Hi-lo/Ten count

[Three]

143-2288? 153-2448? You will not be using these spreads in any casino; they are completely unrealistic. If a reasonable 150-2400 spread is employed Ill18 yields 635.28 and full indices gives 681.65, a difference of 46.36.

The point was not to suggest a bet spread for use in the casino but to keep all rounding errors out of the comparison of optimal 16 to 1 spreads at the same BR and RoR for each PE with the same count. The way to be exact is to not tinker with the optimal spread suggested by CVCX for both scenarios and the parameters set. Why in the world would you want to introduce rounding errors that render results inaccurate? When you set a spread for casino use then you can round and cause suboptimal but practical results.

It could take millions of hands to realize that difference in EV in real life.

The results would be $50/100 hands more for however many hands it would take for the long run to occur. Just like in BJ it will take lots of hands for your results to fall inline with expectation. A million hands will get your results close to expectation for a counter. That is the point you will be $500K behind with the lower PE. That is 10000 hours at 100 rounds/hour. We are talking 10 to 30 years of play at 1000 hours/year (19 hours a week) to 334 hours/year (6.5 hours/week). To me a half million dollars is a pretty significant cut in pay to save some PE. If you are a red chipper at a $15 table it is only about $50K for those 10000 hours at 100 rounds/hour. That's about $5K a year at 1000 hour/year and 100 rounds/hour red chipping.

They like to say a Score of 50 is the minimal playable game you should enter. With the boost in PE the unplayable SCORE of 45 is now a playable SCORE of 50.1. So besides giving you a nice increase in pay it makes more games playable which is another boost to your expectation by making more games playable. It is about the big picture not seeing only the smallest aspect of the value of PE.

[mofungoo]

The point was not to suggest a bet spread for use in the casino but to keep all rounding errors out of the comparison of optimal 16 to 1 spreads at the same BR and RoR for each PE with the same count. The way to be exact is to not tinker with the optimal spread suggested by CVCX for both scenarios and the parameters set. Why in the world would you want to introduce rounding errors that render results inaccurate? When you set a spread for casino use then you can round and cause suboptimal but practical results.

Why not use the spread that will be used at the casino? The purpose of running a sim is to emulate reality as closely as possible so that one may have a better idea what to expect in real life.

The results would be $50/100 hands more for however many hands it would take for the long run to occur. Just like in BJ it will take lots of hands for your results to fall inline with expectation. A million hands will get your results close to expectation for a counter. That is the point you will be $500K behind with the lower PE. That is 10000 hours at 100 rounds/hour. We are talking 10 to 30 years of play at 1000 hours/year (19 hours a week) to 334 hours/year (6.5 hours/week). To me a half million dollars is a pretty significant cut in pay to save some PE. If you are a red chipper at a $15 table it is only about $50K for those 10000 hours at 100 rounds/hour. That's about $5K a year at 1000 hour/year and 100 rounds/hour red chipping.

Once again, your numbers are not realistic. 1000 hours a year? Few amongst us will reach that number of hours in a year. Furthermore, even if you could play that many hours in a year, you would not be able to sustain that pace at the betting levels you specified without getting backed off, or more likely, trespassed. A counter making $2000+ bets can hurt a casino and they will not tolerate that sort of action for long.

Frankly, I would form or join a team if I really wanted to bet at high levels and clean out a casino's bankroll. If it were my team I think I would use the MIT approach: use Hilo and play basic strategy except for the insurance index. Sure, there would be money left on the table, but there would be much more money leaving the casino in our pockets.

[Three]

Why not use the spread that will be used at the casino? The purpose of running a sim is to emulate reality as closely as possible so that one may have a better idea what to expect in real life.

the purpose of the sim was to show the gain from a PE increase with the highest accuracy. Rounding error for practical bets would vary by the situation and be meaningless error. I am guessing you never worked in the math sciences. You would lose your job for introducing unnecessary error into a study. If you want to know the exact value of a PE increase you don't add in error to the process that would vary in all directions in the real world application. You eliminate that noise to get an accurate answer.

