PensiveGerbil
PensiveGerbil
  • Threads: 2
  • Posts: 29
Joined: Oct 13, 2012
October 13th, 2012 at 3:39:00 AM permalink
If you play craps to have fun and try to win money, then variance is your friend. If you have other goals, however, variance can be an enemy. I have been considering craps betting strategies which may reduce variance. Before I share my ideas, let me pose a hypothetical. Suppose you are given $800 in chips to bet at a $10 minimum craps table with 3-4-5x odds. At the end of the session, you will have to return whatever chips you have left. However, if you succeed in placing bets with a total value of $30,000 before going bust, you will win a prize of $1000 minus the amount you lost (you get the full $1000 if your bets yielded a profit). Pretty good deal! :-) So what betting strategy (or strategies) will maximize your expected prize money?
AlanMendelson
AlanMendelson
  • Threads: 167
  • Posts: 5937
Joined: Oct 5, 2011
October 13th, 2012 at 3:43:10 AM permalink
Since craps is a negative expectation game, you are destined to lose. And the longer you play the more you will lose. So what are you getting at? And what are you selling?
PensiveGerbil
PensiveGerbil
  • Threads: 2
  • Posts: 29
Joined: Oct 13, 2012
October 13th, 2012 at 4:18:48 AM permalink
Quote: AlanMendelson

Since craps is a negative expectation game, you are destined to lose. And the longer you play the more you will lose. So what are you getting at? And what are you selling?



I have nothing to sell. I am seeking guidance and discussion about reduced variance betting strategies in a manner which I hope will minimize controversy and off-topic posts.
s2dbaker
s2dbaker
  • Threads: 51
  • Posts: 3259
Joined: Jun 10, 2010
October 13th, 2012 at 6:35:57 AM permalink
Quote: PensiveGerbil

If you play craps to have fun and try to win money, then variance is your friend. If you have other goals, however...

I don't know about the rest of you but when I gamble, I totally hate myself and try to lose all of the money as fast as I can so that I don't have to gamble anymore.
Someday, joor goin' to see the name of Googie Gomez in lights and joor goin' to say to joorself, "Was that her?" and then joor goin' to answer to joorself, "That was her!" But you know somethin' mister? I was always her yuss nobody knows it! - Googie Gomez
MathExtremist
MathExtremist
  • Threads: 88
  • Posts: 6526
Joined: Aug 31, 2010
October 13th, 2012 at 7:33:59 AM permalink
A gazillion years ago I came up with a better (lower-edge) hedge than the typical don't + place system. For a $10 3/4/5x table, do this:
$10 pass
$55 don't pass
Take odds as follows:
4 or 10, $30 for a guaranteed win of $15
5 or 9, $36 for a guaranteed win of $9
6 or 8, $40 for a win of either $3 or $5.

You could also bet 10 pass, 65 don't and take full odds. That wins 5 on every winner and a larger amount on sevens, which might be preferable due to the come out exposure of 7, 11. A long hand just wins 5 per roll, and many established numbers helps you recover a bit from the eventual 7-out.

Do this on every roll with come bets.

But with the terms you cited, you should figure out if the iron cross plus a 7 hedge isn't better. No time now tho.


Edit: I think this is the answer you were looking for...
150 do, 150 don't, 5 midnight. Guarantees you at least 30k in handle with no more than $500 loss for 500 prize win.
"In my own case, when it seemed to me after a long illness that death was close at hand, I found no little solace in playing constantly at dice." -- Girolamo Cardano, 1563
Scotty71
Scotty71
  • Threads: 19
  • Posts: 289
Joined: Mar 5, 2011
October 13th, 2012 at 8:01:06 AM permalink
Quote: PensiveGerbil

I have nothing to sell. I am seeking guidance and discussion about reduced variance betting strategies in a manner which I hope will minimize controversy and off-topic posts.



