I-18 and splitting 10's

[sagefr0g]

KC, very enlightening -- thanks a ton! I think this shows that re-splitting 10's vs. 5 or 6 at TC +7.5 basically doubles the gain beyond standing as compared to splitting one time only. Since standing with 15 vs. 10 is worth 37 thousandths of a percent in the I-18 computations and splitting 10/10 vs. 6 was worth 17 thousandths (with another 17 thousandths for splitting 10/10 vs. 5), splitting and re-splitting the two hands as a combined strategy would seem to be worth more than the 15 vs. 10 index play. Did I jump to any conclusions? Anybody?

I might suggest that those who don't split 10's for cover reasons might re-think it. I know they say that only two kinds of players split 10's -- idiots or counters. But more glaring than that is that only one kind of player insures his 13. In either case you either slip it past'em, or you don't.

i dunno if any conclusions were jumped to but the tc at which one splits tens vs 6 is plus four for I18, no? and the original point was in reference to I18.
but the examples above used tc=5 & tc=7.5 .
don't we want to know the stats for tc=4 and then splitting again if we got another ten and maybe splitting again if we got another ten?

 

[zengrifter]

The problem I have with this type of sim-analysis is that it shows the long-view,
but I feel like it doesn't adequately reflect the actual value IN THAT MOMENT? zg

Firstly, it's not a sim. It's an exact calculation. :whip:

Secondly, you may have a point. The data just shows EV of someone with a pair of tens versus a 6 that chooses a strategy of splitting at every opportunity up to a certain number of splits at the very beginning of the hand and following through with that strategy come hell or high water. Each new splitting opportunity will have it's own modified EV, though, depending upon what additional cards have been seen.

Got it, not a sim. An algorithm?

What I mean to say is that the 10s gain is small per 100 hands, but in that moment when
the opp presents itself the 'subjective' gain is larger, like winning a poker pot say. zg

 

[Renzey]

i dunno if any conclusions were jumped to but the tc at which one splits tens vs 6 is plus four for I18, no? and the original point was in reference to I18.
but the examples above used tc=5 & tc=7.5 .
don't we want to know the stats for tc=4 and then splitting again if we got another ten and maybe splitting again if we got another ten?

If all you are is at +4 TC and you catch a 10 on your first split, then you won't re-split. Depending upon your current depth in the shoe/deck, you'd need to be at something beyond +4.0, in order to be at +4.0 after you catch the 10 and re-split. But sometimes it goes as follows:
You're at +4.0, split and catch a 10 on your first hit, so you don't re-split. On your second 10 you catch a 6 which brings you back to +4.0 TC, so you re-split.

In addition, remember that the index says you'll split at +4.0 TC -- or higher. Assuming that, the average TC you split your 10's at will be something higher than +4 TC. Depending upon pen, it will be somewhere around +6 in a shoe game. That point (whatever it actually is) , is where I think you want to calculate your gain.

 

[sagefr0g]

If all you are is at +4 TC and you catch a 10 on your first split, then you won't re-split. Depending upon your current depth in the shoe/deck, you'd need to be at something beyond +4.0, in order to be at +4.0 after you catch the 10 and re-split. But sometimes it goes as follows:
You're at +4.0, split and catch a 10 on your first hit, so you don't re-split. On your second 10 you catch a 6 which brings you back to +4.0 TC, so you re-split.

In addition, remember that the index says you'll split at +4.0 TC -- or higher. Assuming that, the average TC you split your 10's at will be something higher than +4 TC. Depending upon pen, it will be somewhere around +6 in a shoe game. That point (whatever it actually is) , is where I think you want to calculate your gain.

right, ok, i'm seeing your point. errhh, well if i'm remembering correctly I-18 was set up as a sort of a 'fudge factor' sorta thing, set up for some certain rules & i think for really a four deck game (sort of in the hopes that would be a good trade off for multiple deck games) using Hi/Lo. so i would guess in keeping with the spirit of I-18 one would want to consider this re-splitting thing from the initial conditions Don S. used.
whatever, just playing around some with kc's TDCA program one can see that the idea of re-splitting the ten value cards, or even splitting them in the first place is a rather 'nebulous' thing depending on the pack composition. but i guess, 'nebulous' as it may be, that still the expected values for splitting, while sometimes lower than standing may be still a rather high value, hence not so bad of a move to make, sorta thing.
the images below are some examples first two images for double deck, the last for a four deck game (note the four deck game is tc=8, but splitting isn't a higher ev than standing for that particular pack composition):

 

[k_c]

i dunno if any conclusions were jumped to but the tc at which one splits tens vs 6 is plus four for I18, no? and the original point was in reference to I18.
but the examples above used tc=5 & tc=7.5 .
don't we want to know the stats for tc=4 and then splitting again if we got another ten and maybe splitting again if we got another ten?

