Why Split 9's?

[Texan4Ever]

Can someone tell me why you are suppose to split a pair of 9's and stand on the dealers 7, 10, and A instead of treating the hand as 18 since the dealer draws to 17. If you get a hard 18 you stand on everything so why not treat a pair of 9's the same way?

 

[Sonny]

You split 9's because you will win more money than if you stand in those cases. A 9 is a pretty good starting hand, especially against a weak dealer card.

-Sonny-

 

[QFIT]

Can someone tell me why you are suppose to split a pair of 9's and stand on the dealers 7, 10, and A instead of treating the hand as 18 since the dealer draws to 17. If you get a hard 18 you stand on everything so why not treat a pair of 9's the same way?

18 is not a great hand. But you stand since hitting is suicide. Unless you have the option of splitting and improving the hand.

 

[Cardcounter]

Playable 18 is a hand that most people screw up. If you have an 18 against a 5 you are 55% to win. If you split your 9's against a 5 you have two hands that will each win 55% of the time. What would you rather have 2 winning hands or one winning hand?

 

[zengrifter]

Can someone tell me why you are suppose to split a pair of 9's and stand on the dealers 7, 10, and A instead of treating the hand as 18 since the dealer draws to 17. If you get a hard 18 you stand on everything so why not treat a pair of 9's the same way?

The best answer is because high simulation shows it to be the best play (ie, basic strategy). zg

 

[bjsim]

Simulation of www.bjsim.com

Showing an edge of 4-10% for Splitting 9 on 8. You can take a look at many 1,000,000 rounds simulation or play small simulation yourself to evaluate each staratetgy. You can choose When to Split 9 and when not.
(Dead link: http://www.bjsim.com/sample/BSVGS121.5N3NMD2.htm)

 

[QFIT]

Showing an edge of 4-10% for Splitting 9 on 8. You can take a look at many 1,000,000 rounds simulation or play small simulation yourself to evaluate each staratetgy. You can choose When to Split 9 and when not.
(Dead link: http://www.bjsim.com/sample/BSVGS121.5N3NMD2.htm)

You continue to post incorrect information.

 

[Harry1941]

I would split two 9s against the 7 if the count is +5

Harry
Go Jets

 

[21forme]

18 is not a great hand. But you stand since hitting is suicide. Unless you have the option of splitting and improving the hand.

Suppose a casino had a promo offering you 18 every hand you played. You'd lose. 18 is a net loser.

 

[Sonny]

Can't resist

I was going to let this one go since Qfit already responded to it, but there are just too many issues to ignore.

Showing an edge of 4-10% for Splitting 9 on 8.

That’s a HUGE range! An answer that vague is not helpful at all. The advantage for standing on 9,9 vs. 8 (2D S17 DAS) is about 8.56%. So is splitting really better than standing if we only know that it is somewhere between 4% and 10%? It is impossible to make the right decision with that information.

The results on your website show an advantage of 11.6% so I’m not sure where the 4-10% came from. Maybe that was the result of the individual sims before you averaged them together?

As numerous people have said before, 1M rounds is not nearly enough to get accurate numbers. According to your website you only played 2,726 hands of 9,9 vs. 8 to get your EV estimate. Play another 2k hands and see how different the results are. Then try it again. Do you understand why those results are completely unreliable?

The numbers you give are not correct. If we look at the true EV (Archive copy) of that play we can see that it is around 22.1%, not 4-10% (or 11.6%).

So there are obviously problems with your program and your methodology. You really need to address these issues if you want to make your software practical.

-Sonny-

 

[QFIT]

Showing an edge of 4-10% for Splitting 9 on 8. You can take a look at many 1,000,000 rounds simulation or play small simulation yourself to evaluate each staratetgy. You can choose When to Split 9 and when not.
(Dead link: http://www.bjsim.com/sample/BSVGS121.5N3NMD2.htm)

These are the results from the "sim" that you pointed to for splitting 9's, along with the correct numbers:

 

[London Colin]

Leaving aside the issue of the insufficient sample, and the question of why anyone would want to use a simulation to approximate what can be quickly and exactly calculated by combinatorial analysis, there are two problems apparent from the information provided on that bjsim page -

The % edge for splits and doubles is being calculated as function of the total amount bet, rather than just the initial bets.

Pushes are apparently being excluded from the calculation.

 

[QFIT]

...why anyone would want to use a simulation to approximate what can be quickly and exactly calculated by combinatorial analysis,....

Quick yes, accurate no, unless you are playing single-deck with a fixed number of rounds. CA ignores the cut-card effect.

 

[London Colin]

Quick yes, accurate no, unless you are playing single-deck with a fixed number of rounds. CA ignores the cut-card effect.

That's true; I was forgetting about that subtlety.

But is it fair to say that it generally remains the convention to ignore the cut-card effect when quoting such figures?

 

[QFIT]

That's true; I was forgetting about that subtlety.

But is it fair to say that it generally remains the convention to ignore the cut-card effect when quoting such figures?

Yes, a convention that I find hilarious. How can you quote a number to 8 or 16 decimals that ignores reality. That's why I created actual numbers in my book instead of using the convention that ignores how BJ is actually played. (Sorry no link as I get criticized for links to the correct tables.)

 

[bjsim]

Simulation

Thanks I agree. Bjsim may not be the ideal tool to get precises odds for advantage player. However it is fun simulation and show the average player what can happens over so many rounds. I calculated the EV based on total bets, Are you recommending to use it based on initial bet.
I will fix the EV on Split and double to be based on the initial bets.

 

[London Colin]

I calculated the EV based on total bets, Are you recommending to use it based on initial bet.
I will fix the EV on Split and double to be based on the initial bets.

You need to think in terms of rounds, rather than hands. However many splits and/or doubles take place (0, 1, or more), the net result is an amount won/lost by the player for that round, which needs to be expressed as a percentage of the initial bet. Then you can determine whether splitting/doubling is more profitable than hitting/standing.

And you can't ignore pushes.

 

[Mr. T]

Let me put in my 2 cents worth here if I may.
Wow, this is high brow stuff here. Really outside my depth. What with sims and combinatorial analysis, etc.
From the figures above I make the average gain from splitting 9's is 25.4%.
Permit me to chose an arbitary figuire of pair of 9's occuring 1 in 200. ( Smarter guys than me like Sonny would no doubt have the actual correct number).
Then multiplying this out 25.4/200 is 0.13%. So not splitting a pair of 9's would increase the HA by 0.13% I think this is the answer the original poster is looking for.

 

[London Colin]

From the figures above I make the average gain from splitting 9's is 25.4%.

I think you are confusing the EV of splitting with the difference between the two EVs (splitting and standing), which is what would represent the gain from splitting.

 

[Mr. T]

So Mister. BS says you shoud split 9's.
If you don't split, what is the penalty in terms of the increase in House Advantage in % terms over BS.
That is what the OP, myself and others would like to know.