All this and you just joined the forums today...you should PM 98Steps!
when man determined to destroy himself he picked the was of shall and finding only why smashed it into because." — E.E. Cummings
SanchoPanza
SanchoPanza
  • Threads: 34
  • Posts: 3502
Joined: May 10, 2010
October 13th, 2012 at 8:03:39 AM permalink
Quote: PensiveGerbil

So what betting strategy (or strategies) will maximize your expected prize money?


Can the gambler decide when the session or sessions begin and end? That is no small question.
MrV
MrV
  • Threads: 364
  • Posts: 8158
Joined: Feb 13, 2010
October 13th, 2012 at 8:14:17 AM permalink
Quote: s2dbaker

I don't know about the rest of you but when I gamble, I totally hate myself and try to lose all of the money as fast as I can so that I don't have to gamble anymore.



Actually, I suspect that is true, at least subconsciously, or many who gamble regularly.

No better place to reinforce a deep-seated sense of self-loathing than in a casino.
"What, me worry?"
PensiveGerbil
PensiveGerbil
  • Threads: 2
  • Posts: 29
Joined: Oct 13, 2012
October 13th, 2012 at 8:51:42 AM permalink
Quote: MathExtremist

A gazillion years ago I came up with a better (lower-edge) hedge than the typical don't + place system. For a $10 3/4/5x table, do this:
$10 pass
$55 don't pass
Take odds as follows:
4 or 10, $30 for a guaranteed win of $15
5 or 9, $36 for a guaranteed win of $9
6 or 8, $40 for a win of either $3 or $5.

You could also bet 10 pass, 65 don't and take full odds. That wins 5 on every winner and a larger amount on sevens, which might be preferable due to the come out exposure of 7, 11. A long hand just wins 5 per roll, and many established numbers helps you recover a bit from the eventual 7-out.

Do this on every roll with come bets.

But with the terms you cited, you should figure out if the iron cross plus a 7 hedge isn't better. No time now tho.


Edit: I think this is the answer you were looking for...
150 do, 150 don't, 5 midnight. Guarantees you at least 30k in handle with no more than $500 loss for 500 prize win.



If I understand correctly, you have suggested three strategies which involve making simultaneous Pass and Don't Pass bets on come out rolls? And your first two strategies including taking odds only on your Pass bet? What does "5 midnight" mean?

I believe your strategies fail to maximize the expected prize money largely because the house edge of your bets is far from minimized. I think the simple strategy of betting $10 on both Pass and Don't Pass and then taking/laying maximum odds when a point is established would be optimal. Do you agree?

Potential drawbacks of this strategy in a real live casino craps game would probably include ridicule, hostility, and embarrassment! So let's introduce one more condition: simultaneous Pass Line and Don't Pass bets are prohibited (as well as simultaneous Come and Don't Come bets). With this new condition, what strategies would be optimal?
Ahigh
Ahigh
  • Threads: 90
  • Posts: 5197
Joined: May 19, 2010
October 13th, 2012 at 9:15:51 AM permalink
In general there are two components to a good low-variance strategy:

1) Use low edge bets (unless you are trying to lose)
2) Try not to win too much money

Using a combination of come bets and/or DC bets, try to win as little as possible. If you can imagine every roll of the dice and what would happen, aim towards pushing as an end result or maybe winning a dollar or two.

The variance usually goes away if you do this. There are many details for how to do this, but generally increasing the size of the come bet to win as much on the last come as you lose on travelled bets works.

The tricky part is that you are still betting against the crap rolls with this strategy. So you lose on crap, and you win on yo, and you can also win if you get repeat numbers and don't start reducing the amount of the come bet to compensate for what would happen on a seven to prevent from winning.

