Analysis of Lucky Ladies

#61
The cost of systematically using an index to generously

aslan said:
Are you saying that people tend to minimize the difference between betting the LL at +7 versus +8, while the real advantage is at +8 and higher, so that APs waste their money making this wager too early not realizing the difference? If that is not it, I am confused.
Thanks for trying Aslan. The numbers are arbitrary the point is that the frequency of the count 1 less than a proper index, say+8, is almost as high as the sum of the frequencies of all the TCs +8 or above. Assuming linearity for the advantage gained around the index of TC +8, the advantage gained at TC +8 is less than what you loss by always using the +8 index prematurely at +7. The remainder of what you gave up by always using +7 takes most or all of the advantage gained at +9. An index mistake at high count not only affects your highest bets but the frequency of the incorrect play is higher than the frequency of almost all of the correct use of the index. As you use negative indices the bets are small and this same generous (using at 1 TC error to disadvantage) use of an index is a tiny fraction of the correct use of the index. Think of what the TC frequency bell curve looks like. These high TCs are in the tail of the curve. High indices must be used conservatively. Systematic mistakes on the liberal side are very costly.

This is a true count frequency chart for HILO 8 decks cut 1.5 (Blackjack Attack ed3 p217):
TC +10 or higher, 0.14% 0.14% cumulative
TC +9, 0.10% 0.24% cumulative
TC +8, 0.24% 0.48% cum
TC +7, 0.32% 0.8% cum
TC +6, 0.73% 1.53% cum
TC +5, 1.03% 2.56% cum
TC +4, 2.23% 4.79% cum
TC +3, 3.27% 8.06%cum
TC +2, 6.95% 15.01% cum
TC +1, 11.33% 26.34% cum
TC 0, 27.63% 53.97% cum
TC less than 0, 46.03% 100% cum

Notice the cumulative frequencies as compared to the frequency on the next lowest TC (line below it). They are almost equal for high TC. If we assume linearity of gain in the table above an index of +8 has a frequency of 0.48% of proper use. If you have a systematic error that has you us it at +7 the frequency you use it at disadvantage is 0.32%. If the index is +8 exactly with the assumption of linear gain you can subtract TC +8 frequency from TC +7 as equal advantage and disadvantage. That leaves a frequency of 0.16 of disadvantage you still have to make up. If we assume the TC +9 has twice the advantage you divide this frequency by 2 to have 0.08%, subtract it from 0.10%, the frequency of TC +9, thats 4/5 of the advantage at TC +9. By making this mistake the closest equivalent TC index to +7 rather than +8 (our systematic mistake) is +10. You have given away most of your advantage for this index by being too aggressive.

The mistake always using a TC index of +8 at TC +7 instead of +8 is almost the same net gain in advantage as using a TC index of TC +10. This is a very costly mistake at your highest bets that gives away most of your advantage. As you can see an error of 1 TC cost you the advantage gained for the first two TCs after the index is exceeded. Look at the other high frequency cumulative versus frequency of the next lowest TC and how far up you need to go to recoup your lost advantage of employing indices before they are profitable using the method above. That is the point I have been trying to make.
 

aslan

Well-Known Member
#62
tthree said:
Thanks for trying Aslan. The numbers are arbitrary the point is that the frequency of the count 1 less than a proper index, say+8, is almost as high as the sum of the frequencies of all the TCs +8 or above. Assuming linearity for the advantage gained around the index of TC +8, the advantage gained at TC +8 is less than what you loss by always using the +8 index prematurely at +7. The remainder of what you gave up by always using +7 takes most or all of the advantage gained at +9. An index mistake at high count not only affects your highest bets but the frequency of the incorrect play is higher than the frequency of almost all of the correct use of the index. As you use negative indices the bets are small and this same generous (using at 1 TC error to disadvantage) use of an index is a tiny fraction of the correct use of the index. Think of what the TC frequency bell curve looks like. These high TCs are in the tail of the curve. High indices must be used conservatively. Systematic mistakes on the liberal side are very costly.

