high true count buy pairs ?
- Thread starter beyondbj
- Start date
A traditional card counting system doesn't work well for this bet. Most systems track the number of high cards versus low cards, but this bet would need a system that tracks pairs to non-pairs. A high count (or a low one for that matter) would indicate that there is a higher proportion of certain card groups, but what you really need to know is the proportion of each card rank.
Most pair-related side bets are not very profitable and I doubt they would be a good cover opportunity since the betting opportunities are so rare.
-Sonny-
Most pair-related side bets are not very profitable and I doubt they would be a good cover opportunity since the betting opportunities are so rare.
-Sonny-
Have you simmed this? It seems to me like any time you have a heavy or light deck your probability of getting pair increases drastically. In high counts, you'd see more twenties and pairs of aces, and in low counts more pairs of low cards.
It's mostly about knowing what ranks are removed, not what remains.
It's mostly about knowing what ranks are removed, not what remains.
The problem is it's very non-linear. When you have a pair-based sidebet you will usually have an advantage if one rank is totally removed, but as soon as you add just one card of that missing rank back into the equation the advantage drops a whole bunch. You have something like a quadratic term in the equations for pair bets.
It can be done though, what you need to do is fill up a table with wiseguys and have each one track 2-3 ranks, and give the signal to let the dogs out when one or more ranks are depleted. An additional wiseguy or two tracking suits gives you a powerful force multiplier on some bets.
It can be done though, what you need to do is fill up a table with wiseguys and have each one track 2-3 ranks, and give the signal to let the dogs out when one or more ranks are depleted. An additional wiseguy or two tracking suits gives you a powerful force multiplier on some bets.
Hi,
My friends and I did very simple record before. Most of the casinos allow players to write anything on a score sheet in bacarrat table, as most bacarrat players want to find the trend or pattern, which is kind of silly.
Anyway, sometimes the penatration in bacarrat table is quite good. There might be less then 10 cards left in the shoe. We made a simple excel file and record every card dealt to see the change in the house edge. We made the written record first and then entered them into the excel as it won't be allowed to use computer near the table. The pair bet is 1 to 11.
We found the opportunities to get profit in pair bets is very few. For 20 shoes, only about 30 rounds when players have the advantage to bet pairs (1 to 11). We thought it worthless and gave up the plan.
My point is, even the penetration is good and you can keep track of every rank, I still don't think you can find good opportunities to bet pairs, let along use any bj card-counting method. Don't waste you money and time on it unless you want to make it as a camouflague bet.
My friends and I did very simple record before. Most of the casinos allow players to write anything on a score sheet in bacarrat table, as most bacarrat players want to find the trend or pattern, which is kind of silly.
Anyway, sometimes the penatration in bacarrat table is quite good. There might be less then 10 cards left in the shoe. We made a simple excel file and record every card dealt to see the change in the house edge. We made the written record first and then entered them into the excel as it won't be allowed to use computer near the table. The pair bet is 1 to 11.
We found the opportunities to get profit in pair bets is very few. For 20 shoes, only about 30 rounds when players have the advantage to bet pairs (1 to 11). We thought it worthless and gave up the plan.
My point is, even the penetration is good and you can keep track of every rank, I still don't think you can find good opportunities to bet pairs, let along use any bj card-counting method. Don't waste you money and time on it unless you want to make it as a camouflague bet.
i made a "simulation" of sorts in excel for this a while ago. the count ended up having to be unusefully high i believe.
If you take out all of any one rank the bet would be positive most of the time. There is a small cost of card removal. So if your first card out of the deck is a 9 there will only be 23 [4x6-1] 9s left in the deck where the bet would have been exactly fair if there were 24. but usually there will be imbalance among the other cards so that if the most likely card to get is of rank X then the most likely 2nd card will also be X. And the cards rank Y that is least likely to make a pair is also the least likely to be delt to you anyway.
This all makes the bet seem close to beatable (and it is at extreme counts) since after all taking out all of the 5s for example would be the equivalent of just a +4 true count. But the problem is a truecount of +4 is never all the cards of 1 rank being removed. Consider that each time the count goes up one more low card is gone making the expected value contributed by making your high pairs go up, but it makes the low pairs even more of a long shot on average when you get dealt a 2-6 as your first card which will still happen a lot.
