card eating

[flixo]

hi,I m questionning about increasing my EV by eating card in negative counts but when I do a sim , the EV is great improved BUT ror is 4 times bigger with the same bankroll :I simulate 2x7$ in the first case and 7x2$ in the second case (card eating)my question is : why a so big difference in risk for the same amount of money on the table in bad counts ?Everybody advise to cut his minbet if possible to improve EV without changing the risk and I have differents results accordly to the simsthanks

[bigplayer]

Are you playing seven hands in plus counts as well or are you consolidating your bets down to 1 spot or 2 spots in plus counts. Are you using the same wonging strategies?

[Jabberwocky]

Are you playing seven hands in plus counts as well or are you consolidating your bets down to 1 spot or 2 spots in plus counts. Are you using the same wonging strategies?

My verdict is bet consolidation.

[Norm]

I think that you may have been led astray on the power of using card-eating as a sole player in today's games. There exist examples, not so common, where someone can use card-eating to some advantage when playing alone. But, trying to use the concept in a general manner, without a confederate, provides little gain in SCORE.

[ZenMaster_Flash]

I am glad that the O.P. mentioned the quadrupling in the R.O.R.

Meanwhile here is an old post from elsewhere that may shed a bit of light on this:

I've done a few simulations to illustrate a point raised in a thread below. The point is to illustrate that in negative counts there is a specific value to having a card placed in the discard tray that can be expressed in dollars and cents. When you play an extra hand in a negative count you suffer negative expectation on the play but gain the value of the extra cards being placed in the discard tray. The value of the cards being placed in the discard tray depends, however, on the size of your top bets and not your small bets. The more cards that are used per 100 rounds at negative counts the greater will be the frequency of the rounds at the higher counts where bigger bets are made.

To illustrate this I'm using a big spread, 24 to 1, in the game: 6D,S17,DAS,rspAces,late surr, 230 card penetration. The count system used was halves. The play strategy was generated by SBA for these conditions with a risk aversion of half of one percent for a range -3 to +8. Indices used were exactly as given except for 10 splitting which was eliminated except vs 4, 5, &6 at a true of +8. The simulation estimated true count to the nearest half deck and truncated.

Exact optimal spreads were not calculated. This may change the magnitude of the result but I don't think it would change the order. For each position on which a bet was made the following scheme was used: 1 unit was bet on counts of < +1. At true counts of 1,2,3,4,5,&6 the bets were 4,8,12,16,20,&24. The four Sims differ in the true count at which two hands were played. Except for the last sim, which plays two hands at all counts, we really have a spread of 48 to 1. The 24 to 1 spread actually wins most.
Sim one: 1 hand at all counts <+2, 2 hands at >=+2

W/100= 7.831
Sim two: 1 hand at all counts <+1, 2 hands at >=+1
W/100= 7.954
Sim three: 1 hand at all counts < 0, 2 hands at >= 0
Win/100= 8.73
Sim four: 2 hands at all counts
Win/ 100= 8.934

The two interesting points that may stand out to those who have not been through this exercise before. First, you do make more by increasing your total action while having a negative expectation when doing so gives you positive card eating value. The second thing that stands out is that almost the entire gain is derived from playing the second had at the true count of zero. This is where the greatest portion of the true count distribution falls and where you suffer the least negative expectation when employing the card eating strategy.
I expect to be gone for about five days so I can't respond to comments or questions about procedure until then.

Follow Up:

I apologize for that leaving out that information, especially since I was just critical of someone last week for publishing without them. I had messed up something on my excel spreadsheet and was unable to get the correctly adjusted SD per round after manipulating bet sizes. So I redid the Sims with the bet sizes already fixed to the prescribed levels with the number of hands designated in the first post. The win rates differ only slightly.
2 hands at >= +2: win/rd= .0771 SD/rd=9.831 DI= 7.84
2 hands at >= +1: win/rd= .0818 SD/rd=10.069 DI= 8.124
2 hands at >= 0: win/rd= .0889 SD/rd=10.57 DI= 8.41
2 hands at all: win/rd= .0896 SD/rd=10.925 DI= 8.2

We can see that there is not a huge difference in DI, though the win rate climbs when multiple hands are played at negative counts. As mentioned before, the greatest gain is for playing two hands at TC=0. This also shows the greatest DI. Out of all true counts, positive and negative, your greatest gain in both win rate and DI that is derived from multiple hand spreading occurs at a true count of zero. I don't know if this is new, but I've never seen it mentioned before.

After the simulation I saw that the optimal ramp would call for the max bet at +5 rather than +6. With an optimal spread the gain for multiple hands in the negative counts would be even greater. Generally, the more money you have in action at positive counts the greater the value of card eating at negative counts. Also, this has nothing to do with the cut card effect. You simply increase your win rate per round because you use up the low cards in fewer rounds.

When play indices are optimized for card eating you will also see gains in win rate and DI. But there will also be a gain in cover value as many of these plays will differ greatly from normal strategies, such as 10 splitting in negative counts to eat cards.

[flixo]

Are you playing seven hands in plus counts as well or are you consolidating your bets down to 1 spot or 2 spots in plus counts. Are you using the same wonging strategies?

Of course, just cutting my minbet in negative count and back to 2 boxes in plus count.
And I get EV 25% greater with eating cards BUT ROR 4 times bigger with same level of betting in plus count and just cutting ma minbet in 7 boxes...
I don't understand

[bigplayer]

Of course, just cutting my minbet in negative count and back to 2 boxes in plus count.
And I get EV 25% greater with eating cards BUT ROR 4 times bigger with same level of betting in plus count and just cutting ma minbet in 7 boxes...
I don't understand

I don't understand it either. Double check everything to make sure everything is the same. Same + Count bet ramps, Same Bankroll size, Same Betting Amounts on a Dollar Basis at each + True Count, etc. ROR should be lower with the eating approach because everything should be the same except you're betting 7@$2 instead of 2@$7. You've got some setting or something that is creating an apples to oranges type of situation.

[Norm]

The TC frequencies will be dramatically different, raising the average bet and therefore the variance and risk unless the betting ramp is adjusted.

[Norm]

There are some other differences. If you are at a count of -1, and play seven hands, you are very likely to go positive during the round, changing your play and betting. That is, you are more likely to double and split, but gain less by doing so since your original bet is so low. You are far more likely to take insurance than if you were playing two hands, but with a very low bet. Basically, you are losing out on some good opportunities.

When you make most any change in a sim, the results will be different. There are so many interactions in Blackjack.

[flixo]

There are some other differences. If you are at a count of -1, and play seven hands, you are very likely to go positive during the round, changing your play and betting. That is, you are more likely to double and split, but gain less by doing so since your original bet is so low. You are far more likely to take insurance than if you were playing two hands, but with a very low bet. Basically, you are losing out on some good opportunities.

When you make most any change in a sim, the results will be different. There are so many interactions in Blackjack.

okay norm but it s not normal ROR go 4 times bigger , there should be something missing or wrong...
EV is bigger, SD/hour is bigger but ROR shouldn't change so dramatically by cutting our minbet ...

[Norm]

It won't change so much, in fact barely at all, if you use optimal bets for each of the situations.