I ran into this recently (player's first two cards pays: pair 3:1; BJ 6:1; suited pair 10:1; suited BJ 15:1; Push dealer BJ 25:1). Saw a sim evaluation elsewhere for HiLo indicating 11.6% house advantage off the top and a player advantage in 6D (H17) of 3% at HiLo TC+5 and increasing an additional 3.5% for each TC >+5. My very limited experience playing it at KO +3 to +6 was very nice. If I'd have had an idea of the advantage (saw the above analysis later), I would have been putting more out.
High Tie Bonus bet
- Thread starter Diver
- Start date
Automatic Monkey
Banned
I saw that too in CBJN, and that pay table can't possibly be right. I get a house edge over 30% with those rules.
The Wizard Of Odds lists a couple of more realistic pay tables:
http://wizardofodds.com/blackjack/appendix8.html#hightie
The Wizard Of Odds lists a couple of more realistic pay tables:
http://wizardofodds.com/blackjack/appendix8.html#hightie
The pay table is right
At least for the game I played recently. I haven't run a sim so can't comment on that. My own experience in playing it at a positive KO count (+3 & up) was tainted by ridiculously good results that would only have been exceed by a BJ match with the dealer.
Automatic Monkey said:
At least for the game I played recently. I haven't run a sim so can't comment on that. My own experience in playing it at a positive KO count (+3 & up) was tainted by ridiculously good results that would only have been exceed by a BJ match with the dealer.
The HA for that paytable is 11.6% like u said in the first post.Diver said:
As to the value of it at higher TC's, I have no idea. Were you really saying that the player's advantage increases by 3.5% for each TC above +5, assuming you made the side bet?
Pretty wild if so.
That's how it was described at a noted BJ newsletter this week, citing the increased prevalence of BJ and pairs at higher counts. It does open some interesting possibilities, doesn't it? Might even offset the clobbering I'm taking now on CV when my paired faces are getting drubbed by dealer BJ and draws to 21.Kasi said:
Well it was the increased chances of BJ at higher counts that made me wonder and even dream it might be possible to overcome a 11.6% HA. But I think I've seen BJ's going up to 7% or more at higher counts. Frequency of pairs at higher counts I don't know much about except of course there are alot more pushes at 20.Diver said:
Interesting possibilities? - are u kidding me? ! lol.
EmeraldCityBJ
Well-Known Member
I'm the one who ran the sim and reported the info to CBJN. The paytable I observed at this particular casino is different from any of the ones published at Wizard of Odds, or anywhere else I could find information about the game. However, you can use the probabilities posted by Wizard of Odds to arrive at the 11.6% reported by my sim. The casino's paytable is the same as "Version 1" on the Wizard of Odds page, except the top payout for a BJ push is 25:1 rather than 50:1. This means the return for the top payout is 5.4% rather than 10.8%. This increases Wizard's HA calc of 6.2% by another 5.4%, which matches the 11.6% I observed by running the sim.Automatic Monkey said:
I am reasonably confident in my sim and the other numbers I reported, but would welcome and encourage anyone to run a similar sim or to do further analysis to prove/disprove my results. Using HiLo, my sim suggests that the edge is 0.00% at TC=+4, and the player has a 3.2% edge at TC=+5. At TC=+6, the player edge was somewhere around 6.7-6.8% (I don't have the detailed results handy at the moment, so this is from memory).
EmeraldCityBJ
Well-Known Member
Certainly, the increased probability of a BJ is what makes this bet interesting during high counts. The probability of a pair also increases marginally with high counts (and also very negative counts) since you're more likely to pair up two cards if the remaining pack is defficient in some ranks and has a surplus of other ranks. The same effect impacts the Match the Dealer side bet, but the count has to get pretty extreme before the player has an advantage.Kasi said:
Dog Hand
Well-Known Member
My CVData Sim Results are Quite Different...
Hmm... I ran a billion-hand sim and got a house edge of 9.992% with a standard error of 0.03. Also, my player didn't have an advantage on the side bet until HiLo +6, at which his edge was only 0.078%. His edge did grow rapidly with TC: 2.9% at +7, 5.8% at +8, 8.9% at +9, and double-digit edges for +10 & above. For reference, I was simming a 6D, S17, DA2, DAS game with 75% pen for a heads-up player.
I'm puzzled... your HA results are in good agreement with the WoO's calcs. Can you describe your sim params?