The point of this example isn't that this is necessarily a good or the best strategy to decrease variance. Just that variance is the result of attempting to win and having it sometimes go your way and sometimes not. If you try _not_ to win _or_ lose but still play the game, the variance goes away for a zero edge game. For a small edge game the variance _usually_ goes away except for the case where you lose _sometimes_. The lower the edge, the less frequently you lose, and the rest of the time you have a low to no variance "game."
aahigh.com
PensiveGerbil
PensiveGerbil
  • Threads: 2
  • Posts: 29
Joined: Oct 13, 2012
October 13th, 2012 at 9:58:04 AM permalink
Quote: Ahigh

In general there are two components to a good low-variance strategy:

1) Use low edge bets (unless you are trying to lose)
2) Try not to win too much money

Using a combination of come bets and/or DC bets, try to win as little as possible. If you can imagine every roll of the dice and what would happen, aim towards pushing as an end result or maybe winning a dollar or two.



This sounds like the right track to me.

Quote: Ahigh


The variance usually goes away if you do this. There are many details for how to do this, but generally increasing the size of the come bet to win as much on the last come as you lose on travelled bets works.



I do not know what "travelled bets" are. However, it sounds like you are advocating increasing the size of bets following losses. I think this will increase variance. Perhaps I do not understand what you are suggesting.
Ahigh
Ahigh
  • Threads: 90
  • Posts: 5197
Joined: May 19, 2010
October 13th, 2012 at 10:59:53 AM permalink
The lowest variance strategy for craps is minimum on the pass or don't pass and nothing else.

All I'm saying, and you can figure it out however you want, is you can increase and decrease the subsequent bets on the come bet and still have a relatively low edge strategy where each bet has a good chance to win and therefore your variance is relatively minimized.

This also allows you to play a game where you try to guess how long the shooter is going to last, and try to bet the come bet right before he sevens out without looking like a player who is betting against the shooter.

I hope this makes sense. There's lot of variations on combinations of come bets and DC bets after a $5 line bet that are very low edge and low variance.

Good luck experimenting and have fun. If you keep your bet sizes down, and don't think too much just randomly pick the pass line, the don't pass line, the come bet, or the DC, you will be just fine. Don't think too much and try not to win too much and you will be fine.
aahigh.com
TIMSPEED
TIMSPEED
  • Threads: 89
  • Posts: 1246
Joined: Aug 11, 2010
October 13th, 2012 at 1:03:28 PM permalink
Partner up and play the doey don't for $10 flat and 3/4/5...thats what I do in Reno..i know im going to lose, but then again, im going to get back my losses in comps and freebies...easy choice
Gambling calls to me...like this ~> http://www.youtube.com/watch?v=4Nap37mNSmQ
AlanMendelson
AlanMendelson
  • Threads: 167
  • Posts: 5937
Joined: Oct 5, 2011
October 13th, 2012 at 1:34:20 PM permalink
I know that some casinos allow the doey/dont bets, but they will not credit your account for $$ played as the bets cancel out each other.

This happened at Mandalay Bay a few years back. They allowed the bets so that the player could move his odds based on the point, but they gave him no credit towards comps since the flat bets cancelled out and there was no house advantage on the odds.

So the question was, did he win any money? Not enough to cover his bill.
MathExtremist
MathExtremist
  • Threads: 88
  • Posts: 6526
Joined: Aug 31, 2010
October 13th, 2012 at 2:38:27 PM permalink
Quote: PensiveGerbil

If I understand correctly, you have suggested three strategies which involve making simultaneous Pass and Don't Pass bets on come out rolls? And your first two strategies including taking odds only on your Pass bet? What does "5 midnight" mean?

I believe your strategies fail to maximize the expected prize money largely because the house edge of your bets is far from minimized. I think the simple strategy of betting $10 on both Pass and Don't Pass and then taking/laying maximum odds when a point is established would be optimal. Do you agree?

Potential drawbacks of this strategy in a real live casino craps game would probably include ridicule, hostility, and embarrassment! So let's introduce one more condition: simultaneous Pass Line and Don't Pass bets are prohibited (as well as simultaneous Come and Don't Come bets). With this new condition, what strategies would be optimal?


