It still depends on hands played in the time period of the rebate. If you get your losses rebated every year, but you play BS for 10M hands every year, you will always be down no matter what the loss rebate is, even 99%. As long as you play enough hands to be certain you will be down, no loss rebate can help you, even 99%. It's still the same idea. The loss rebate can be +EV, but it does not make the game itself plus EV.zengrifter said:
Get a Mathematical Advantage in Blackjack!
- Thread starter vincebreezy
- Start date
NDN21
Well-Known Member
Get a mathematical advantage.
Suppose you lose 100 units. If you were essentially betting 0.9 units to win one then you would still have 10 units right? But you don't because they have taken the full unit at the time of the loss. Therefore you are not essentially betting 0.9 units to win 1 unit.
The rebate is not awarded until the month's wagers are calculated (thereby factoring in a time constraint) the phrase could be better stated as "you are EVENTUALLY betting 0.9 units to win 1 unit" but there is a time variable you have to consider.
Also, you could win and the rebate doesn't come into play at all. In that case it's still 1 unit bet, 1 unit won.
Since the rebate is conditional (depends on a loss) wouldn't the correct formula be
-U +0.1U=rebate
Where U=unit
-U because you have to lose first for the rebate to come into play. Also because the casino takes the full loss at the time of the loss, the whole unit (not just 0.9 unit) then gives back the rebate which is +0.1 unit.
And expected value definition is the average of all possible outcomes. Since this rebate ONLY occurs after a loss then ALL possible outcomes are in the negative EV. The rebate given back to the player is not greater than the initial loss so the EV remains negative.
The game may be +EV, the rebate is definitely -EV.
Assume a coin toss is a 50/50 event. Say we flip a coin 100,000 times at $1 flip and you win 53,000 and I win 47,000. I am down 6,000 so you give back $600. The results are skewed in favor of one side, say tails. But a coin toss is still a 50/50 event, even after one side came up 6,000 more times than the other. That is tails is still just as likely to come up as heads, right? I believe they call that "independent trials", no previous trials has any influence on the outcome of following trials.
But some people incorrectly assume that the results would skew back toward even, 50 heads/50 tails (to keep the event at a 50/50 event, right?).
There is no guarantee that our results would skew back to 50/50. The results could remain "off" by 6,000 for 100,000 more trials. There is no mysterious "law of averages" that says things have to even out in the end. There is no rule that says the results must now skew heads because the last results skewed tails.
Just like at a craps table where there is no "law" that says "because a 7 hasn't come up in 85 throws one is due", there is no "law" that says "if in a heads-or-tails series one side has come up more than the other then the side that hasn't come up as often now must come up in order to keep heads-or-tails at a 50/50 event".
By accepting a 10% rebate on losses there is no guarantee that the results ever have to get back to 50/50 and thus ensure you recoup your losses plus the 10% rebate. You could start losing and keep losing indefinitely.
Heads or tails is a 50/50 event. The results of a series of heads or tails doesn't have to be a 50/50 event.
(P.S.- Heads-or-tails doesn't have to be a 50/50 event. If you use a quarter and place it heads-up then heads will turn up slightly more often than tails. The slower the coin turns then more often heads will show up.
Some people can flip a coin so that it looks like it's flipping along a horizontal axis but in reality it is just moving around a vertical axis that is slightly tilted, it's not really flipping but rather "waving" (like spinng a quarter on a desk, right before it stops it is waving with the same side always facing up, rather than flipping over and switching sides).)
I disagree. Try to tell the casino to let you bet 0.9 units. They won't let you do so. At the time of the wager, on a loss they take the full unit. It is still one unit risked, one unit won/lost.
Suppose you lose 100 units. If you were essentially betting 0.9 units to win one then you would still have 10 units right? But you don't because they have taken the full unit at the time of the loss. Therefore you are not essentially betting 0.9 units to win 1 unit.
The rebate is not awarded until the month's wagers are calculated (thereby factoring in a time constraint) the phrase could be better stated as "you are EVENTUALLY betting 0.9 units to win 1 unit" but there is a time variable you have to consider.
Also, you could win and the rebate doesn't come into play at all. In that case it's still 1 unit bet, 1 unit won.
Since the rebate is conditional (depends on a loss) wouldn't the correct formula be
-U +0.1U=rebate
Where U=unit
-U because you have to lose first for the rebate to come into play. Also because the casino takes the full loss at the time of the loss, the whole unit (not just 0.9 unit) then gives back the rebate which is +0.1 unit.
And expected value definition is the average of all possible outcomes. Since this rebate ONLY occurs after a loss then ALL possible outcomes are in the negative EV. The rebate given back to the player is not greater than the initial loss so the EV remains negative.
The game may be +EV, the rebate is definitely -EV.