Dog Hand
ECBJ,EmeraldCityBJ said:
Hmm... I ran a billion-hand sim and got a house edge of 9.992% with a standard error of 0.03. Also, my player didn't have an advantage on the side bet until HiLo +6, at which his edge was only 0.078%. His edge did grow rapidly with TC: 2.9% at +7, 5.8% at +8, 8.9% at +9, and double-digit edges for +10 & above. For reference, I was simming a 6D, S17, DA2, DAS game with 75% pen for a heads-up player.
I'm puzzled... your HA results are in good agreement with the WoO's calcs. Can you describe your sim params?
Dog Hand
Automatic Monkey
Banned
If that's the case, then I probably screwed up the sim. I'll try it again.EmeraldCityBJ said:
Automatic Monkey
Banned
All right, I modified one of the canned sidebet sims in CVData and I get something much like what Emerald City BJ gets:Automatic Monkey said:
Code:
TC Side Bet 1
EV
TOT -11.63%
18 62.77%
17 68.72%
16 36.71%
15 29.69%
14 39.18%
13 33.63%
12 29.40%
11 24.06%
10 20.29%
9 16.97%
8 12.61%
7 9.12%
6 5.67%
5 2.14%
4 -0.73%
3 -4.07%
2 -6.84%
1 -9.33%
0 -11.73%
-1 -14.23%
-2 -16.73%
-3 -19.16%
-4 -21.24%
-5 -23.51%
-6 -25.41%
-7 -27.28%
-8 -29.18%
-9 -30.98%
-10 -31.54%
-11 -32.92%
-12 -35.57%
-13 -36.02%
-14 -37.33%
-15 -36.94%
-16 -32.36%
-17 -40.15%
-18 -44.11%For a KO player, the advantage also starts to kick in at a RC of +5:
Code:
Side Bet 1
EV
TOT -11.61%
29 56.14%
28 37.35%
27 34.94%
26 39.34%
25 42.53%
24 37.15%
23 37.76%
22 29.71%
21 27.50%
20 27.09%
19 24.96%
18 20.47%
17 21.31%
16 20.96%
15 16.69%
14 15.05%
13 13.30%
12 11.62%
11 9.83%
10 7.89%
9 6.31%
8 5.38%
7 3.25%
6 1.92%
5 0.21%
4 -1.37%
3 -2.58%
2 -3.83%
1 -5.02%
0 -6.32%
-1 -7.09%
-2 -8.23%
-3 -9.17%
-4 -10.12%
-5 -10.75%
-6 -11.37%
-7 -12.08%
-8 -12.51%
-9 -12.84%
-10 -13.18%
-11 -13.52%
-12 -13.63%
-13 -13.94%
-14 -14.19%
-15 -14.11%
-16 -14.30%
-17 -14.42%
-18 -14.60%
-19 -13.85%
-20 -14.67%
-21 -14.44%
-22 -15.62%
-23 -16.49%
-24 -16.94%
-25 -17.83%
-26 -18.37%
-27 -19.32%
-28 -19.79%
-29 -20.44%
-30 -20.87%
-31 -21.30%
-32 -22.42%
-33 -22.03%
-34 -23.75%
-35 -23.56%
-36 -24.34%
-37 -24.63%
-38 -25.53%
-39 -26.24%
<-39 -26.86%EmeraldCityBJ
Well-Known Member
I actually ran my own program rather than CVData. There are some differences between my sim and your sim with CVData:Dog Hand said:
1) My sim draws four cards from a 6D shoe and evaluates the side bet. It does not actually play the BJ hand.
2) My Pen was set to 5.0/6.0 (a bit better than the 75% you used).
3) The game rules for the game in question are H17, but that was not a factor given the way I did my sim.
4) My sim was also for 1 billion hands, so I expect my standard error to be comparable to yours.
I don't think any of these items would make a significant difference in the calculated HA, although they might cause me some error in determining the frequency of certain counts.
My best guess is that one of the payouts in your CVData sim is set too low. The difference between your sim and the WoO house edge is 1.61%. This is approximately the probability of catching a suited pair. I believe that if the suited pair paid 9:1 rather than 10:1, your sim would be very close to the WoO calcs. This is the first payout I'd review.
Another payout to look at is the BJ Tie. If I recall CVData correctly, its payouts are additive, meaning that the BJ Tie needs to pay 10:1 if the player's BJ is suited, and needs to pay 19:1 if the player's BJ is not suited to get the correct total payout of 25:1.
EmeraldCityBJ
Well-Known Member
My KO Numbers
I ran a separate sim for KO, and got numbers which are close to what Automatic Monkey posted.