That's not a realistic condition. I can bet however I want at a dice table, and if I have the edge I don't care about ridicule. With a 30k playthrough requirement where odds count, make $60 each do & don't plus $2 12, and take and lay full odds. (near the end, if necessary, reduce the line bets a few bucks to maximize odds.) The 12 hedge trades off edge for variance, but that's more important when you have a prize on the line. You'll hit 30k in around 45 bets at a cost of around 90. So with the prize you win over $900.

But you need to define optimal. Maximizing EV with a promo means increasing the odds of meeting its criteria, and just making unhedged bets doesn't do that. It's not that hard to lose 800 at a 10 table with 3/4/5 odds, and if you do that you fail. So playing 10 don't and laying full odds might be the lowest EV, but there's also a non-zero chance that you lose altogether and miss out on the prize. Are you willing to accept that chance?
"In my own case, when it seemed to me after a long illness that death was close at hand, I found no little solace in playing constantly at dice." -- Girolamo Cardano, 1563
PensiveGerbil
PensiveGerbil
  • Threads: 2
  • Posts: 29
Joined: Oct 13, 2012
October 13th, 2012 at 7:11:07 PM permalink
Quote: TIMSPEED

Partner up and play the doey don't for $10 flat and 3/4/5...thats what I do in Reno..i know im going to lose, but then again, im going to get back my losses in comps and freebies...easy choice



I believe you have cleverly deduced the practical application of my hypothetical problem. :) Of course, it is very common for casinos to offer prizes for making a certain quantity of bets. They are called comps. Like you, my goal in playing craps is to win prizes whose value exceeds the expected loss of my bets, and to accomplish this without losing a painful amount of money. I suspect this is usually only possible at casinos that give comp credit for your free odds bets. I chose to present the problem as I did to highlight the evolution of my reasoning and to postpone my thread being derailed by posts declaring it is impossible to receive comps exceeding the expected loss of your bets. [So far, I have achieved this goal. However, I recognize that the casino's generosity may have been related to my being a new customer. If I stop receiving desired comps whose value exceeds the expected loss of my bets, I will stop playing craps.]

When you play the doey-don't with a partner, do you attempt to disguise your partnership from the casino staff? Do you think most casinos would penalize you (i.e., reduce your comps) upon discovering you are betting with a partner to guarantee yourself a very small loss? [I think they would be wise to penalize this strategy, since it so clearly demonstrates a reluctance to gamble.] So while your partner strategy may be optimal (if it is not detected and penalized), I would like to explore reduced variance strategies you can use when you do not have a partner or do not want to subject her to a boring craps session! :O
PensiveGerbil
PensiveGerbil
  • Threads: 2
  • Posts: 29
Joined: Oct 13, 2012
October 13th, 2012 at 7:13:02 PM permalink
Quote: AlanMendelson

I know that some casinos allow the doey/dont bets, but they will not credit your account for $$ played as the bets cancel out each other.

This happened at Mandalay Bay a few years back. They allowed the bets so that the player could move his odds based on the point, but they gave him no credit towards comps since the flat bets cancelled out and there was no house advantage on the odds.

So the question was, did he win any money? Not enough to cover his bill.



The doey/don't bets do not really cancel out; they double the player's expected loss (assuming he makes two opposing $10 bets on come-out rolls instead of one $10 bet). Nevertheless, I can understand why a casino may want to penalize players that use this strategy.
PensiveGerbil
PensiveGerbil
  • Threads: 2
  • Posts: 29
Joined: Oct 13, 2012
October 13th, 2012 at 7:34:46 PM permalink
Quote: MathExtremist

That's not a realistic condition. I can bet however I want at a dice table, and if I have the edge I don't care about ridicule.



I added the "no simultaneous Do/Don't bets" condition because I DO care about social relations at the craps table and it is MY hypothetical! [And now that I have revealed the practical application of this exercise, you may appreciate my concern that the casino might reduce my prizes if my variance reduction strategy is excessively obvious.]

