73 pages? Really? OMG!!
There are many problems in suggestions that one strategy is better than another:
- First, a common mistake (famously made by the authors of OPP and Speed Count) is that you do not adequately measure the effect of risk.
- Second (made by the authors of both), is the incorrect assumption that you must apply the same parameters to both systems, like betting strategies, ignoring that optimal betting can be dramatically different between strategies.
- Third, assuming that you can correctly compare wonging strategies.
- Most importantly, what are the exact circumstances? Different strategies have different objectives.
When someone makes a statement that A is better than B, red flags pop up in my mind.
"I don't think outside the box; I think of what I can do with the box." - Henri Matisse
Norm, very good post! I would like to know why other than the extremely rare remark from you that the Hi Low count gets the money, that you have remained mainly silent during like six years of continuously repeated counting system debates? You obviously could have at least kept these debates on an even plateau.
(1) if you pay the LS game then my prediction is the my KO system will outperfrom the HO2 w ASC by an even greater degree and most likely will beat HO2 w ASC even in teh play all situations. So be selective and play the LS game.
(2) The amount my KO system underformed the HO2 for the play all game was meniscal. Any theoretical underperformance in power would be made up in the other 4 items that I mentioned in the five criteria I used to rate a system.
(3) I was asked the reason why I believe my KO system underperformed the HO2 w ASC for the play all game. I stated two reasons.
(a) First my KO has a pivot of a true count of 4 which is great for accuracy when you have positive true counts and making a large bet. But what you gain on the positive true counts you lose on negative true counts. When you are in a play all situation and playing with negative true counts, the accuracy of my KO system true count is much less than the accuracy of the balanced HO2 true count. Your minimum bet is out but still the reduced accuracy of the negative true count for the KO system as compared to the HO2 system did take a toll on SCORE in play all game I believe. There is not much I could do about that and I would not change to a balance count for negative true count accuracy where you minimum bet is out or when back counting, is not even played at all.
(b) Perhaps more important that the decrease in accuracy at negative true count of the KO system is the fact that I gave very, very few negative indices for my KO system. I had listed all of the indices that I gave. The largest negative index that I gave to Gronborg was -2.5 for standing on hard 13 v 3 when KO + (1/2)*(AA89mTc) >= crc(-2.5). If that one play is ignored then there were just a three or four induces at -1 and all other indices that I gave to Gronbog were non-negative indices. So what I need to ask Gronbog is what did he do with indices that I did not give to him. Did you just use basic strategy on those negative plays? I could calculate negative indices for KO with AA89mTc and 5m7c but I never did because I only back count and so to me they were not important. However considering some players do play all game then it would be important to have these negative indices for these players. When Gronbog comes back I will ask him what he did with plays for indices that I did not give to him. If he just used basic strategy then that would explain why my KO system underperformed the HO2 with ASC for the play all situation,. But even with basically no strategy change for negative indices, my KO system almost tied the HO2 w ASC for the no LS play all game. If I had negative indcies you will see a change. I can easily calculate negative indices and values of k1 and k2 my KO system and give them to Gronbog to add to his simulation. Then you will probably see that my KO system will outperform the HO2 w ASC even in the play all stiautoin when I give Gronbog negative indices.
So basically my KO system was at a disadvantage to the HO2 w ASC for the play al game since the HO2 w ASC had it full set of positive and negative indices and my KO system had very, very few negative indices.
I will attach to this post (as I attached before also) the six simulations that list indices and values of k1 and k2 that I gave to Gronbog that he put into his simulation. He did all six indices but you will see very few negative indices.
So when Gronbog come back I will ask if he can add some negative indices as sim 7 and then redo the sim with these negative indices as well. My prediction is that my KO system, once these negative indices are added, will outperform the HO2 w ASC even in the play all game.
And if you look at the indices I gave to Gronbog, I left out some of the positive indices for doubles and splits that were high, such as doubling hard 8 v 4. I had only included hard 8 v 5 and hard 8 v 6 doubles. If I included some of these more rare doubles and splits, which I am sure were included in the HO2 w ASC since they used the full set of indices, then the SCORE even for the back counted game would probably increase even more over the HO2 w ASC.
And when LS is added, you will see a big jump in my KO system outperforming the HO2 w ASC.
So let me calculate and add negative indices and values of k1 and k2 for my KO system and give them to Gronbog to add to his simulation and then we will see how my KO system does for the play all game.
So here are the indices and values of k1 and k2 that Gronbog put into his simulation. You will see that almost all indices are non-negative. So I will need to add some negative indices, say KO negative indices to -5 or so for the play all game which I believe will then make a difference in the play all game sims.
KO AA89mTc 5m7c sims (1).jpg
KO AA89mTc 5m7c sims (2).jpg
Last edited by bjanalyst; 02-16-2019 at 07:51 PM.
Firstly, you have to stop skewing and rigging the performance of each system. If you say that your KO system will outperform the Hi-OPT II ASC. Secondly, if you want to skew and rig performance than I can guarantee you that Hi-OPT II ASC with LS and RSA rule will outperform your KO system. You know why? Because unbalanced counts doesn't work well with resplit aces. You want to be selective the game you should be playing is LS and RSA not just LS. Which is available in most high limit tables and you KO system doesn't outperform Hi-OPT II ASC with play all option.
