same table could have the exact same hands, both make different plays, and both be correct, for their individual bankrolls, bet sizes, and risk level.

 

[tthree]

risk averse plays I am aware of in my indices:
Doubling 10 v T
Doubling A6 v 2
Doubling A5 v 3
Doubling A4 v 3
Doubling A3 v 3
Doubling A2 v 3
Splitting 88 v T
Splitting 44 v 3
Splitting 44 v 4
Hard doubling 44 v6 at high enough TC
Insuring a blackjack
Insuring a 20
There are probably others like surrender decisions.

Why the hand decision has a slow rate of gain after the index is exceeded is because of the key cards for the hand match up. Any hand match up were the key cards are 2,7,8 or 9 has your index as a poor indication of the actual situation. Therefore any hand decision that increases your stake at risk on this sketchy information of your actual situation beg for a conservatively adjusted index.

You may be right about how you get to a larger bet size to increase your return exponentially. With ramps and other aspects of the games determined by sims based on your indices the actual reason for things are not always apparent. I just see exponential gain in bankroll growth as you resize the optimal ramp with kelly for your growing bankroll over non RA indices.

 

[MangoJ]

Thank you tthree for your nice explanation about RA indices, and how they work (i.e. maximizing CEV for given RoR). As I now understand you basically avoid risky doubles or splits for only marginal increase of EV in terms of lower variance, which in turn allows you to up your betsize (which in turn might up your EV/hour).

I have a question about an opposite system, which actually favors doubles and splits for a marginal decrease in EV. There would actually be an application to that style: Bonus with wagering requirement (WR), where each split, double and insurance would contribute to WR. Reducing the WR by marginal plays would earn real money, as you would avoid playing the WR on a losing game.

Any thoughts from you ?

 

[tthree]

Thank you tthree for your nice explanation about RA indices, and how they work (i.e. maximizing CEV for given RoR). As I now understand you basically avoid risky doubles or splits for only marginal increase of EV in terms of lower variance, which in turn allows you to up your betsize (which in turn might up your EV/hour).

I have a question about an opposite system, which actually favors doubles and splits for a marginal decrease in EV. There would actually be an application to that style: Bonus with wagering requirement (WR), where each split, double and insurance would contribute to WR. Reducing the WR by marginal plays would earn real money, as you would avoid playing the WR on a losing game.

Any thoughts from you ?

I had to do a little research to understand you question. Ecasinos match your first deposit as a bonus. You have to play a certain amount to cash out. I have never played on line so I may be a little ignorant so forgive any misconceptions. I assume counting is useless except maybe for those on the table during that hand. You are playing BS.

This is mostly an RoR problem with a twist. You need to have some money left after the wager requirement. Preferably more than your deposit so you show a profit. I assume you are asking if you should double or split if the EV is larger than the average of -0.05%. I am not sure that you can make that work because you are now playing at a larger average house edge. If I follow the reverse logic you would maintain your RoR by betting smaller but doubling and splitting more loosely. The trouble with this is increased negative variance. I assume the bonus has a positive EV or you wouldn't try this.

You could make the obviously flawed argument of reevaluating more doubles and splits at the new house edge and have a regression that continued to loosen your play.

I guess if you were to search for hand match ups that would be beneficial to double or split liberally look for a tiny house edge and small sum of squares for the EOR. Even with a blackjack the insurance penalty is to large to justify deviating from BS. I am not sure if they deal infinite deck or 8 deck but I don't think any of the BS margins are that close for doubles and splits that you would consider adding them as a deviation from BS.

The weakness of your index as related to reflecting the under or over representation of certain key cards (2,7, 8 and 9 for HIOPT II) is what keeps the penalty low enough for RA play while counting to increase profit. With BS and no counting possible this effect is not present.

 

[MangoJ]

Yes, it is about online casino matched bonus.
The setting is, the casino gives you, say $100 for your $100 deposit. Say, you need to play for $5000 (the wagering requirement WR) to cashout your deposit, bonus, and any winnings. So one needs to play 50x the bonus amount. Any game with a house edge below 2% beats this bonus by simple flat betting.
(such bonuses are pretty rare nowadays...)
We chose to play a blackjack game with house edge of 0.5%, thus we are playing a +EV game, as we expect a $75 win. We flat-bet $10 a hand. Let us further assume that our bankroll (which is not the $200 "session bankroll" in casinos account) is large enough to absorb any variance while flat-betting. So let us ignore all RoR questions.