This is a true count frequency chart for HILO 8 decks cut 1.5 (Blackjack Attack ed3 p217):
TC +10 or higher, 0.14% 0.14% cumulative
TC +9, 0.10% 0.24% cumulative
TC +8, 0.24% 0.48% cum
TC +7, 0.32% 0.8% cum
TC +6, 0.73% 1.53% cum
TC +5, 1.03% 2.56% cum
TC +4, 2.23% 4.79% cum
TC +3, 3.27% 8.06%cum
TC +2, 6.95% 15.01% cum
TC +1, 11.33% 26.34% cum
TC 0, 27.63% 53.97% cum
TC less than 0, 46.03% 100% cum

Notice the cumulative frequencies as compared to the frequency on the next lowest TC (line below it). They are almost equal for high TC. If we assume linearity of gain in the table above an index of +8 has a frequency of 0.48% of proper use. If you have a systematic error that has you us it at +7 the frequency you use it at disadvantage is 0.32%. If the index is +8 exactly with the assumption of linear gain you can subtract TC +8 frequency from TC +7 as equal advantage and disadvantage. That leaves a frequency of 0.16 of disadvantage you still have to make up. If we assume the TC +9 has twice the advantage you divide this frequency by 2 to have 0.08%, subtract it from 0.10%, the frequency of TC +9, thats 4/5 of the advantage at TC +9. By making this mistake the closest equivalent TC index to +7 rather than +8 (our systematic mistake) is +10. You have given away most of your advantage for this index by being too aggressive.

The mistake always using a TC index of +8 at TC +7 instead of +8 is almost the same net gain in advantage as using a TC index of TC +10. This is a very costly mistake at your highest bets that gives away most of your advantage. As you can see an error of 1 TC cost you the advantage gained for the first two TCs after the index is exceeded. Look at the other high frequency cumulative versus frequency of the next lowest TC and how far up you need to go to recoup your lost advantage of employing indices before they are profitable using the method above. That is the point I have been trying to make.
View attachment 7811
So, if +8 TC is the correct point at which to bet the LL, you would be ill-advised to try to get away with initiating this bet at +7. Is that right? I mean, the frequency curve is not my bailiwick-- I just want to know the impact of the point you are making where the rubber meets the road. :laugh:
 

Attachments

#63
aslan said:
View attachment 7811
So, if +8 TC is the correct point at which to bet the LL, you would be ill-advised to try to get away with initiating this bet at +7. Is that right? I mean, the frequency curve is not my bailiwick-- I just want to know the impact of the point you are making where the rubber meets the road. :laugh:
My point has to do with any high index. This is a gross generalization. Gain is not always linear. It is costly to use a play before it is profitable ALL THE TIME. Rounding TC estimates or deck estimates in the wrong side of advantage gives this away. Lucky Ladies advantage grows quick once the index is exceeded. As long as you don't have a systematic issue that has you applying your index too soon it should be profitable. If your estimate of the TC is generous you need to be conservative about using your indices.
 

aslan

Well-Known Member
#64
tthree said:
My point has to do with any high index. This is a gross generalization. Gain is not always linear. It is costly to use a play before it is profitable ALL THE TIME. Rounding TC estimates or deck estimates in the wrong side of advantage gives this away. Lucky Ladies advantage grows quick once the index is exceeded. As long as you don't have a systematic issue that has you applying your index too soon it should be profitable. If your estimate of the TC is generous you need to be conservative about using your indices.
I think that's what I said. You're too brainy for me. The long and the short of it is not to bet your money too soon or you will squander your advantage. IE, don[t bet too soon, whether it is knowingly or due to being too loose with your TC estimates. Being on the money is critical when it comes to these kinds of bets.
 
#65
tthree said:
My point has to do with any high index. This is a gross generalization. Gain is not always linear. It is costly to use a play before it is profitable ALL THE TIME. Rounding TC estimates or deck estimates in the wrong side of advantage gives this away. Lucky Ladies advantage grows quick once the index is exceeded. As long as you don't have a systematic issue that has you applying your index too soon it should be profitable. If your estimate of the TC is generous you need to be conservative about using your indices.
Whoa wait a minute- no play is profitable "all the time," only on the hands we win! Your simulator takes care of rounding and deck estimation resolution issues, and if the average return is positive, your EV for that hand is positive providing your simulator is playing the game the same way you are. If you're concerned about accuracy, use higher resolution or a different TC estimating procedure, and if you are concerned about variance :cat: make your strike point higher.