If you take out all of any one rank the bet would be positive most of the time. There is a small cost of card removal. So if your first card out of the deck is a 9 there will only be 23 [4x6-1] 9s left in the deck where the bet would have been exactly fair if there were 24. but usually there will be imbalance among the other cards so that if the most likely card to get is of rank X then the most likely 2nd card will also be X. And the cards rank Y that is least likely to make a pair is also the least likely to be delt to you anyway.
This all makes the bet seem close to beatable (and it is at extreme counts) since after all taking out all of the 5s for example would be the equivalent of just a +4 true count. But the problem is a truecount of +4 is never all the cards of 1 rank being removed. Consider that each time the count goes up one more low card is gone making the expected value contributed by making your high pairs go up, but it makes the low pairs even more of a long shot on average when you get dealt a 2-6 as your first card which will still happen a lot.
6 Deck "Perfect Pairs" Sim
I actually just built a python sim for the "Perfect Pairs" sidegame offered at a nearby casino...
Game: 6 decks, HiLo count used
Payout: Suited match 30:1, colored match 10:1, non-colored match 5:1
Results:
https://docs.google.com/viewer?a=v&...lhZTAtNTVhYzllNWY5MTRj&hl=en&authkey=CMDwrt0K
The sim:
A six deck shoe is created and shuffled. Medium and either high or low cards are randomly removed until the shoe reaches the target TC. This shoe is then shuffled again and used to deal 3 hands to the player in the correct order while removing a dealer card before dealing the player the second cards for his hand. Payouts are determined, and the cards used for the hand are then reinserted into the deck and shuffled, before dealing again. Each datapoint represents the results of 3 hands played per shuffle * 5000 shuffles per shoe * 1000 random shoes generated per TC for a total of 15 million hands per count.
As you can see the results aren't perfect but they suggest that this is a break-even game at TC +/- 10. At around TC -15 the sidebet pays for the losses you should expect to recieve from your blackjack bet, and somewhere after TC < -15 you can justify EV-wise raising your minimum bet up to the max side bet amount employing "cover betting" in the process. However, if the camera's watching you, the chances that you'll be able to play this bet before getting kicked out is somewhere around 0%... Good luck
I actually just built a python sim for the "Perfect Pairs" sidegame offered at a nearby casino...
Game: 6 decks, HiLo count used
Payout: Suited match 30:1, colored match 10:1, non-colored match 5:1
Results:
https://docs.google.com/viewer?a=v&...lhZTAtNTVhYzllNWY5MTRj&hl=en&authkey=CMDwrt0K
The sim:
A six deck shoe is created and shuffled. Medium and either high or low cards are randomly removed until the shoe reaches the target TC. This shoe is then shuffled again and used to deal 3 hands to the player in the correct order while removing a dealer card before dealing the player the second cards for his hand. Payouts are determined, and the cards used for the hand are then reinserted into the deck and shuffled, before dealing again. Each datapoint represents the results of 3 hands played per shuffle * 5000 shuffles per shoe * 1000 random shoes generated per TC for a total of 15 million hands per count.
As you can see the results aren't perfect but they suggest that this is a break-even game at TC +/- 10. At around TC -15 the sidebet pays for the losses you should expect to recieve from your blackjack bet, and somewhere after TC < -15 you can justify EV-wise raising your minimum bet up to the max side bet amount employing "cover betting" in the process. However, if the camera's watching you, the chances that you'll be able to play this bet before getting kicked out is somewhere around 0%... Good luck
farmdoggy said:
I actually just built a python sim for the "Perfect Pairs" sidegame offered at a nearby casino...
