when to split

I'm trying to figure out when exactly not to use basic strategy. Hence this question:

I'm going to define the term "win %" to mean the percentage of hands on which you are paid off, whether it was a plain ol' win, DD, or split. In other words, the amount of the pay-off is unimportant, just whether you won or lost.

In tournament BJ, I realize it's often a bad idea to DD on the classical DD hands. This is because your win % decreases as a result of your inability to hit again should the DD card be bad. In casino BJ your long term winnings still go up, but in TBJ we don't care about the long term. So I often forego the basic strategy DD to keep up my win %.

I am unsure, however, when it comes to splits. Some splits are sure to raise your win % (for example, splitting 8s against a dealer 7 is better than playing them as a 16). However, some are bound to lower it slightly (like perhaps splitting 4s against a 6).

I'm curious whether anyone would do the calculations to find out when splitting will in fact increase your win %, regardless of what it does in the long term. This may include splits that are not in basic strategy (because they increase the win % of a losing hand slightly, such that it's still a losing hand, which would be bad in basic strategy but potentially good in TBJ). This would be useful information.

(Note, there is *never* a time that a DD will increase your win %. However, against a dealer low card it will either decrease it only slightly or leave it unchanged.)

 

Of recent

only a few days ago there was a thread that discusses splitting. The main concern with splitting or doubling is exposure us to greater losses with greater losses taking precedent over greater wins because its so hard to dig out of a deficit while gaining chips makes you a target by several competitors. I think the time to split or double is fairly simple, when you are being buried by a wildcat and the game is at least 2/3 start pounding your opportunities. If I hold 250 in chips and the leader is at 600 I cant wait to see a pair of deuces, if double after split is allowed, and I have a 50 bet out. The time to make any offensive move depends on the situation and the severity of your deficit.

Interestingly I have been thinking about this while doing a study of players on bet21. Some people are playing several tourneies per day and SnGs too. The hands really add up. So, to defy the probability charts makes me wonder, what good are the charts? The players are playing thousands of hands unlike in the old days when BJTs were found by mostly Vegas players a couple times per week. Now they play the same per day. When does it become a wise move to just follow the charts? If we are playing more BJT than regular BJ why not just forget the charts during BJ?

Wump, here is a link to Ken Smiths double charts. I dont think he has figures for splitting, but have a look around.

http://www.blackjackinfo.com/bjtourn-doublechart.php

 

knowledge

Good to see you in the Teaser section Barney, sharing your experiences

Ken did a nice article in All-In on doubling, and if my memory serves, sometimes doubling does increase your chances of winning over hitting.

Cheers

Reachy

 

Thanks Barney, I have seen threads on splits, but never something that specifically gave the win% for them. I'll check out that link.

-----

Doubling can certainly increase your chances to win a tournament, there's no doubt about that. But it can never increase your chances of winning a given hand.

 

Ask Ken

Double your Fun by Ken Smith; All-In Blackjack, VOl III; Issue 8, Pg 40-42

"Group 2
Doubling Is More Likely to Win Than Just Hitting
10 vs. 2,3
11 vs. 2,3
12 vs. T
13 vs. T
14 vs. 9,T
15 vs. 8,9,T,A
A,6 vs. 2,3
A,7 vs 2,3,4,5,8

Wait! How can this be? By limiting our ability to draw a card, we can actually improve our chances to win the hand? That's right. Group 2 hands actually have a negative one-card cost. Welcome to the Tournament Zone. I told you this game was rarely simple. Here's how it works.

Look at hard 10 vs. 2. If you double, you can't bust: if you hit instead and draw a two, basic strategy would have you hit again and you might bust. Doubling a 10 vs. 2 actually has a higher win percentage than just laying basic strategy."


Hope this helps Wumpie.

Cheers

Reachy

 

Okay, somebody's got some 'splainin' to do. This just plain makes no sense. Here's why:

It is basic strategy to hit a 12 against dealer 2. That *must mean* that if you hold a 12 against 2 you have a better win % if you hit, despite the chance of a bust. So, *any time* you hold 12 against 2, you ought to hit. If not, then your basic strategy has been calculated incorrectly.

Now, if you hold a 10 against dealer 2, basic strategy is to double. If you double and hit a 2, now you're in a situation where you would *like* to hit again according to basic strategy, but may not because you doubled. That means that your options have been reduced by the DD. This is okay in casino BJ, but not necessarily in a tournament.

The key lies in the words "if you hit instead and draw a two, basic strategy would have you hit again and you might bust." True, you might bust, but basic strategy tells you that despite that risk, your win % is still better hitting a 12 against a 2.

So somebody has to be wrong here. I don't want to be the newbie who gets up on his high horse only to get shot down, but it must be that either basic strategy is wrong, Ken is wrong, or we've defined win % in a different way. But if we're talking about purely % of hands won regardless of the payout, the information you quoted is incorrect.


So, to clarify, based on 6-deck H17 basic strategy, the only times doubling down does not decrease your win % are:

10 vs. 4,5,6
11 vs. 4,5,6
A,6 vs. 4,5,6

and in those cases, your win % just stays the same, it never improves.

Doubling 10 vs. 2,3; 11 vs. 2,3; and A,6 vs. 3 (all basic strategy plays) only reduce your win % minutely, because there's only 1 card you don't want to see. However, they *do* reduce it. Not to say doubling isn't the correct move, of course. It often is.