Quote: MathExtremist

With a 30k playthrough requirement where odds count, make $60 each do & don't plus $2 12, and take and lay full odds. (near the end, if necessary, reduce the line bets a few bucks to maximize odds.) The 12 hedge trades off edge for variance, but that's more important when you have a prize on the line. You'll hit 30k in around 45 bets at a cost of around 90. So with the prize you win over $900.



We are now in close agreement...except perhaps for the $2 hedge bet against a 12

Quote: MathExtremist

But you need to define optimal. Maximizing EV with a promo means increasing the odds of meeting its criteria, and just making unhedged bets doesn't do that....



I specified that the prize money would be reduced by the value of the chips that you lost so that the optimal strategy would need to keep the house edge low while reducing variance.

To be honest, I did not consider the 12 hedge because I did not know such a bet existed. I am a craps newbie and I only just glanced at the "house edge" column in the wizard's "Proposition Bets" table before concluding I would not make any of those bets so no need to learn them. So the question now is whether the 12 hedge bet is worth the cost. Two $60 bets and one $2 bet adds up to $122, and 45 x $122 = $5490. Since another $24510 is needed to hit 30k, your projection assumes that around 82% of your bets following this strategy will be on the odds. Based on the craps simulation results recently posted by 7craps here: https://wizardofvegas.com/forum/gambling/craps/11507-high-variance-right-vs-darkside/ , I think the true proportion is closer to 76% Excluding the $2 hedge bets, your average bet size will be $120 plus $380 in odds for a total of $500, requiring about 60 bets to reach 30K at a cost of about $101 Adding the $2 hedge bet would bring your average bet size up to $502. You will still need about 60 bets to reach 30K but the 12 hedge bets will cost up to an addition $20 in lost expectation (based on a maximum house edge of 16.67%). Since I believe it would be extremely unlikely to go bust using our doey-don't strategy without the 12 hedge, I doubt it would be worth the cost. An increased likelihood of winning a smaller prize does not necessarily increase your prize money expectation.
Mission146
Mission146
  • Threads: 142
  • Posts: 16832
Joined: May 15, 2012
October 13th, 2012 at 7:44:06 PM permalink
Quote: PensiveGerbil

The doey/don't bets do not really cancel out; they double the player's expected loss (assuming he makes two opposing $10 bets on come-out rolls instead of one $10 bet). Nevertheless, I can understand why a casino may want to penalize players that use this strategy.



Emphasis added.

I'm afraid that's slightly incorrect. The Expected loss on a Don't Pass bet for a $100 bet is $1.40, so $0.70 for $50. (Based on bet resolved) The Expected loss on a Pass Line Bet for a $100 bet is $1.41 (Bet Resolved) or $0.705 on $50.

If you make a $50 PL and a $50 DP bet, then everything except for the 12 self-negates, and the 12 only loses on the PL, so you lose $50.

($100/2) * 1/36 = 1.38888888888 Expected Loss or $1.39.

However, I do not personally see why a casino would want to penalize this style. The Expected Loss is REALLY close to being the same as just picking one or the other, but the player can literally (discluding odds on one or the other, which still have the same theoretical result, because Odds are technically a seperate bet with a 0 HE) only lose or break even. Break even is as good as it gets for the player, so I'm not sure why the casino wouldn't want to encourage this behavior.

EDIT: In fairness, I am being somewhat semantical, here. What you are looking at is the House Edge for Pass Bet made of 1.414% + House Edge for Don't Pass bet MADE of 1.364% divided by 2 = 1.389% or -$1.39 (rounded) EV per the above. Technically speaking, the Don't Pass Bet is unresolved in this scenario, but I don't look at it as such because I would view the Doey-Don't 12 as the only outcome that resolves anything. If you do not make any Odds bets, the only thing that can result in other than a break-even is a 12 on the CO, so I'm calling the overall bet resolved in this scenario.
https://wizardofvegas.com/forum/off-topic/gripes/11182-pet-peeves/120/#post815219
Ahigh
Ahigh
  • Threads: 90
  • Posts: 5197
Joined: May 19, 2010
October 13th, 2012 at 7:59:33 PM permalink
Short answer why is because the casino is also leveraging a relatively unlimited bankroll in addition to the edge (IE: their betting system aside from the edge is to have a bigger bank than you) in order to make more money off the average player.