Game: 6 decks, HiLo count used
Payout: Suited match 30:1, colored match 10:1, non-colored match 5:1
Results:
https://docs.google.com/viewer?a=v&...lhZTAtNTVhYzllNWY5MTRj&hl=en&authkey=CMDwrt0K
The sim:
A six deck shoe is created and shuffled. Medium and either high or low cards are randomly removed until the shoe reaches the target TC. This shoe is then shuffled again and used to deal 3 hands to the player in the correct order while removing a dealer card before dealing the player the second cards for his hand. Payouts are determined, and the cards used for the hand are then reinserted into the deck and shuffled, before dealing again. Each datapoint represents the results of 3 hands played per shuffle * 5000 shuffles per shoe * 1000 random shoes generated per TC for a total of 15 million hands per count.
As you can see the results aren't perfect but they suggest that this is a break-even game at TC +/- 10. At around TC -15 the sidebet pays for the losses you should expect to recieve from your blackjack bet, and somewhere after TC < -15 you can justify EV-wise raising your minimum bet up to the max side bet amount employing "cover betting" in the process. However, if the camera's watching you, the chances that you'll be able to play this bet before getting kicked out is somewhere around 0%... Good luck
Game: 6 decks, HiLo count used
Payout: Suited match 30:1, colored match 10:1, non-colored match 5:1
Results:
https://docs.google.com/viewer?a=v&...lhZTAtNTVhYzllNWY5MTRj&hl=en&authkey=CMDwrt0K
The sim:
A six deck shoe is created and shuffled. Medium and either high or low cards are randomly removed until the shoe reaches the target TC. This shoe is then shuffled again and used to deal 3 hands to the player in the correct order while removing a dealer card before dealing the player the second cards for his hand. Payouts are determined, and the cards used for the hand are then reinserted into the deck and shuffled, before dealing again. Each datapoint represents the results of 3 hands played per shuffle * 5000 shuffles per shoe * 1000 random shoes generated per TC for a total of 15 million hands per count.
As you can see the results aren't perfect but they suggest that this is a break-even game at TC +/- 10. At around TC -15 the sidebet pays for the losses you should expect to recieve from your blackjack bet, and somewhere after TC < -15 you can justify EV-wise raising your minimum bet up to the max side bet amount employing "cover betting" in the process. However, if the camera's watching you, the chances that you'll be able to play this bet before getting kicked out is somewhere around 0%... Good luck
I knew this thing would turn positive at high or low true counts. Any ideas about the match the dealer sidebet in blackjack and spanish? The effect has to be similar for that bet.
MtD for Blackjack and SP21
Thanks for the kind words, moo
I just got back from a 3 day snowboarding trip, otherwise I would have responded sooner.
I ran the sims using pretty much the same program, just changed the payouts and cards to match (and the deck/count type used for SP21).
Results:
6-deck Match the Dealer, payout for each player card that matches the dealer card is 11:1 (suited) or 4:1 (not suited):
https://docs.google.com/viewer?a=v&...JlNWQtZGUxZmUyNzgxYTlh&hl=en&authkey=CL2_nLgJ
6-deck SP21 MtD, payout is 9:1 (suited) or 4:1 (not suited). The SP21 HiLo count I used treats the 2 as a "medium" card:
https://docs.google.com/viewer?a=v&...gwOTctYWIwNWIyYTVmNjU4&hl=en&authkey=CJ6ul6kC
And here's the numerical data for both in case the graphs are hard to read...
https://docs.google.com/document/d/...hIP7YODQ8Ph9Y8770/edit?hl=en&authkey=CJqRt8kI
MtD turns out to be a better bet than perfect pairs, with a break-even game around TC +/- 8.5. And Spanish 21... Wow! The sidebet can actually be played semi-frequently! Break-even at TC +/- 6.
I was already thinking about learning SP21, but this certainly adds some incentive :devil: Just think of how perfectly you'd fit in with the plops while sitting at the SP21 table yelling "Yeah!!! Suited baby!!!!"
Thanks for the kind words, moo
I just got back from a 3 day snowboarding trip, otherwise I would have responded sooner.
I ran the sims using pretty much the same program, just changed the payouts and cards to match (and the deck/count type used for SP21).