If one looks into details of the wagering requirement, one will notice that all bets that hit the felt will contribute to the wagering requirement, so a Double Down will not only increase our betsize at stake, it will also lower the remaining wagering requirement (so does insurance and splits - not sure about even money, though).
Since we play the bonus until WR is met (not a specific number of hands), we should be concerned in maximizing EV per total bet (not EV per hand).

My question reformulated would be:
If I'm facing a possible Double down decision where doubling down gives me a marginal 0.2% less return (and hence a Hit (or stand) would be basic strategy), would it be better to double down ?
My guess would be to double down, on the following reasoning:
If I'm not doubling down that hand, I collect full EV and decrease my WR for just $10 (the initial bet for that hand).
If I would double down, I collect reduced EV ($0.02 less) from that hand, but I reduce WR by $20 (as the doubled amount contributes to WR).
The additional WR reduction of $10 is worth $0.05 (as the house edge is 0.5%, the expected loss when I would need to play this $10). Since I play until WR is met (and not beyond), the value in reducing WR is quite real.
Hence doubling down that marginal hand is a $0.03 favourite, despite not being basic strategy.

(I know, if the game itself had a 1.5% advantage for a specific number of hands, I would not want to double down. But the situation is different here, the bonus advantage of 1.5% is per bet, not per hand.)

I don't know if this reasoning is ploppyish (a simulation should proof it).
What is your educated opinion on such a play ? I wonder if there is a "wagering-optimized" basic strategy (not necessary based on indices). It would of course depend on bonus turnover and house edge.

 

[tthree]

As you didn't state all the rules of the game I will assumed 8D S17 DAS. I don't think your 0.2% criterion is optimal. Using that criterion I would sim these hands:

10,11,77 vs A
8,A4,A5,A6,A7,A8,44 vs 2
8,A2,A3,A4,A5,44 vs 3
8,A2,A3,A8,A9,44 vs 4
7,8,A8,A9 vs 5
7,8,12,A8,A9,TT vs 6
9,A5,A6,A7,66,99 vs 7
9,A6,A7,22,33,66,77 vs 8
A7,22,33,77 vs 9
22,33,77 vs T

These splits have a negative expectation for each additional split. This means each matchup has a different answer to whether or not to resplit. Many should only be split once. Again I think a smaller criterion than 0.2% should be employed. Many of these matchups turn a positive expectation into a negative one but fall in the 0.2% range. Someone has some sims to run. Let me know what you find out. I am curious and with travel costs going up I may want to take advantage of +EV home play.

 

[MangoJ]

It is rather rare to find a game with resplit allowed, but yes 8D S17 DAS is pretty the norm of these games. Thanks for your table of critical plays, I will run it through my CA to check tomorrow.

 

[Albee]

You must be playing in a PA casino.

There are a few stores in PA that will not allow 'even money' on a BJ. You have to pay insurance on your BJ. If the dealer has it, your hand is a push and they pay you on the insurance bet. (your only getting half of what you should) If the dealer does not have it, they take your insurance bet and pay your BJ at 3:2. (in a sense, a push).

I know of four PA stores that do this and two who do not. I have not been to the others.

It was told to me the other day the PA gaming board is trying to change it, but as of now that's how it is.

 

[tthree]

You must be playing in a PA casino.

There are a few stores in PA that will not allow 'even money' on a BJ. You have to pay insurance on your BJ. If the dealer has it, your hand is a push and they pay you on the insurance bet. (your only getting half of what you should) If the dealer does not have it, they take your insurance bet and pay your BJ at 3:2. (in a sense, a push).

I know of four PA stores that do this and two who do not. I have not been to the others.

It was told to me the other day the PA gaming board is trying to change it, but as of now that's how it is.

Albee, I must be misreading your post or you are not trying to say what you are saying.