IE: Martingale does work as a strategy provided you have an infinite bank. And built into the casinos typical bets (the flip side of 32 across with pressure for example) is a martingale betting strategy with a relatively unlimited bankroll compared to whoever is making some multiple of 32 across on them.

If you gamble without any chance to win at all, this generally destroys the casino's ability to out-bank you to leverage a bankroll to make more money. They're simply counting 12's and have no other way to make a cent off of you. IE: most people lose more than the average edge times the average bet because they get greedy and expect to win an unrealistic amount and instead lose it all.

If this were not the case, the Santa Ana Star probably wouldn't still be offering a zero edge field bet.

But most people get greedy and forget that even with no edge, the chance of winning is still only 44% and martingale in the field even with zero edge is an expensive proposition. Even Fibonacci has a rough time in the field. So the bigger bankroll always wins!

But reverse martingale against the house has a much better chance of getting to max bet at Santa Ana Star, so if you have a big bank, and want to hit the house maximum bet from the minimum, better to get paid triple on the two.
aahigh.com
sodawater
sodawater
  • Threads: 64
  • Posts: 3321
Joined: May 14, 2012
October 13th, 2012 at 8:05:37 PM permalink
deleted
Last edited by: sodawater on Oct 1, 2018
Mission146
Mission146
  • Threads: 142
  • Posts: 16832
Joined: May 15, 2012
October 13th, 2012 at 8:11:40 PM permalink
Quote: sodawater



Did you just prove the point you were trying to argue against? Playing a $50 pass and a $50 don't is the same player disadvantage as playing a $100 pass OR don't. It does double the expected loss.

A $50 line bet has an expected loss of ~ 70 cents. As you just proved, $50 opposing line bets have an expected loss of $1.39



I didn't disprove my own point, which is why I said, "Slightly incorrect." When you talk about the $0.705 Expected Loss, then if you are going to round that rounds up to $0.71. Not only are you rounding down on that, but in addition to that, we're also (correctly) rounding up the Overall from an Expected Loss of $1.3888888 to $1.39. It's not double.

EDIT: If you look at my (unquoted) Edit from the previous post, which I doubtlessly Edited too late, you'll see that I'm basically just looking at this from a standpoint of semantics claiming that Midnight on the CO is the only outcome that resolves the overall bet, so I'm considering the DP portion of the bet (and essentially considering the whole thing as one bet) resolved, where technically, one might say the DP is not resolved.
https://wizardofvegas.com/forum/off-topic/gripes/11182-pet-peeves/120/#post815219
sodawater
sodawater
  • Threads: 64
  • Posts: 3321
Joined: May 14, 2012
October 13th, 2012 at 8:17:30 PM permalink
deleted
Last edited by: sodawater on Oct 1, 2018
Mission146
Mission146
  • Threads: 142
  • Posts: 16832
Joined: May 15, 2012
October 13th, 2012 at 8:21:18 PM permalink
Quote: sodawater

Oh... I didn't know you were talking about the 1-2 cents... the difference obviously comes from the fact that the DP is slightly better than the pass



Actually, no, because then you would still be looking at an average of $1.40+ per $100 bet resolved. The difference comes from the fact that an, "Unresolved," DP (considering Bet MADE) has a HE of 1.36% because of Midnight being a Push. That's where the semantics comes into play. Midnight does not resolve a DP bet, but since it is the only outcome that can resolve the overall bet (in this case) in anything other than a break-even, I'm calling the DP bet resolved in this scenario.