0, a card counter should already have an advantage from counting cards, right? If you don't have an edge from card counting then why the heck are counting cards in the first place.zengrifter said:
Not necessarily.zengrifter said:
Assume a coin toss is a 50/50 event. Say we flip a coin 100,000 times at $1 flip and you win 53,000 and I win 47,000. I am down 6,000 so you give back $600. The results are skewed in favor of one side, say tails. But a coin toss is still a 50/50 event, even after one side came up 6,000 more times than the other. That is tails is still just as likely to come up as heads, right? I believe they call that "independent trials", no previous trials has any influence on the outcome of following trials.
But some people incorrectly assume that the results would skew back toward even, 50 heads/50 tails (to keep the event at a 50/50 event, right?).
There is no guarantee that our results would skew back to 50/50. The results could remain "off" by 6,000 for 100,000 more trials. There is no mysterious "law of averages" that says things have to even out in the end. There is no rule that says the results must now skew heads because the last results skewed tails.
Just like at a craps table where there is no "law" that says "because a 7 hasn't come up in 85 throws one is due", there is no "law" that says "if in a heads-or-tails series one side has come up more than the other then the side that hasn't come up as often now must come up in order to keep heads-or-tails at a 50/50 event".
By accepting a 10% rebate on losses there is no guarantee that the results ever have to get back to 50/50 and thus ensure you recoup your losses plus the 10% rebate. You could start losing and keep losing indefinitely.
Heads or tails is a 50/50 event. The results of a series of heads or tails doesn't have to be a 50/50 event.
(P.S.- Heads-or-tails doesn't have to be a 50/50 event. If you use a quarter and place it heads-up then heads will turn up slightly more often than tails. The slower the coin turns then more often heads will show up.
Some people can flip a coin so that it looks like it's flipping along a horizontal axis but in reality it is just moving around a vertical axis that is slightly tilted, it's not really flipping but rather "waving" (like spinng a quarter on a desk, right before it stops it is waving with the same side always facing up, rather than flipping over and switching sides).)
Last edited:
There is no question that a loss rebate on a coinflip game is +EV. The EV of the coinflip game is 0, so the 10% cushion for negative variance has to increase that.NDN21 said:
It's not that big of a deal, but if you want to avoid the time value problem, just do all your betting on the last day of the month. No biggie.NDN21 said:
zengrifter
Banned
Hep me out here, people. Toss me a friggin bone!
I'm the boss, need the info. zg
I'm the boss, need the info. zg
NDN21
Well-Known Member
zengrifter said:Hep me out here, people. Toss me a friggin bone!
I'm the boss, need the info. zg
Here is the bone.
http://en.wikipedia.org/wiki/Gambler%27s_fallacy
This coinflip proposition is not a +ev. You could lose......... and continue to lose............and continue to lose..................and continue to lose..........etc. because the probability of heads-or-tails is ALWAYS 50%.
From the link-
Last edited:
zengrifter said:
It depends on when they pay the rebate and how the BS bets.
Assuming a coin-toss game, your edge would be only 5% if they paid u after every loss and u flat bet. If u set a goal of winning say 10 times your winning bankroll, and the rebate would be paid every time u lost ur starting bankroll, it would be worth more like 90% of the rebate.
The idea is bet big, don't flat bet, lose often.
I've seen it mostly offered to big-betters who lose their $1MM bankroll and maybe the casino says "Ok just pay us $900,000."
zengrifter
Banned
Huuuhh, from Dr. Evil to no respect at all. I tell ya, its not easy being me! zgScottH said:
EasyRhino
Well-Known Member
Okay, I wish I could actually do the formulas, but I can't. First, we assume that a fixed amount of money is deposited for the entire month. Let's say it's only $100.
The value of the insurance itself is going to depend on a few things from your style of play:
a) Your odds of losing during the month
b) If there is a loss, its magnitude.
If the insurance has a nominal value of 10% of losses, then it can never be worth more than that (and that would result from a manner of play that ground every bankroll into dust)
So you can see, the resulting value of the insurance itself it worth more if you are more likely to have a large loss. However, the value of the insurance must be compared with the EV of the blackjack play itself. EV is a calculation we're much more accustomed to, but that depends on:
a) The odds of winning or losing for the month
b) The expected magnitude of the wins or losses
These two formulas are linked by the variance of possible win/loss for the month. If you were good at the math, you could come up with a break-even point. But I can't do that.
But at the common senese level, if a player plays a lot of hands (a few hundred), there is a fairly good chance that they will lose, and the EV of the loss will be fairly large, easily larger than 10% of bankroll for the month. Thus, it's still a net loss for the player.
But if the player plays only a single hand of $100, then the expected loss is only about 50 cents or a buck. The value of the insurance is about $5 in this case (given a 50% chance of loss). So, in this case, $5 is more than $1, so it's a net gain for the player.
The value of the insurance itself is going to depend on a few things from your style of play:
a) Your odds of losing during the month
b) If there is a loss, its magnitude.
