BJ Surrender Playtech

Hi all,
I'm experiencing a very bad run in blackjack surrender at a play tech casino. Results are in 2nd deviation maybe?
I'm not good at it so I report the result
Wager ~40.000
Loss 950
Flat bet 10

I know I can't be suspicious yet. Anyway this situation made me wondering if I was failing in optimal strategy and that's why I'm writing here.

I used the one computed using the site beatingbonuses (didn't know yours) that is computed by applying the rules and I selected per unit wagered instead of using per hand.
This is the result:

*wanted to upload table of beatingbonuses*

That is slightly different from the one here

*wanted to upload your table reported in playtech review*

That is again different with the one obtained using wizard of odd calculator.

Differences are in
16 vs 10 surrender if possible otherwise hit or stand?
Soft 18 vs 2 double if possible or stand?
Soft 18 against A hit or stand?

So which one is correct and why? Can you help me with the math?
If the one I used is wrong could it affect so much the results?

I was not able to upload the images

Hi and welcome to WoV,

I can't run play-tech on my pc anymore, so are these the correct rules*** (if not, post the rules thanks) ?

***: 8-decks, BJ pays 3/2, Stand on all 17, Double after split, Split 3 times max (Aces once), Double after split, and later surrender

House edge: about 0.35%

Quote: Giorgior27

Hi all,
I'm experiencing a very bad run in blackjack surrender at a play tech casino. Results are in 2nd deviation maybe?


I'm not good at it so I report the result
Wager ~40.000
Loss 950
Flat bet 10

I think it is within ~1.21 SD, going by the beating bonuses website.

Edit/update (845 pm):

For comparison 68.26....% of the time, the results should be within 1 SD and 80% of the time the results should be within 1.281552 SD.

Note: since I don't know how to convert SD to % I am using the table at the bottom of the "Rules for normally distributed data" section from this link below:

https://en.wikipedia.org/wiki/Standard_deviation#Rules_for_normally_distributed_data

Quote:

I know I can't be suspicious yet. Anyway this situation made me wondering if I was failing in optimal strategy and that's why I'm writing here.

I used the one computed using the site beatingbonuses (didn't know yours) that is computed by applying the rules and I selected per unit wagered instead of using per hand.
This is the result:

*wanted to upload table of beatingbonuses*

That is slightly different from the one here

*wanted to upload your table reported in playtech review*

That is again different with the one obtained using wizard of odd calculator.

Differences are in
16 vs 10 surrender if possible otherwise hit or stand?
Soft 18 vs 2 double if possible or stand?
Soft 18 against A hit or stand?

So which one is correct and why? Can you help me with the math?
If the one I used is wrong could it affect so much the results?

I was not able to upload the images

If you make a 3 card hard 16 vs 10, compared to hitting it costs you about 0.01%^^^ to about 0.06%^^^, every time you stand on that hand.

^^^: Cost of "errors" is ~ 0.01% for a player with 10, 4 and 2 in their hand, and ~ 0.06% for a player with 10, 3 and 3 in their hand.

If you have an A and 7 (Soft 18) vs a 2, compared to standing it costs you about 0.27% every time you double.

If you have an A and 7 (Soft 18) vs A, compared to hitting, it costs you about 0.56% every time you stand.

----
Update (about 830 pm):

going by the playtech rules, as listed on the WoO website,

you would stand on a 16 if you had 10, 4 and 2 in your hand, hitting costs you about 0.01%
you would hit on a 16 if you had 10, 3 and 3 in your hand, standing costs you about 0.06%

doubling on soft 18 vs a 2 costs you about 0.3%, when compared to standing

standing on soft 18 vs A costs you about 0.5%, when compared to hitting.

Also, to work out the "true cost" of these errors, you would have to know the chance of getting these hands and then multiply that figure by the ones I provided above.


----
Update (about 9pm)

So you can work out the strategy and figures

Here are some basic strategy calculators >>> https://wizardofodds.com/games/blackjack/strategy/calculator/

http://www.beatingbonuses.com/bjstrategy.php (you would probably know about this one already).

Here is a site that will tell you the correct play for a specific hand, and to help work out the cost of deviating from the best play for that hand >>> http://www.bjstrat.net/cgi-bin/cdca.cgi

Thanks for your answer.
I continued playing and the result are still strange,
I played 48000 and I'm down of 1200 using always flat bet of 10.

If I insert these results in the "are my result fair?" section of the beatingbonuses site I get:
Chance of loss =8%
Chance is 1/12
N of SDs 1.4

Is it possible that I'm still so unlucky? I made different sessions at different times so the probability should go down of a long bad run like this one, am I wrong?