Results:
6-deck Match the Dealer, payout for each player card that matches the dealer card is 11:1 (suited) or 4:1 (not suited):
https://docs.google.com/viewer?a=v&...JlNWQtZGUxZmUyNzgxYTlh&hl=en&authkey=CL2_nLgJ
6-deck SP21 MtD, payout is 9:1 (suited) or 4:1 (not suited). The SP21 HiLo count I used treats the 2 as a "medium" card:
https://docs.google.com/viewer?a=v&...gwOTctYWIwNWIyYTVmNjU4&hl=en&authkey=CJ6ul6kC
And here's the numerical data for both in case the graphs are hard to read...
https://docs.google.com/document/d/...hIP7YODQ8Ph9Y8770/edit?hl=en&authkey=CJqRt8kI
MtD turns out to be a better bet than perfect pairs, with a break-even game around TC +/- 8.5. And Spanish 21... Wow! The sidebet can actually be played semi-frequently! Break-even at TC +/- 6.
I was already thinking about learning SP21, but this certainly adds some incentive :devil: Just think of how perfectly you'd fit in with the plops while sitting at the SP21 table yelling "Yeah!!! Suited baby!!!!"
Thanks to those who have input here already, but:
I have pondered over this myself, and actually used it for cover, and i used it for cover when the count was WAY WAY in the tank, or WAY in through the roof. (as if those are the times if an AP is watching they would know, or more think I am not counting). But also w/ the thought process that the chances are better it's at least not "AS" non-profitable.
But even w/ lets say a high-low, and if the above stated sim is correct, that it needs TC +/- 10....
I don't know about you, but a TC of that high is once maybe every few HOURS. Maybe see that for 4-5 hands.....in a 4-5 hr time frame.
So even if TC 10 is the threshold, that just so insanly high (or low) it just never occurs. We are talking the same threshold you would split 10's against a dealer 4 almost...astronomical.
I don't know, but the common sence to me just says, they PAY 11 to 1, and their are 13 cards in the deck, hence the odds are 13 to 10
odds of 13 to 1, with a payout o 11 to 1. That's so much juice (like 18% i believe) off the bat....
Even if the TC was at +11, there has to be TONS of varience in that, b/c that is just a gauge. It's like using Hi-Low for insurance, and being pissed off when it does not work. It's just NOT that accurate at all.
The only way to truly do this is to account for EVERY value seperatly, and when their became inbalances (such as 14Jacks remain and 16Queens remain, but on average on 5 of all other cards remain, etc)
.....just brainstorming, but one would have to be rainman like.
Thank to whom came up with the prior post. But in all practical applicability, TC of +/- 10 is just so rare, it is not even profitble to plan what has a 51% probability at those rare times.
If they (casino) were not so stingy and make it 12:1 payout, then...THEN...it's time to get cracking on some systems, but at 11:1, its mute
I have pondered over this myself, and actually used it for cover, and i used it for cover when the count was WAY WAY in the tank, or WAY in through the roof. (as if those are the times if an AP is watching they would know, or more think I am not counting). But also w/ the thought process that the chances are better it's at least not "AS" non-profitable.
But even w/ lets say a high-low, and if the above stated sim is correct, that it needs TC +/- 10....
I don't know about you, but a TC of that high is once maybe every few HOURS. Maybe see that for 4-5 hands.....in a 4-5 hr time frame.
So even if TC 10 is the threshold, that just so insanly high (or low) it just never occurs. We are talking the same threshold you would split 10's against a dealer 4 almost...astronomical.
I don't know, but the common sence to me just says, they PAY 11 to 1, and their are 13 cards in the deck, hence the odds are 13 to 10
odds of 13 to 1, with a payout o 11 to 1. That's so much juice (like 18% i believe) off the bat....
Even if the TC was at +11, there has to be TONS of varience in that, b/c that is just a gauge. It's like using Hi-Low for insurance, and being pissed off when it does not work. It's just NOT that accurate at all.
The only way to truly do this is to account for EVERY value seperatly, and when their became inbalances (such as 14Jacks remain and 16Queens remain, but on average on 5 of all other cards remain, etc)
.....just brainstorming, but one would have to be rainman like.
Thank to whom came up with the prior post. But in all practical applicability, TC of +/- 10 is just so rare, it is not even profitble to plan what has a 51% probability at those rare times.