 

Aah, that is vital information! However, it doesn't change the fact that the double-down did not increase your win %! If you hit to 12 you can just stand to have the same win % as if you had doubled down to 12. Reducing your options can *never* increase your win % unless it opens up some other option, but DD doesn't. So I maintain that I was correct, though I missed a nuance that is vital to a complete understanding.

Now here comes the next issue. As tournament players we really shouldn't be basing any decision we make on basic strategy tuned to high EV. We should have our own "basic strategy" tuned to maximizing win %. We should also have a separate "basic strategy" tuned to lowering the loss % (those two might conflict). The latter would be most effective earlier on when we'd just like to maintain a bouyant bankroll, whereas the former would be more effective on crucial hands where we need to get paid off. Such a "basic strategy" would answer my original question about splitting as a result of its creation.

 

tournament strategy

wump -

most basic strategy guidelines - for playing 'straight hands' - maximize win % - or combination of win%/push% - and are a pretty good guide for tournament play - the problem with doubles and splits is that you are increasing your bet - and you have to think risk of ruin - I figure a rough guide to ror is that losses hurt you twice as much as wins help you - so to put more money on the table - I want 2-1 odds of winning on that extra money - that eliminates almost all doubles and most splits

however - as anyone on this site will tell you - sometimes any chance is better than no chance - and you just have to take the risky, long odds play - but for the early hands - I limit splitting to A,A against 2-6; 8,8 against 2-8; and 7,7 against 5,6 and limit doubles to 11 against 5,6. will only vary from this to correlate bets/results, if highly desirable - then only if one of the more advantageous splits/doubles - such as 10 against 5,6; 11 against 3; 9,9 against 6 - have to be pretty close to 2-1 advantage -

 

Thanks, that does give me something to go on. I'd still like to see all of the stats (because sometimes the stats surprise me), but your logic seems sound.

One of the reasons I'm interested in the stats is that I think it would be useful to know how splitting affects win % for elimination hands. If you're ever a lowish bankroll but have the high, it would be quite useful to know whether a split will give you a slightly greater chance to win. I think that would be hard to predict without calculation given the complication of having two different hands.

 

Tss

Wumpie

Of course the act of "doubling" itself doesn't increase the %chance of winning a hand, it's simply that strict adherance to BS reduces your %win rate so as to maximise EV, as you have already pointed out.

The use or otherwise of BS in tournaments is a popular topic that has been discussed many times, most recently in the following thread - http://www.blackjacktournaments.com/bb/showthread.php?t=2963. I have been working on a range of variations on BS myself but it is a big job and I've had to put it on a back-burner at the moment. I do however use a couple of simple variations of BS, that are easy to remember, at key points in the game. I summarise this information in BS-like charts and currently there are 2 variations I use; 1) Push as good as a win, 2) Push as bad as a loss. I also use my own version of Wongs/Kens "Win both ways" strategy that includes surrender strategy. And of course we mustn't forget that "Curts Revenge" is a BS variation.

Cheers

Reachy

 

On a related note

Do the rules ever allow surrender of one or more of the hands resulting from splitting? (in tournaments and/or in regular play.)

 

No. Surrender (if allowed) is only permitted on the original 2 cards before taking any other action. Weird rules excepted, of course.

 

Not in most games

Usually not, but there is a variation called super fun 21 with crazy rules where you can do just about everything.

Here is a link to the rules:

http://wizardofodds.com/superfun21

 

Thanks both

I realised it would have to be a rare option, but it seems like it's more-or-less unheard of, which is what I was hoping for, since it would add some unwanted complexity to my next software project.

 

Whatever that software project is, surrender should be something on a menu that can be added or removed at will for maximum flexibility. Outside of Las Vegas and I guess on-line, most local B&M casinos in the US do not have surrender as a player option.

 

When it comes to DD I have done a 270 degree on that subject.

When I first started playing in BJT I would DD on ANY 11 (even against Aces which is a violation of BS). I also DD all 10's against Ace-9. I DD on 9 versus 4-7 and so on and so forth.

After losing many hands due to the overly agressive strategy, many members here clued me into better ways to think about DD. For example when I place a bet now I do it for very specific reasons - whether correlation, in a progression, to "catch up" on a BR that I am chasing or to protect a lead over a BR that I am leading - OR to have a situation in which I am LOOKING for a DD.

As a general rule of thumbs I have adopted this: The number of chips I have bet is the maximum that I want to lose and the only reason for ME to DD is to have a very specific plan/purpose.

When it comes to splits - once again not a cut and dry subject, at least not for me.

There are many people here, much wiser and brighter than me who will have memorized the statistical advantage you have when splitting 8-8 versus a 7 (which is ALWAYS a good play) or when splitting 3-3 versus dealers 7.

I take a much simplier approach to the problem. Do I need extra chips? Was my bet originally placed in the context that it "allows" a split. Will a split effectively cripple me when heading into an EH? Does my progression allow for a split? Is this a split that I am doing in hopes of getting a DD?

Then there are times that I have two face cards and I will split them versus a dealer's bust card because: the BR that I am chasing just did a DD and I want to continue with the correlation or I'm betting min and the table bets large and I want to "take the dealers bust card" - yes the math guru's here scoff but I've seen it done, had it done to me and tried to replicate it myself

Good topic