About the rules, I don't know if the one listed in WoO are right, here the rule taken directly from the game i'm playing:

Decks 6, shuffle after each hand
Double on any 2 cards and doblue after split allowed
Late surrender (dealer first peeks for BJ both on ace and ten and then if it's not BJ you can surrender)
Can't hit on splitted aces
No resplit
Dealer stands on soft 17
7 card charlie pays 1:1
BJ pays 3:2

Using these settings in the calculator of beatingbonuses that you linked and setting per unit wagerd (it is correct or should i use per hand?), i get
- Hard 16 against 10 surrender if possible otherwise stand
That is always the same, why don't i get different info about the different combination for the 16s

-Soft 18 double against 2
I get this if i set it per wager while using per hand i get what you said so stand. Can you explain me why? And which one should i consider?

-Soft 18 hit against ace
Here we all agree then


Using wizard of odds calculator it says for hard 16 to surrender or always hit, so it's different.

Maybe the best one to use is the last one bjstrat? Which adapts every hand?


EDIT:
To compute the true cost as you said what should i do? For example starting from A/7 vs 2 that i alwasy doubled and instead i should always stand it costs 0.003.

Probability of A * Probability of seven * probability of 2 dealer * 0.003?

No of rounds = n = 40000/10 =4000, ev = -0.0035, SD =1.14, Loss = 950/10 = -95 units

For case of loss more than 95 unit,

Z < (-95 +4000*0.0035)/1.14/(4000)^0.5

Z< -1.123

Probability = 13.07%

Thanks ssho88, so now it's correct that the probability of my new results are 8% right? I know that someone will fall in this % but i can't get if my errors could cause that.

Offtopic:
Using the combinatorial simulator i noticed that i can take out cards from the shoes and this could be usefull in the splitting hands correlation. Is this also usefull if i split my bet in 5 of 2$?
Variance should decrease and maybe I would not faced this badrun.
The 5th hand will have an advantage in house hedge comparing to the same at first hand? After 4 hand i will know at least 8 cards more and probably something >8.
I was wondering if this could cut a bit the house edge during wagering requirements.

Quote: Giorgior27

Thanks ssho88, so now it's correct that the probability of my new results are 8% right? I know that someone will fall in this % but i can't get if my errors could cause that.

Offtopic:
Using the combinatorial simulator i noticed that i can take out cards from the shoes and this could be usefull in the splitting hands correlation. Is this also usefull if i split my bet in 5 of 2$?
Variance should decrease and maybe I would not faced this badrun.
The 5th hand will have an advantage in house hedge comparing to the same at first hand? After 4 hand i will know at least 8 cards more and probably something >8.
I was wondering if this could cut a bit the house edge during wagering requirements.

No of rounds = n = 48000/10 =4800, ev = -0.0035, SD =1.14, Loss = 1200/10 = -120 units

For case of loss more than 120 unit,

Z < (-120 + 4800*0.0035)/1.14/(4800)^0.5

Z< -1.307

Probability = 9.56%

Why do you use standard deviation of 1.14 and in beating bonuses for playtech surrender is 1.13?
And can you explain me why this SD 1.13 is different from the one per wager SD that is set to 1.09 ?
Which one should i consider?

Quote: Giorgior27

Using these settings in the calculator of beatingbonuses that you linked and setting per unit wagerd (it is correct or should i use per hand?), i get
- Hard 16 against 10 surrender if possible otherwise stand
That is always the same, why don't i get different info about the different combination for the 16s

I would personally use "Per Hand (standard)" if using the beating bonuses bj strategy page (I don't know why the strategy changes for "Per Wager"

Also to answer your next question, because it is a basic strategy calculator.

Quote:

-Soft 18 double against 2
I get this if i set it per wager while using per hand i get what you said so stand. Can you explain me why? And which one should i consider?

Sorry, I don't know enough to answer that.
But going by the bjstrat figures, my answer near the bottom of this post about the "true cost" may help you realise that it is not a "major mistake" to double instead of stand with the above hand.
In other words, if you are trying to get through wagering as quickly as possible" ( say to do with a bonus) then I would consider going against basic strategy, and double a soft 18 vs a 2.
Note: I would always try to play best or basic strategy, when I am NOT trying to "wager through a bonus".

Quote:

Using wizard of odds calculator it says for hard 16 to surrender or always hit, so it's different.

Maybe the best one to use is the last one bjstrat? Which adapts every hand?

If you want to play each hand "perfectly" then yes bjstrat is probably the best one to use.
If you want to play each hand as "quickly/efficiently as possible" then I would use a basic strategy, like the one on Wizard of Odds or beating bonuses websites.

Quote:

EDIT:
To compute the true cost as you said what should i do? For example starting from A/7 vs 2 that i alwasy doubled and instead i should always stand it costs 0.003.

Probability of A * Probability of seven * probability of 2 dealer * 0.003?

For six-deck BJ the probability of getting a player A-7 vs dealer 2 = about 1 / 1,088

So going by BJStrat website, the cost is about 0.003, so the "true cost" is 0.003 x 1 / 1008 or about 1 / 362,653

In other words, doubling a " player A-7 vs a dealer 2" instead of standing will cost you 1 unit every every 362,653 units bet.