Are you serious? I showed you that any unbalance count can be converted into a equivalent balance count. Balanced/unbalance makes no difference. Your statement about unbalanced counts not working for spltting Ace is as ridiculous as Norman's statement that EoR do not work with the KO count!
I used EoR to calculated the indices and values of k1 and k2 for my KO system which outperformed the HO2 w ASC for back counted game and once I add negative indices I believe it will also outperform the HO2 w ASC for the play all game. And if LS is added, my KO system will bury the HO2 w ASC. I was correct at every other prediction that I made and once simulations are done you will see that I am also correct about my LS prediction of my KO system performance .I have yet to be shown to have made any mistake (other than\ typos) in any of my calculations or any of my predictions! It should also be noted that EoR have been posted to four of five significant figures. If EoR are worthless and only give you an indication of playing strategy then why publish a rough estimate to five significant figures which does not seem like a rough estimate to me.
I am amazed by supposedly educated and talented players making such ridiculous, untrue and ludicrous statements. I will attach the exhibit of the unbalanced KO and its balanced version again which I attached before. You can make the KO count into a balance count so your argument about unbalance counts not working well with splitting Aces make no sense at all.
Here is the exhibit showing the balanced and unbalance version of the KO count. Ay unbalanced count can be made into a balance count and the fact that a count is unbalanced in no way affects the indices or values of k1 or k2 which includes splitting Aces.
KO & KO.bal.jpg
Last edited by bjanalyst; 02-16-2019 at 07:55 PM.
It is not ridiculous because I did a lot of simulation to verify that. Using unbalanced counts with true counts and without true counts and both came out to have lower SCORE that balanced counts with resplit aces rules. I don't know why but that is what the simulations show. If you turn KO count, an unbalanced count, to a balanced count it is no longer unbalanced. So my statement still holds true.
All valid points which are addressed by the SCORE methodology. To be specific, I never said that bjanalysts KO system was better. I said that it had a higher SCORE in the back-counting scenarios. Given the common loose usage of the term SCORE (much to Don's chagrin) these are all fair questions to ask.
I don't know why I am posting in this thread again. BJanalyst never listens.
Determining which combined count is best is a lot more complicated than what has the highest CC. Different counts will have different Root mean squares. So TCs are not equivalent. Doubles splits and surrenders will have different frequencies for doubling, splitting and surrendering above the index. Plus bet size for each TC in your spread must be factored in. For example you might have a double decision with one count that has an overall EV from simming flat betting of .2563 that doubles 67% of the time. And another that doubles 55% of the time and has an EV of .2560. So EV says the former is best but variance would say the latter is stronger. Then a SCORE for the playing decision for each combined count would say the latter is even stronger than the variance said. Then you factor in bet size for your bet ramp and everything will change again since the combined count is probably not your betting count.
Determining the optimal combined count is a lot more complicated than finding the count with the highest CC. The highest CC is the place to start investigating. But you need to do sims for that combined count and other combined counts that are possible for the two counts being combined using some multiple of one count added to the other. All your logic and charts are preliminary versions that represent the start of investigation, not the end of investigation.
You compare you count to Hiopt2/ASC with limited indices. People don't put a ton of effort into learning to master a high level count with side counts and decide a few hours learning all the indices isn't worth the effort. Yet you compare your count to HIopt2/ASC with way less indices than any Hiopt2/ASC user would use. Which brings up the question of why don't you use a lot more indices. You go to all this effort to have the ability to have a lot of strong indices even having indices that don't exist for other counts but you never do the necessary investigation to determine the strongest combined count for a play.
With the ease to learn and use Hiopt2/ASC compared to your count and the ease to implement it, why would anyone want to use your count even if it does outperform Hiopt2/ASC. Someone starting with Hiopt2 and doing a balanced ace side count that was chosen to strengthen betting even more than the ASC and make many plays stronger would have more gain than you got over KO, but the higher percentage gain would be added to Hiopt2/ASC. So if you improved KO by 7%, the same adjustments to Hiopt2 would get at least 7% gain over Hiopt2/ASC SCORE but it would only need one side count.
Simpler and much more powerful beats more complicated and less powerful hands down.
I never heard of such drivel. I just showed you that the unbalanced KO count can be made into a balanced count. So now you have KO.bal and a KO unbalanced count. The indices of KO.bal annd KO are the same. Makes no difference if balanced or unbalanced. You can see that the CC between the blaance and unbalanced KO count is 100%. The counts are equvialtent. There must have been some other reason why there was a problem with splitting Ace that has nothing to do with the counts being balanced or unbalanced. I just do not know what happened with these simulations of splitting Aces but something other than balanced or unbalanced must have been in play.
Maybe it has to do with accuracy of true count calculations. You need a large negative index not to split Aces, So with a large negative index the true count calculations of the balanced KO for example which has a pivot at a true count of 4 and so is 4 true count points farther from the negative index of not splitting Aces as a balanced count with a pivot at a true count of zero would be. Maybe that is what is happening. But this has to do with inaccurate true counts due to errors in estimating decks remaining. If there was no error in estimation of decks remaining, then theoretically balanced or unbalanced would not make a difference.
Bookmarks