I chose those compositions for no virtuous reasons other than they are reflective of an average HiLo composition. The compositions are set pre-deal so after player picks up his cards and views dealer's up card, the pair of tens and up card still needs to be included in the count. cdca computes differently than tdca and to input the identical situation in tdca you would first input the composition and then deal player cards and dealer up card. cdca removes cards after 'Compute' is clicked while tdca removes them and alters count display before 'Calc' is clicked.

My method is to input a number of decks and then remove half of the cards of each rank, leaving half of a shoe. In order to increase count I remove (2-6) in equal amounts, leave (7.8.9) unchanged, add tens in the amount of 4 times what was removed for each (2-6), and add aces in an equal amount of what was removed for each (2-6).

For example 8 decks:
alter composition to 16,16,16,16,16,16,16,16,64,16 (2-A)

Remove 1 ea (2-6), add 4 tens and 1 ace
New composition is 15,15,15,15,15,16,16,16,68,17 which is HiLo RC=+10, HiLo TC=10/4 = 2.5

Remove 2 ea (2-6), add 8 tens and 2 aces
New composition is 14,14,14,14,14,16,16,16,72,18 which is HiLo RC=+20, HiLo TC=20/4 = 5

etc... (RC increases by 10, TC increases by 2.5 for each successive entry)

All I was looking to do is output some data to show where a trend may exist.

By the way there is an online version of cdca on my website so anyone can use it to duplicate the data or come up with other data. There is also a downloadable free for personal use console version there.

 

[Renzey]

I - 18 revisited

Don Schlessinger says he'll be revisiting the Illustrious 18 along with Zenfighter. This may shed new light on the component value of splitting 10's as part of one's index repertoire.

 

[alwayssplitaces]

There will be a lot of heat if you re-split a max bet pair of tens. But it sure will be a good way to annoy the ploppies.

 

[zengrifter]

There will be a lot of heat if you re-split a max bet pair of tens. But it sure will be a good way to annoy the ploppies.

Typically there is not a lot of heat, at modest betting levels at least. zg

 

[Blue Efficacy]

Twice in recent months, since growing the balls to do it, splitting 10's to four hands has saved me from almost losing my trip roll. It did get me some heat one of those times though but have since played that store and hit them hard while rated many more times without incident.

Also both of those times the other players were cheering me on to re-split, somewhat unusual.

 

[Renzey]

Bringing Heat & Annoying Ploppies w/ Index Plays

There will be a lot of heat if you re-split a max bet pair of tens. But it sure will be a good way to annoy the ploppies.

Let's review. Standing w/ 16 vs 10, Doubling w/ 10 vs 10, Insuring a 13, Doubling w/ 8 vs 6, Doubling on A/8, Hitting 13 vs 2, Splitting 10's. In addition, Doubling for less w/ 12 vs 2 or 3, Often Insuring a random hand for a tenth of a bet and tipping the dealer the original Insurance bet if it wins, Sharing your 7/7 vs 3 with the ploppy next to you, Jumping in on his 9 vs 7 double, Going bust then buying in for your max bet and going all in with it -- maybe three or four times in a row.

Will splitting and re-splitting 10's really jump out here??

 

[Automatic Monkey]

Don Schlessinger says he'll be revisiting the Illustrious 18 along with Zenfighter. This may shed new light on the component value of splitting 10's as part of one's index repertoire.

I've been doing some work on indices, as used by a counter instead of a flat better.

Turns out, splitting 10's is mighty, especially if you are counting the ace as a positive card. And for a counter, the decisions he has to make when there is a lot of money on the table mostly treat the ace as a high card- splitting 10's, doubling 8-10, doubling A8 and A9. Whether or not to hit or stand on 16 vs. 10 may be the most powerful index for a flat better but it is insignificant for a counter in a multideck game, who always knows exactly how he is going to play that hand when he has more than a minimum bet down. Thus he can forget about that index, and forget about the contributions the EoR of each card make to that index.

So for a shoe player using a large spread, I calculated the playing decision value of each card weighted for the amount of money he has on the table when he has to make the decision. For a level 2 system, I found out that for maximum playing efficiency the best system tags to use are {A...10} {-1,1,2,2,2,2,1,0,-1,-2}

Look familiar? :toast:

 

[blackjack avenger]

I Think You Would Agree

I've been doing some work on indices, as used by a counter instead of a flat better.

Turns out, splitting 10's is mighty, especially if you are counting the ace as a positive card. And for a counter, the decisions he has to make when there is a lot of money on the table mostly treat the ace as a high card- splitting 10's, doubling 8-10, doubling A8 and A9. Whether or not to hit or stand on 16 vs. 10 may be the most powerful index for a flat better but it is insignificant for a counter in a multideck game, who always knows exactly how he is going to play that hand when he has more than a minimum bet down. Thus he can forget about that index, and forget about the contributions the EoR of each card make to that index.