If they (casino) were not so stingy and make it 12:1 payout, then...THEN...it's time to get cracking on some systems, but at 11:1, its mute
Even if one rank was depleted, that would mean it's now 12:1 odds
(yet still with an 11:1 payout)
even w/ 2 ranks depleted, (which would be rare, unless very very deep pen).....you then just hit breakeven
for clarify.....my post is in relation to PAIRS to the player, not "dealer match"
have never seen or played dealer match before. But w/ 13 ranks, and to have it match, it sounds like 13:1 odds and the same ratio as having a pair yourself. Now if BOTH your cards have to match the dealer....then it sounds even worse.
again, the opportunity for +EV is so rare it seem unreachable
(yet still with an 11:1 payout)
even w/ 2 ranks depleted, (which would be rare, unless very very deep pen).....you then just hit breakeven
for clarify.....my post is in relation to PAIRS to the player, not "dealer match"
have never seen or played dealer match before. But w/ 13 ranks, and to have it match, it sounds like 13:1 odds and the same ratio as having a pair yourself. Now if BOTH your cards have to match the dealer....then it sounds even worse.
again, the opportunity for +EV is so rare it seem unreachable
For the "Perfect Pairs" game the average payout on a pair is actually 12.5:1 for an infinite deck given the pay table that I described [0.25*30(for suited) + 0.25*10(colored but not suited) + 0.5*5(not colored)]...
For Match the Dealer the Payout on 6 decks is 11:1 on a suited match, but there are 2 chances for a match per hand, and this game has more playing opportunities and a lower variance than perfect pairs.
The standard deviation on the game is HUGE, but I think you were more referring to the side game correlation to the HiLo count. Yes, there are certain deck compositions at TC +5 that can be favorable to the player, just as there are deck compositions at TC +15 that are unfavorable to the player to play this bet... For example (assume ridiculous pen) if the remaining cards are [10, J, Q, K, A] (whatever suits) the TC is +52 and the player has a 0% chance of winning the pair game. This is the reason why I had 1000 random decks generated for each count... To get the average advantage. And BTW, my HiLo insurance bets seem to do just fine There are better methods to beat this game, but it's still better than nothing to use information you already have... Even though you won't always be making the right decision, you will enough of the time.
For Match the Dealer the Payout on 6 decks is 11:1 on a suited match, but there are 2 chances for a match per hand, and this game has more playing opportunities and a lower variance than perfect pairs.
The standard deviation on the game is HUGE, but I think you were more referring to the side game correlation to the HiLo count. Yes, there are certain deck compositions at TC +5 that can be favorable to the player, just as there are deck compositions at TC +15 that are unfavorable to the player to play this bet... For example (assume ridiculous pen) if the remaining cards are [10, J, Q, K, A] (whatever suits) the TC is +52 and the player has a 0% chance of winning the pair game. This is the reason why I had 1000 random decks generated for each count... To get the average advantage. And BTW, my HiLo insurance bets seem to do just fine
There are better methods to beat this game, but it's still better than nothing to use information you already have... Even though you won't always be making the right decision, you will enough of the time.You're using incorrect terminology here. If there's a one in twelve chance of something happening, that means that the odds against it are 11:1.
But if one rank WAS depleted; the odds STILL wouldn't be 11:1 (the break even point, or what's known as "fair" odds), unless you were playing with an infinite deck. If you were playing SINGLE deck with one rank depleted (and all other cards remain), then the chance of getting a pair would only be three in forty seven, or almost one in sixteen. Of course; the more decks are being used, the closer you would COME to the break even point. For eight decks with one rank missing; the chance of getting a pair is 31 in 384 - getting closer, but STILL less than "fair" odds.
But if one rank WAS depleted; the odds STILL wouldn't be 11:1 (the break even point, or what's known as "fair" odds), unless you were playing with an infinite deck. If you were playing SINGLE deck with one rank depleted (and all other cards remain), then the chance of getting a pair would only be three in forty seven, or almost one in sixteen. Of course; the more decks are being used, the closer you would COME to the break even point. For eight decks with one rank missing; the chance of getting a pair is 31 in 384 - getting closer, but STILL less than "fair" odds.