1) If you take even money obviously you win your bet amount.
2) If you take insurance for half your bet your bet pushes and your insurance wins your bet amount.
3) If you lose your insurance bet of 1/2 your bet amount you win 3/2 your bet amount on your blackjack so you win your bet amount.

All three of these have the same outcome of you win your bet amount. There is no difference between even money and full insurance on a blackjack. I do play in Pa.

 

[Albee]

Albee, I must be misreading your post or you are not trying to say what you are saying.

1) If you take even money obviously you win your bet amount.
2) If you take insurance for half your bet your bet pushes and your insurance wins your bet amount.
3) If you lose your insurance bet of 1/2 your bet amount you win 3/2 your bet amount on your blackjack so you win your bet amount.

All three of these have the same outcome of you win your bet amount. There is no difference between even money and full insurance on a blackjack. I do play in Pa.

The stores I talked about in PA will not give you even money.

My mistake.....if you take insurance on your BJ and the dealer has it, your hand is a push and the insurance is paid 2:1....making it 'even money'.


I have no clue on why they do this, yet they do. Thank you for noting my screw up.

 

[21forme]

The stores I talked about in PA will not give you even money.

My mistake.....if you take insurance on your BJ and the dealer has it, your hand is a push and the insurance is paid 2:1....making it 'even money'.


I have no clue on why they do this, yet they do. Thank you for noting my screw up.

The only conceivable reason to not offer even money is if a player is down to his last bet, he can't take insurance, so best he can do is a push. If he is able to take even money, he'd win the hand.

 

[Automatic Monkey]

The only conceivable reason to not offer even money is if a player is down to his last bet, he can't take insurance, so best he can do is a push. If he is able to take even money, he'd win the hand.

Makes no sense from the casino's perspective. Insurance or "even money" is +EV for the house except when used by an AP, and the chances of an AP being down to his last dollar are infinitesimal.

 

[Gamblor]

Yes makes no sense why a casino would not allow even money since its +EV for casinos, always thought maybe PA had some confusing wacky law concerning even money, and some casinos interpret differently.

But then this wouldn't explain why Tropicana doesn't offer even money.

 

[CORed]

It was actually a 3/2 game for this instance.

Does the casino you were playing at deal 6/5 games? If so, two possibilities come to mind:

The dealer was confused and thought that the "no even money" rule for 6/5 applied to 3/2 games.

Management decided to apply the "No even money" rule to all games in the hope of preventing dealers from offering it by mistake on 6/5 games.

 

[Shoofly]

Does the casino you were playing at deal 6/5 games? If so, two possibilities come to mind:

The dealer was confused and thought that the "no even money" rule for 6/5 applied to 3/2 games.

Management decided to apply the "No even money" rule to all games in the hope of preventing dealers from offering it by mistake on 6/5 games.

Just the other day at a place that only offers 3/2, a dealer told me that they were instructed not to offer even money. They will pay it if the player asks for it, but the dealer only offers insurance. I asked why, and she said, "I guess they are just greedy." I decided to widen my spread there. They are clueless.

 

[tthree]

Just the other day at a place that only offers 3/2, a dealer told me that they were instructed not to offer even money. They will pay it if the player asks for it, but the dealer only offers insurance. I asked why, and she said, "I guess they are just greedy." I decided to widen my spread there. They are clueless.

At least the dealer was clueless.

 

[Sucker]

Casinos that do not offer even money implement this rule (or should I say; NON-rule) in an effort to prevent shot-taking by someone who would wait for the dealer to flip over the BJ, and then swear up & down that they DID ask for even money. Penny-wise; pound foolish. :laugh:

 

[Remember Remember]

Casinos that do not offer even money implement this rule (or should I say; NON-rule) in an effort to prevent shot-taking by someone who would wait for the dealer to flip over the BJ, and then swear up & down that they DID ask for even money. Penny-wise; pound foolish. :laugh:

Actually this rule will foil counters entirely. I'm hoping it doesn't spread to more stores. (Just pretend it's the truth everyone...maybe the PCs will believe it.)

 

[Shoofly]

At least the dealer was clueless.

And management. This was a management decision. I hope to ask some of the other dealers about this, to see what their rationale was. I can't think of any reason to do this other than stupidity.