Bump.
Does anyone have any ideas I can try to get this side bet simulation to work? Is there any information I could provide that would help? I didn’t show the strategy that I attached, but that is simple enough I don’t think I messed that up. As far as I can tell the attached strategy just needs to have the system tags, but I attached a full strategy with all the indexes too. This is partly to double check my LL numbers, but also to generically have confidence if I do other side bet sims.
Sent from my iPhone using Tapatalk
Install the latest update. If that doesn't fix it, go to Tools-Export, create an export file and email it.
https://qfit.com/downloads.htm#free
"I don't think outside the box; I think of what I can do with the box." - Henri Matisse
Alright, I think I figured it out. It has to do with the "Side Bets are Merged" button at the bottom, and the lower-left sector of the screen named Side Bet. In my first side bet tab (first pic in the original post), the bet is set to 1.0. In the next 4 side bet tabs, the "Side Bet" window is blocked out because the bets are merged. What I did was un-check "Side Bets are Merged", and I noticed that side bet tabs 2 through 5 also had the bet set to 1. I thought that with side bets merged, only the bet on tab 1 would count, so I paid no attention to these other hidden bets. But with the side bets un-merged, I changed the bets in tabs 2 through 5 from 1.0 to 0, then re-merged the side bets and that's when I finally got normal looking EV numbers. According to this sim, doubled halves tags in 6D with the LL9 payout has an index of about 13. So when the sim wasn't working, it was because I was betting 5 times the side bet amount with only the pay table of 1 bet.
LL fix.jpg
6 decks (1 deck cut off), Hi-Lo, Truncating TC, Truncating deck estimate, quarter deck resolution, TC divisor is cards in tray, 2 billion rounds, LL10 pay table.
The EV is slightly positive at +4, but really kicks in at +5 & above. The EV changes very rapidly with the count.
Hi-Lo LL10 6D.jpg
It seems something is not right. For a 6-deck shoe, EV should be -17.6%. At TC +4, EV should be slight negative, I believe. For an 8-deck shoe, EV should be slightly positive.
If we use ASC, the critical TC should drop one when aces per deck drop one. Can you double check?
Look how quickly the EV changes with the count. Yes, Shackleford's analysis shows a -17.6 EV off the top for this pay table. The sim I show above at TC 0 shows -18.0%. Don't you recognize those as the same? TC = zero isn't a single point off the top of the deck, rather it is a whole bunch of simulated situations over a range of TC from -0.5 to +0.5. The EV changes by 6% for each 1 TC, and the small difference you pointed out is only 0.4%. The sim above would show an EV of -17.6 above at a TC of +0.07. What is the TC off the top with a single low card as the burn card? TC +0.17. So the difference you are focused on is a fraction of the difference of a single card.
Then look at TC +4. It's close to the break even point. Simulations depend on the simulated conditions, which I noted. If I changed things at all, like rounding the deck estimate instead of truncating, it is going to change the exact results around a break-even point. The simple idea is that the EV changes very quickly with the count. If you make this bet at +4, you might have a negative expectation if your deck estimation is off or if you determine true count differently than the simulated results. You are better off waiting until +5 to bet this one, to be sure about it.
Wonderful work! This is convincing. The rapid growth of EV with TC is easily understood. The density of 10s is roughly proportional to the TC, so the probability of two-10 hands is proportional to the square of TC. It quadratically increases. Is it possible to do an ASC on top of this?
Don’t hold me to this, because I side count differently, but I think you could make a simple adjustment. Let’s say you’ve played four of six decks, and you’ve seen 20 aces played when you expected to see 16. You would want to count those four extra played aces as a non-ten valued card, so you would multiply that number of aces by 2 and add 8 to the running count. You would probably use the same index. More aces played means more tens remain, increasing the EV.
Sent from my iPhone using Tapatalk
I have done some investigation on ASC on LL10. The result is simple. For each deficit of aces per deck, the critical HiLo TC drops one. For example, if the critical TC for a 6-deck shoe is +4, then it becomes +3 when ASC shows one deficit of aces in the remaining deck. I just need somebody to double check. Thank you for your nice work.
Bookmarks