My sim shows a 0.2% edge at IRC+25, and a 2.2% edge at IRC+26 (or, +5, and +6 respectively if your IRC is -20). What I found interesting is how the HA was remained relatively constant in the 13-14% range during many lower counts (anything from IRC until about IRC+13), and it's encouraging to see another dataset showing the same trend.
While I'm glad that we're coming up with the same count to start making this bet, I'm bothered that my numbers are consistently higher than the CVData results. Given the calculated overall house edge, both sets of data seem plausible. It would be nice to have a third set of data for comparison. At this point, I think we know the count at which the bet becomes playable, but there is still some doubt regarding how much of an advantage we have at such counts.
I ran a separate sim for KO, and got numbers which are close to what Automatic Monkey posted.
My sim shows a 0.2% edge at IRC+25, and a 2.2% edge at IRC+26 (or, +5, and +6 respectively if your IRC is -20). What I found interesting is how the HA was remained relatively constant in the 13-14% range during many lower counts (anything from IRC until about IRC+13), and it's encouraging to see another dataset showing the same trend.
While I'm glad that we're coming up with the same count to start making this bet, I'm bothered that my numbers are consistently higher than the CVData results. Given the calculated overall house edge, both sets of data seem plausible. It would be nice to have a third set of data for comparison. At this point, I think we know the count at which the bet becomes playable, but there is still some doubt regarding how much of an advantage we have at such counts.
Automatic Monkey
Banned
What you're seeing with the KO sim is a characteristic of unbalanced counts that they don't give very much information on the low counts, just with the high ones. A count at the IRC represents a normal situation at the beginning of a shoe, but a terrible count at the end of it. There's a similar disparity at high counts but not nearly as large.EmeraldCityBJ said:
Just for the heck of it, I created a count just for playing this side bet. Aces are -4, tens are -1, and all other cards are +1. I left the Units Bet column to give the information about count frequency:
Code:
Side Bet 1 Side Bet 1
Units Bet EV
TOT 249,999,982 -11.64%
26 4,014 71.95%
25 6,682 67.97%
24 9,267 59.33%
23 13,613 57.13%
22 21,877 54.25%
21 32,669 53.63%
20 46,776 45.27%
19 66,357 40.42%
18 102,637 41.08%
17 132,778 37.62%
16 200,217 33.22%
15 280,792 29.19%
14 394,091 26.19%
13 549,718 22.94%
12 762,051 19.15%
11 1,066,271 16.67%
10 1,466,694 13.15%
9 2,002,321 10.67%
8 2,777,050 8.10%
7 3,916,350 5.26%
6 5,233,547 2.51%
5 7,351,637 -0.27%
4 10,430,012 -2.68%
3 14,708,349 -5.05%
2 20,957,902 -7.38%
1 31,523,238 -9.61%
0 40,563,766 -11.84%
-1 30,772,115 -14.14%
-2 20,943,565 -16.47%
-3 15,020,566 -18.71%
-4 10,851,810 -20.91%
-5 7,756,410 -23.02%
-6 5,784,261 -25.12%
-7 3,998,103 -27.09%
-8 2,975,154 -28.88%
-9 2,138,197 -30.86%
-10 1,552,265 -32.20%
-11 1,110,976 -34.00%
-12 777,693 -35.94%
-13 548,390 -37.20%
-14 381,829 -38.35%
-15 261,275 -40.61%
Code:
Side Bet 1 Side Bet 1 Side Bet 1 Side Bet 1
Units Bet # Won EV WR
TOT 19,102,883 2,774,638 9.62% $3.68
24 9,227 1,872 63.80% $0.01
23 13,708 2,652 56.53% $0.02
22 21,914 4,030 50.48% $0.02
21 32,872 6,151 50.87% $0.03
20 47,043 8,597 45.92% $0.04
19 66,816 11,996 43.22% $0.06
18 102,127 18,209 41.70% $0.09
17 132,402 22,559 34.23% $0.09
16 201,468 33,945 32.02% $0.13
15 281,650 46,474 28.72% $0.16
14 394,687 63,642 25.47% $0.20
13 551,232 87,457 22.61% $0.25
12 762,317 118,975 20.02% $0.31
11 1,064,619 161,947 16.30% $0.35
10 1,468,023 219,225 13.71% $0.40
9 2,005,491 292,385 10.06% $0.40
8 2,776,180 397,799 7.67% $0.43
7 3,917,312 550,479 5.08% $0.40
6 5,236,366 722,632 2.56% $0.27So this looks like a good way to attack the game with a partner, one of you using a normal count and the other using this special count to determine when to play the sidebet.