No, You did not offend me. I like challenges and I want my calculations to be correct of course. Actually I am glad that you are challenging and questioning my results instead of blinding taking what I say at face value. I want to know the truth and if I made a mistake I would like to know.
You say EoR are "squishy". But Don Schlesinger's has EoR to six significant digits. If they were that "squishy" then why calculate to six significant digits?
I do not have access to any simulations. I was hoping you did. Plus I would like independent verification of my work. Even if I had simulation software I should not be trying to prove my own work. My work should be independently verified.
Even though I do not have actual simulations of HL using Tc = HL + AA78mTc for the six situations mentioned previously, I do have indirect simulation verification in that I used my LSL program to calculate indices for the HL and the indices generated from my LSL program exactly match the published HL indices obtained through simulations. Every index that I tested with my LSL program agreed with published indices. My LSL program gave correct results for dozens of cases and I did not find one case that disagreed with the published HL indices. Thus I am very confident that my results are correct.
Another indication that my LSL program is correct is that I calculated weighted CC (which shows playing efficiency) for various counts such as HL, KO, Hi Opt 1, Hi Opt 2 and more and as expected weighted CC of Hi Opt 2 was the highest which was higher than Hi Opt 1 which was higher than KO which was higher than HL which are the same results obtained through simulations.
So basically, I have not found a single situation where results from my LSL program disagree with published results for known counts.
Based on my calculations presented earlier that showed that the weighted CC of the HL + k*(AA78mTc) are approximately equal to the weighted CC of Hi Opt 2 with side count of Aces, my guess is that simulations would show that the HL + k*(AA78mTc) has approximately the same power as Hi Opt 2 with a side count of Aces.
I have heard that with less then one deck remaining, strange things happen and non-linearity may set in so you may be correct that there are some errors in EoR with less than one deck remaining. But I play the shoe game and the dealers all cut off alt least one deck. So for the shoe game with the cut card at one deck or more, the linearity of the EoR is a very good approximation. Also it should be noted that the insurance bet is the one blackjack bet that is totally linear and insurance is the most important playing strategy variation which is improved by using Tc = HL + AA78mTc. .
If someone could plug in the six changes I mentioned previously into a canned HL simulation program using Tc instead of HL only for those six situations that would be great. Such a sim should show a substantial improvement. The CC of Tc = HL + AA78mTc with EoR for those six situations are of the order of 20% to 30% more than the CC of HL with EoR for those situations. The differences are big and should show up in simulations.
If you know anyone who has simulation software and they can run a HL simulation with just these six changes and compare it with a HL simulation with no changes that would be great. I would like to see the results at which point we can see who is correct and if the differences you point out are significant or not. I think the huge increase in CC for these six situations will overwhelm any "errors" that you mentioned above.
Below are calculations of indices for doubling on hard 9 v 2 that I gave before. I am listing it again to address one of the points that you made and to show you clearly how I calculated indices.. For a given playing strategy situation let EoR(c) = EoR for card 'c" removed from one deck. Then there are 51 cards remaining. This is how the one deck EoR was calculated for each playing strategy situation. So what I did was calculate the slope of the LSL between EoR and the tag values of he derived count. Then I calculate AACpTCp = Average Advantage Change per True Count point as AACpTCp = (51/52)*(LSL slope). Then I calculated Index as Full Deck House Advantage (FHDA) / AACpTCp to get the infinite deck index. As the CC increases the difference in indices between different number of decks decreases and they all converge to the infinite deck indices.
But if you can provide simulations for the above six situations I mentioned that would be great. I would be very interested in the results.
Bookmarks