----
Lastly, here is another link that should be good to use for working out the house edge, correct strategy, etc

https://www.blackjackinfo.com/free-blackjack-combinatorial-analyzer/

---
I was in a hurry with this post as I have to go to work, so spelling etc not checked (99.99% sure that the figures are correct).

Is this a live dealer, hand shuffle? If so for each next questions never mind...

Can I ask what state you are located in?

And do you play EXACT BASIC STRATEGY to the tee? If not how do you deviate? Do you hit 16 against 10 sometimes?

Thank you ksdjdj, you were so clear.

I'm using BJ to clean a cashable bonus that "in theory" was really easy because was a 1000 bonus on a 1500 deposit with wager 70xb.
Roughly 70.000* (-0.0039) = -273 over 1000 of bonus it was acceptable.

Probably to move a bonus using bjstrat is too much time consuming and does not compensate for the gain in terms of house edge. I will think about that.

Then A-7 vs 2 is something so small that i could not even notice and so probably is for the others differences, thank you for the explanation.

The program that you linked is even better than the bjstrat online analyzer because splitting pairs consider automatically the other hand on the table.

In this regard, do you think that playing 5 hand could bring an advantage? In which way should i do that? I mean a flat 10 should be splitted in 5*2$ or maybe i can also use something like 5*4$? The total bet would be higher but the variance is similar to one single of 10$.
Plus i should add the advantage obtained at each hand that know at least 8 cards more and increase from the 1st to the 5th.
I know it's not something that give a positive advantage to player but is this usefull to cut more the house hedge?

Quote: heatmap

Is this a live dealer, hand shuffle? If so for each next questions never mind...

Can I ask what state you are located in?

And do you play EXACT BASIC STRATEGY to the tee? If not how do you deviate? Do you hit 16 against 10 sometimes?

It is playtech bj surrender RNG so every each hand the shoes are shuffled.

I'm in Italy and yes I played exact basic strategy (the one on beating bonuses) or at least I don't deviate deliberately, if i did that it would be a mistake/misclick (very rare)

Quote: Giorgior27

(snip)
In this regard, do you think that playing 5 hand could bring an advantage? In which way should i do that? I mean a flat 10 should be splitted in 5*2$ or maybe i can also use something like 5*4$? The total bet would be higher but the variance is similar to one single of 10$.
Plus i should add the advantage obtained at each hand that know at least 8 cards more and increase from the 1st to the 5th.
I know it's not something that give a positive advantage to player but is this usefull to cut more the house hedge? (snip)

Here is my opinion, I will split this into two parts/scenarios,
The first part is if you have enough reserve bank and can keep getting the bonus (eg if it is a "reloadable bonus")

Part 1:
Any advantage of knowing more cards by playing 5 x $2 would probably be cancelled out by the "slower play". (5 x $4 could be acceptable, but you need to work out your "turnover per minute" and compare it to $10 x 1).

If I had started with the same deposit and bonus, and was down about $1200, then I would do this:
Bet the maximum allowed*** by the terms and conditions set out in the bonus, or $130 (whichever is less^^^)

***: Always read the terms and conditions before considering anything like this (The majority of bonuses at the moment that I have seen are currently "weak" in terms of "max bet per game". so this strategy may not be "doable")

^^^: When playing a bonus on blackjack, never bet more than 1/4 of your remaining $ in your casino account (this is so that you can still double and split correctly).


The second part, is if you don't have enough bank in reserve, and/or the bonus is a "one off promo" (eg. a "1st/new member deposit bonus", with no more bonuses on offer after receiving that one).

Part 2:
I would bet no more than 5 x $2 or 1 x $5, so that I can "lower the variance" and "cut my losses"

Important: This is only applicable for the above scenario you are in, "being down $1200, with about $22,000 wagering to go"

It is a one time bonus for new member so I will follow the second part of you message. I will try to decrease variance.

I tried to use 5x4$ one minute but it is too slow anyway: it takes one minute almost for one play because each seat I have to cancel the player hand and copy it in "other players hand".
I was just curious but maybe the advantage is not enough to accept being so slow.

I though that I should keep the high variance to hope in a come back in the expected value but maybe it requires too much hand and I could end up without see it right?

Quote: Giorgior27

I'm using BJ to clean a cashable bonus that "in theory" was really easy because was a 1000 bonus on a 1500 deposit with wager 70xb.

Are you sure that they don't have a reduced 'contribution to wagering requirement' for Blackjack? It usually only counts 5% to 10%. Check the small print.

Thank for your advice OnceDear. I'm sure because it's a pending bonus and each "chunk" was released correctly according to my hand history.
This bonus won't probably last much i know. The first one in a while that still didn't adapt to other casino's policy.