We see the same thing in blackjack proper- there are plenty of games that are +EV with a High-Low TC of +1, but there will be situations where despite the TC being +1, if you were to do a combinatorial analysis on the cards actually remaining in the shoe it would be -EV. You can use a level 5 multiparameter count to intercept many of those traps, or you can just accept it as well as the fact that a TC of +1 will be more profitable than you expected. In fact about 1 out of 21 times your advantage will be 150%!
 
#66
This is getting frustrating so I will not try to explain any more after this. Either I am not good at explaining what I am saying or it is going over peoples heads. By all the time I am talking about a systematic error that has you applying an index to early. It could be your deck estimation or TC conversion process or an over aggressive mentality toward using indices or something else. There is about the same gain or lose on either side of the index 1 TC off from the index. The frequency of each these TC ranges is very different unless the index is zero. The TC range closer to zero occurs much more frequently than the other one so it costs that much more.
 

rrwoods

Well-Known Member
#67
Automatic Monkey said:
Whoa wait a minute- no play is profitable "all the time," only on the hands we win!
I don't think that's what he's saying.

Say you have a bet/play/whatever that is +EV at TC +7.1. If you chop this to +7 for simplicity, because of the sharp slope of the TC distribution curve, you are giving back way more EV than you think, and if you're rounding at 6.5 you are probably giving all of it back and then some.

(I don't actually know the exact strike point for LL9, but I think it's between 6.5 and 6.8 somewhere. I play it at floored TC +7.)
 
#68
rrwoods said:
I don't think that's what he's saying.

Say you have a bet/play/whatever that is +EV at TC +7.1. If you chop this to +7 for simplicity, because of the sharp slope of the TC distribution curve, you are giving back way more EV than you think, and if you're rounding at 6.5 you are probably giving all of it back and then some.

(I don't actually know the exact strike point for LL9, but I think it's between 6.5 and 6.8 somewhere. I play it at floored TC +7.)
What I'm saying is your simulator has already taken care of that. If you have a positive number in from of TC +7 and you are calculating the TC the same way the software is and it's +7, you have an advantage. Bet.

Also, you do not give back very much by betting a little too soon. Certainly more than I would want to give back, but it takes several TC worth of bad betting to make it -EV.
 
#69
I've seen various Hi-Lo TC trigger points for LL10 for 8D and 6D (1,000 QQh+BJ; 200 QQh, 25 suited match; 10 suited; 4 all other 20s). What is the consensus here for the trigger? +5?
 
#75
bjplayer123 said:
So the advantage is at tc of +7. Thats very hard to happen though
That is for reduced pay LL.
For full pay LL it would be +5 HiLo.

A more powerful method is to key-card blocks of 2+ 10s.
 
Last edited:
#76
xengrifter said:
That is for reduced pay LL.
For full pay LL it would be +5 HiLo.

A more powerful method is to key-card clocks of 2+ 10s.
Please explain "key-card clocks". I understand the importance of key cards, but not "key-card clocks".

Thank you for indulging my unfamiliarity with that term usage.
 
#77
Spyros Acebos said:
Please explain "key-card clocks". I understand the importance of key cards, but not "key-card clocks".
Oops, it was a typo - meant CLUMPS ... two or more contiguous tens.
key-carding them with reasonable accuracy can be stronger than the 10-count.
 

Toxic

Well-Known Member
#78
What advantage can be gained from QH tracking? I don't know much about ace tracking, but I was flashed the QH during shuffle and successfully predicted which hand it would be dealt. I happened to get it, that was luck I'm sure. It was a low count so I didn't play the bet
 
#79
Toxic said:
What advantage can be gained from QH tracking? I don't know much about ace tracking, but I was flashed the QH during shuffle and successfully predicted which hand it would be dealt. I happened to get it, that was luck I'm sure. It was a low count so I didn't play the bet
At a neutral count, knowing that your first card will be ANY 10, you have almost a 100% advantage on LL...
... knowing that it will be the Qh gives nearly a 400% adv.
 
Top