Once again, your numbers are not realistic. 1000 hours a year? Few amongst us will reach that number of hours in a year. Furthermore, even if you could play that many hours in a year, you would not be able to sustain that pace at the betting levels you specified without getting backed off, or more likely, trespassed. A counter making $2000+ bets can hurt a casino and they will not tolerate that sort of action for long.

Again the large betting levels are to allow for accurate estimate of what PE is worth. If you scaled everything down you would introduce rounding error that could all stack up in one direction. The increase in accuracy for the answer is the equivalent of why you don't use a 20,000 round sim. The error grows enough to make the results unreliable. To get the highest degree of accuracy you use billions of rounds and in this case you use a very large BR and whatever the calculated optimal bets are. Every deviation from that introduces more error to the answer of what the PE is worth. If you want an accurate answer you don't introduce error.

[Three]

Frankly, I would form or join a team if I really wanted to bet at high levels and clean out a casino's bankroll. If it were my team I think I would use the MIT approach: use Hilo and play basic strategy except for the insurance index. Sure, there would be money left on the table, but there would be much more money leaving the casino in our pockets.

the discussion was about the value of PE. The MIT team trained disposable people quick and burned them. That made the backers the most money. But the players had a very short playing career. The players put in the fray so quickly would take longer to train for index plays. The way they threw them out so quickly index plays might hurt their results because of the misuse of the plays. They accuracy of the players use of the plays are in question. Index plays can cost you more than they gain if you employ them before the play is warranted by the count. Griffin wrote of this and there is a condition that has been called new counter syndrome where a new counter costs himself money by trying to impress himself with his new knowledge by the use of playing deviations too often (when they are incorrect plays). The nature of the bell curve will have the frequency of use early be higher than the same span of TC on the correct side of the play. For example if the index was exactly +2 and you used it at +1 all the time either through inaccurate estimates of TC or to show off your new knowledge to yourself the frequency of misuse (all of TC 1) where you use the index as a negative EV is 7.48% in a 6 deck 1 deck cut off game. The frequency of all TCs that warrant the play combined (TC >= +2) is 10.67%.

[mofungoo]

the purpose of the sim was to show the gain from a PE increase with the highest accuracy. Rounding error for practical bets would vary by the situation and be meaningless error. I am guessing you never worked in the math sciences. You would lose your job for introducing unnecessary error into a study. If you want to know the exact value of a PE increase you don't add in error to the process that would vary in all directions in the real world application. You eliminate that noise to get an accurate answer.

You must be a bean counter, right? Bean counters are famous in my world (design engineering) for only looking at numbers without understanding what they mean. Why not run a sim based on real world situations instead of setting up phony sims to get big numbers? Well, old bean, if you had set the sim for a 5% ROR and 200 rounds per hour (your original setup had you playing heads up, right?) with bankroll and game conditions the same, you would have gotten bigger EV numbers, and those numbers would be closer to the results you would get during actual play.

Again the large betting levels are to allow for accurate estimate of what PE is worth. If you scaled everything down you would introduce rounding error that could all stack up in one direction. The increase in accuracy for the answer is the equivalent of why you don't use a 20,000 round sim. The error grows enough to make the results unreliable. To get the highest degree of accuracy you use billions of rounds and in this case you use a very large BR and whatever the calculated optimal bets are. Every deviation from that introduces more error to the answer of what the PE is worth. If you want an accurate answer you don't introduce error.

This is about the most ridiculous thing I have ever read. If you want an accurate answer, you must set up the sim so that the sim is relevant to the real world situation you are simming. Otherwise the results you get are not accurate in the real world context. For me, the math is what I use to develop a winning strategy. It is not an end unto itself. For me, the only thing that matters is that I beat the casino. This is not about beating you at all; if you make more money than I do, that's perfectly fine. I can live with that.

[BlackHead]

[QUOTE=Tthree;136978]Why not. PE just gets you more money at no cost. Only a fool would not want a raise for doing the same thing.