If the insurance has a nominal value of 10% of losses, then it can never be worth more than that (and that would result from a manner of play that ground every bankroll into dust)
So you can see, the resulting value of the insurance itself it worth more if you are more likely to have a large loss. However, the value of the insurance must be compared with the EV of the blackjack play itself. EV is a calculation we're much more accustomed to, but that depends on:
a) The odds of winning or losing for the month
b) The expected magnitude of the wins or losses
These two formulas are linked by the variance of possible win/loss for the month. If you were good at the math, you could come up with a break-even point. But I can't do that.
But at the common senese level, if a player plays a lot of hands (a few hundred), there is a fairly good chance that they will lose, and the EV of the loss will be fairly large, easily larger than 10% of bankroll for the month. Thus, it's still a net loss for the player.
But if the player plays only a single hand of $100, then the expected loss is only about 50 cents or a buck. The value of the insurance is about $5 in this case (given a 50% chance of loss). So, in this case, $5 is more than $1, so it's a net gain for the player.
NDN21
Well-Known Member
There is no guarantee that you will ever get back to even. There is no guarantee because a coin flip is always 50/50.ScottH said:
So if you are down then you are down even with a rebate. Even if you just lost a $1 you are still down 90 cents.
All a rebate would do is lessen your losses.
For it to be +EV a player would have to be guaranteed that the results get to
- 50/50
or - within the amount of the rebate (assuming rebate has been awarded, that the player has rebate in pocket)
There is no guarantee so it is not +EV.
A player could lose $100 and theoretically never get back to even or get back to a $10 loss (assuming the rebate has been awarded).
Sonny
Well-Known Member
I have to disagree with that. Your Expected Value is simply that, what you expect. It makes no mention of variance or risk of ruin. It simple states that, on average, you will expect to make a profit. Even though a blackjack player has an advantage doesn’t mean that he won’t go broke or quit with a lifetime loss. Even after a series of losing sessions there is no guarantee that he will ever get back to even, yet he still has a +EV. The term EV refers to the theoretical advantage, not the actual outcome.NDN21 said:
Right, but that may be all you need in some cases. If you win $1 half the time and lose $0.90 the other half of the time, you will have a +EV. As Scott said, you are risking $0.90 to win $1. Since you are getting paid more than 1:1 odds on a fair bet you have a +EV.NDN21 said:
-Sonny-
I'd be willing to bet my entire bankroll that a coinflip with you giving me a loss rebate is +EV for me.NDN21 said:
OR, if you don't think it's +EV, why don't you try it with me? But hey, give me a 50% loss rebate since loss rebates don't make this +EV in your opinion.
NDN21
Well-Known Member
ScottH said:
If it's +EV then give me a 100 unit "headstart". Why not 500 units. Heck since it's +EV and will eventually end up with you making money then why not give 1000 units?
Start from 1000 units down. There is no guarantee you get back to even because that coin flip is still a 50/50 proposition.
NDN21
Well-Known Member
Yes, IN SOME CASES.Sonny said:
The casino pays the rebate at the end of the month. At the time of the bet you are not risking .90 units, you are risking one whole unit. The casino takes the whole unit at the time of the loss. You eventually get it back but at the time of the bet it's one unit risked. If you win that unit back before the end of the month then you get no rebate.
Here's a better explanation of why you can't be guaranteed to get back to even or within the 10% rebate (assuming it's been paid).
http://www.math.niu.edu/~rusin/known-math/99/averages
Sonny
Well-Known Member
So just be sure to limit the amount of action you give them each month. Assuming a 10% monthly rebate, if you bet one big hand every month then you are risking 90% to win 100%. Essentially the rebate is acting like a match play coupon. The more hands you play, the lower the EV becomes until it eventually turns –EV.NDN21 said:
Sure, playing one hand per month is a slow grind, but with such a big advantage you can still make some decent money with very little time invested.
The wizard has a nice analysis of this on his website:
http://www.wizardofodds.com/general/rebate.pdf
-Sonny-
Last edited:
NDN21
Well-Known Member
From the link
If you only bet once a month it is possible that you lose your first 6 bets and never get back to even, isn't it?
In that sense you cannot guarantee that you will make money off of this. It's impossible to guarantee that because the coin flip will always be 50/50.
How is that you can guarantee that if a player is down that they will always, always get back to even or within that 10% rebate (assume it's been paid).
What am I missing?
Then it's not a 50/50 situation. I am only talking about the coin flip scenario, not the blackjack or other casino games.
If you only bet once a month it is possible that you lose your first 6 bets and never get back to even, isn't it?
In that sense you cannot guarantee that you will make money off of this. It's impossible to guarantee that because the coin flip will always be 50/50.
How is that you can guarantee that if a player is down that they will always, always get back to even or within that 10% rebate (assume it's been paid).
What am I missing?