So for a shoe player using a large spread, I calculated the playing decision value of each card weighted for the amount of money he has on the table when he has to make the decision. For a level 2 system, I found out that for maximum playing efficiency the best system tags to use are {A...10} {-1,1,2,2,2,2,1,0,-1,-2}

Look familiar? :toast:

For shoes the counts with the highest betting correlation would outperform those with high playing correlations.:joker::whip:

 

[Southpaw]

For shoes the counts with the highest betting correlation would outperform those with high playing correlations.:joker::whip:

Not necessarily true. Not to long ago, IC&T really showed me up on this. If LS is involved, a system with a high PE will outperform one that has a high BC in a 6-deck game, even when a spread as large as 1-20 is used. We were comparing HO2 (no ASC) to Hi-Lo.

SP

 

[Automatic Monkey]

For shoes the counts with the highest betting correlation would outperform those with high playing correlations.:joker::whip:

Well sure, but the system tags listed have really good betting correlation too, especially in H17.

 

[blackjack avenger]

Nobility in Complexity

Not necessarily true. Not to long ago, IC&T really showed me up on this. If LS is involved, a system with a high PE will outperform one that has a high BC in a 6-deck game, even when a spread as large as 1-20 is used. We were comparing HO2 (no ASC) to Hi-Lo.

SP

The difference in the 2 counts is because HO2 is a higher level system. It should be better, even with no ASC.

If you have a lower level system that outperforms a higher level system, then that is interesting because it's suggested in literature a higher level system is harder to use.

I am not saying that no lower level system ever outperforms a higher level system; but as groups if you compare them, the higher level counts perform better.

or

Halves rule, the others drool:joker::whip:

 

[Automatic Monkey]

The difference in the 2 counts is because HO2 is a higher level system. It should be better, even with no ASC...
...

But in a realistic situation, it is not.

 

[Renzey]

Well sure, but the system tags listed have really good betting correlation too, especially in H17.

AM, The Mentor Count whose tags you quoted, has a 97% BC and a 62% PE with no side counts -- a pretty good all-around athlete for DD and shoes, I think.

 

[zengrifter]

AM, The Mentor Count whose tags you quoted, has a 97% BC and a 62% PE with no side counts -- a pretty good all-around athlete for DD and shoes, I think.

For anyone considering a move to level-two, I consider Mentor to be the best
all-around system, with the added increased granularity of the unique 2DTC.

AutoMonk's findings also concur with Brett Harris, who found that ZEN would
out perform RPC even at ENHC shoes, which surprised him. zg

 

[blackjack avenger]

?

But in a realistic situation, it is not.

?:joker::whip:

 

[BJgenius007]

I've been doing some work on indices, as used by a counter instead of a flat better.

Turns out, splitting 10's is mighty, especially if you are counting the ace as a positive card. And for a counter, the decisions he has to make when there is a lot of money on the table mostly treat the ace as a high card- splitting 10's, doubling 8-10, doubling A8 and A9. Whether or not to hit or stand on 16 vs. 10 may be the most powerful index for a flat better but it is insignificant for a counter in a multideck game, who always knows exactly how he is going to play that hand when he has more than a minimum bet down. Thus he can forget about that index, and forget about the contributions the EoR of each card make to that index.

So for a shoe player using a large spread, I calculated the playing decision value of each card weighted for the amount of money he has on the table when he has to make the decision. For a level 2 system, I found out that for maximum playing efficiency the best system tags to use are {A...10} {-1,1,2,2,2,2,1,0,-1,-2}

I am using Omega II, the tags are {0,1,1,2,2,2,1,0,-1,-2}. The only difference is that ace is neutral 0. (To balance the count, 3's tag is 1 instead of 2 in your system.) It is because...

For betting purpose, Ace should be treated like ten.
For playing purpose, Ace should be treated like a small card.
If you can keep the ace side count, you can enjoy the best of two worlds.

Add ace side count when you bet. (Four more aces in remaining shoe increases true count by 1. You bet accordingly.)
Adjust playing decisions based on ace side count. (Shifting indexes when you have hard 16. Ace rich in remaining deck helps doubling on 9, 10, soft 18, soft 19, soft 20. Ace poor in remaining deck helps doubling on 11, other soft hands, etc. IMO the last one is the most helpful. When you know there are very few aces left and many tens left. doubling on bad soft hand like soft 13, soft 14 against dealer OK up-card make you look like a dumb ploppy not knowing your basic strategy. After a while, you will notice ace poor and ten rich make dealer so easy to bust.)