Would you recommend optimizing my team play action in the long run by going Zen/Ten count to improve my PE. My local casino rules are all 8 deck,6 deck, H17 no surrender, %0.69 house advantage, but good penetration. I remember reading somewhere that at TC1 the Zen Count has a %0.70 advantage so I would imagine this would be a good step in having a chance at beating this tuff game. I ran some simulations and noticed my win rate was somewhere around 4 dollars an hour but around 12 dollars an hour with the Zen count without my partner using a ten count added.

[Three]

You must be a bean counter, right? Bean counters are famous in my world (design engineering) for only looking at numbers without understanding what they mean. Why not run a sim based on real world situations instead of setting up phony sims to get big numbers?

If I were to do I you say I could cherry pick the rounding to get bigger numbers. The way I did it is the most accurate way. I am no bean counter. I am the one always pointing out how the unintended consequences of bean counter suggestions will cost hundreds to thousands of times more than the peanuts they save. It is about seeing the big picture which has been my point all the time. The bean counter would say the increase from the PE isn't worth it because you only get 3.9%. The big picture allows you to bet more to reap an 11.1% gain and makes some unplayable games playable while getting you to the long run in 87% of the time it would take otherwise. That is seeing the big picture that the bean counter misses.

Well, old bean, if you had set the sim for a 5% ROR and 200 rounds per hour (your original setup had you playing heads up, right?) with bankroll and game conditions the same, you would have gotten bigger EV numbers, and those numbers would be closer to the results you would get during actual play.

Well I was keeping everything optimal and 13.5% RoR gets you optimal BR growth for both situations. Using what you call real world numbers and bets 200 rounds/hour, 106K BR spread optimal 16 to 1 spread same 6 deck game as before:

HILO I18 and Fab4 heads up:
EV $821.22, spread 100 to 1600, RoR 5%, N0 8680

HILO full indices heads up:
EV $929.63, spread 110 to 1760, RoR 5.3% (I had to round the optimal bet from 108 to 110 by tweaking the RoR a hair but he wanted practical error introduced), N0 7796

Now we get a gain from PE of 13.2% along with all the other benefits. The 200 rounds an hour shouldn't make any difference for the percentages but that tiny tweak introduced almost 20% error for what the increase in PE is worth. I hope that answers why you don't include it in the study to determine the value of PE. The rounding could have had the error be as high as 40% and it could have been in either direction. That only becomes relevant if you are looking at a specific situation not the general gain from the PE increase. So basically the gain in PE for practical applications to HILO is 6 to 14%. If you used a much smaller BR the gain would have a much wider range of error since the rounding error is much more significant compared to the bet size.

This is about the most ridiculous thing I have ever read. If you want an accurate answer, you must set up the sim so that the sim is relevant to the real world situation you are simming. Otherwise the results you get are not accurate in the real world context. For me, the math is what I use to develop a winning strategy. It is not an end unto itself. For me, the only thing that matters is that I beat the casino. This is not about beating you at all; if you make more money than I do, that's perfectly fine. I can live with that.

So the accurate answer is that the gain is anywhere from 5 to 20% depending on how all the rounding errors affect the specific situation. Now isn't that a useful answer when speaking generally. I already knew the affect of rounding error so the noise it introduces is not a useful metric when trying to quantify a specific average number.

This is not about beating you at all; if you make more money than I do, that's perfectly fine. I can live with that.

I agree BJ is not a competition between APs but I don't want to leave money on the table that becomes effortless to pick up with enough practice. You spend 12 years in school and then 4 to 10 in college to be ready for work. What is an extra few hours learning more indices and the time to make the TC calculation effortless. You need to do the latter anyway. I don't even do math to get the true count. I know the bin ranges for all levels of pen to be each TC. So the math isn't neccesary at the table. If there are 3.5 decks yet to be played and the RC is +23 I know the TC is near the middle of the TC +6 bin (range of TC +6 bin at 3.5 decks remaining is RC +21 to +24) without doing any math.

[BlackHead]

Bump

Anyone have any ideas for optimum same table team play vs 8-6 deck H17?


Sent from my iPhone using Tapatalk

[mofungoo]

HILO I18 and Fab4 heads up:
EV $821.22, spread 100 to 1600, RoR 5%, N0 8680

HILO full indices heads up:
EV $929.63, spread 110 to 1760, RoR 5.3% (I had to round the optimal bet from 108 to 110 by tweaking the RoR a hair but he wanted practical error introduced), N0 7796

Now we get a gain from PE of 13.2% along with all the other benefits. The 200 rounds an hour shouldn't make any difference for the percentages but that tiny tweak introduced almost 20% error for what the increase in PE is worth. I hope that answers why you don't include it in the study to determine the value of PE. The rounding could have had the error be as high as 40% and it could have been in either direction. That only becomes relevant if you are looking at a specific situation not the general gain from the PE increase. So basically the gain in PE for practical applications to HILO is 6 to 14%. If you used a much smaller BR the gain would have a much wider range of error since the rounding error is much more significant compared to the bet size.

Well, the first thing you need to do is to understand the meaning of PE. PE is a measure of the efficiency, or accuracy, of a count's tags regarding card play strategies. PE has nothing to do with the number of indices used at all. Hilo using the Ill18 has the same PE as Hilo using all indices.

So all your sim work attempting to asses PE is invalid. Your use of odd bet amounts to "reduce noise" is plain silly. You can't use precisely optimal bets in a casino for long. Wasn't it you who cautioned against "barber pole bets", saying how the dealers would get irritated having to break down the bets to pay them? Besides that, doing this would slow the game and reduce hands per hour.

If you want to explore the effect of reduced indices on results, why not use the standards outlined in the SCORE method? Your sim setup is close, but SCOREs are not determined by heads up sims. There are specific numbers of players for different numbers of decks used.

[moo321]

Bump

Anyone have any ideas for optimum same table team play vs 8-6 deck H17?


Sent from my iPhone using Tapatalk

Parameters? You could do opposition betting to disguise/increase spread if you have a massive bankroll relative to the table. One could keep insurance count. One could keep a side-bet count.

[BlackHead]

Parameters? You could do opposition betting to disguise/increase spread if you have a massive bankroll relative to the table. One could keep insurance count. One could keep a side-bet count.

How would I go about using parameters in conjunction in aiding someone else using Hi-lo or any other count. Also, my local casino has a match the dealer bet which %75 percent of the people are betting on it. I haven't found much online about a side-count to beat the match the dealer bet is it possible.


Sent from my iPhone using Tapatalk

[Three]

Well, the first thing you need to do is to understand the meaning of PE. PE is a measure of the efficiency, or accuracy, of a count's tags regarding card play strategies.

Well if you don't understand the purpose of the sim then I don't know what to say. Griffin defined PE for everyone using 70 playing deviations but also included insurance decisions with them. Today just the 70 playing deviations are used with a separate metric for insurance correlation. Now I suppose you could say the HILO player using basic strategy for decisions and the HILO player playing !18 and Fab4 and another using full indices all have a PE of .51 but we know that statement has no real would application. HILO has the potential to be played to the PE of .51 but obviously the BS player using HILO doesn't have a PE in the actual casino play of .51 nor does the I18 and Fab4 player. You need to use all the 70 deviations to be playing with HILO's full potential, a PE of .51. I think arguing that someone using HILO and basic strategy for playing decisions has a PE of .51 indicates a lack of understanding of the application of PE to real world play. I have to assume either I have a couple idiots posing what ifs, which I don't believe, or I am being trolled. Demanding real world applications to fuzzy up the generalization of what higher accuracy of playing decisions is worth and then acting like a player playing BS has a PE of .51 because he uses the HILO count is being very inconsistent mentally.

This is pointless. I have made the point to those smart enough to understand. What is left are trolls and people that aren't smart enough.

[mofungoo]

PE and EV are different concepts. PE is the same for a given count regardless of the number of indices used. Using a different number of indices will affect the EV, not the PE. If you believe differently, please use your math skills to compute what you think the PE would be for the Ill18.

BTW, the only time I mentioned Basic Strategy was in regard to forming a team. That would for training and camouflage purposes. Money would be left on the table, of course, relative to using more indices. There are more things to consider with a team approach than simply maximizing EV.