The HE per bet MADE is actually worse in this scenario than DP alone. The HE per bet resolved is better than either, but that's only because I'm calling the whole thing one bet and the DP resolved.
https://wizardofvegas.com/forum/off-topic/gripes/11182-pet-peeves/120/#post815219
PensiveGerbil
PensiveGerbil
  • Threads: 2
  • Posts: 29
Joined: Oct 13, 2012
October 13th, 2012 at 8:26:03 PM permalink
Quote: Mission146

Quote: PensiveGerbil

The doey/don't bets do not really cancel out; they double the player's expected loss (assuming he makes two opposing $10 bets on come-out rolls instead of one $10 bet). Nevertheless, I can understand why a casino may want to penalize players that use this strategy.



Emphasis added.

I'm afraid that's slightly incorrect. The Expected loss on a Don't Pass bet for a $100 bet is $1.40, so $0.70 for $50. (Based on bet resolved) The Expected loss on a Pass Line Bet for a $100 bet is $1.41 (Bet Resolved) or $0.705 on $50.

If you make a $50 PL and a $50 DP bet, then everything except for the 12 self-negates, and the 12 only loses on the PL, so you lose $50.

($100/2) * 1/36 = 1.38888888888 Expected Loss or $1.39.

However, I do not personally see why a casino would want to penalize this style. The Expected Loss is REALLY close to being the same as just picking one or the other, but the player can literally (discluding odds on one or the other, which still have the same theoretical result, because Odds are technically a seperate bet with a 0 HE) only lose or break even. Break even is as good as it gets for the player, so I'm not sure why the casino wouldn't want to encourage this behavior.



Indeed, I should have said the strategy "NEARLY doubles" the player's expected loss.

I think casinos may be wise to penalize players using this strategy (i.e., offer less comps) for a couple of reasons. First, I think this strategy may detract from the fun party atmosphere and camraderie they hope to achieve at a craps table. Secondly, it should alert them to the possibility that this person is not playing craps to have fun or win money; he may have other goals. He may be an advantage gambler who expects to receive freebies worth more than the theoretical loss of his action. Of course, if this doey-don't player starts making high HE prop bets, it would be a different story!
Mission146
Mission146
  • Threads: 142
  • Posts: 16832
Joined: May 15, 2012
October 13th, 2012 at 8:39:56 PM permalink
Quote: PensiveGerbil

If you play craps to have fun and try to win money, then variance is your friend. If you have other goals, however, variance can be an enemy. I have been considering craps betting strategies which may reduce variance. Before I share my ideas, let me pose a hypothetical. Suppose you are given $800 in chips to bet at a $10 minimum craps table with 3-4-5x odds. At the end of the session, you will have to return whatever chips you have left. However, if you succeed in placing bets with a total value of $30,000 before going bust, you will win a prize of $1000 minus the amount you lost (you get the full $1000 if your bets yielded a profit). Pretty good deal! :-) So what betting strategy (or strategies) will maximize your expected prize money?



The Minimum Bet is $10.00 and you must win $30,000 before you go bust. If you slap $10 on DP + PL, then you lose only $10 of your $800 bankroll everytime a Twelve comes up. You would make a total of 30,000/20 = 1,500 such bets in order to complete the objective.

Of the 1,500 bets, 1/36 should result in a CO of 12.

1,500 * 1/36 = 41.66666667 times one should expect to see Midnight. The Expected Loss on that is -$416.67.

You'd have to see Midnight 79 times in the 1500 rolls (It wouldn't be 80, because after 79 you no longer have enough to make both bets. Specifically, it would be 79 Midnights and then hope for a win on DP or PL)

You'd have to see Midnight, then, nearly twice as often as you should. Specifically, at an absolute minimum (to fail) you would have to see Midnight 79 times in 1,499 rolls (and still lose the last one) which would be 1,499/79 = 1:18.975 rolls.

-----

I hope 7Craps is following this thread. I'd be interested in determining whether it would be better to just bet the minimum all the way, or to start off betting $400 on each ($800 total bet) and then $200 ($400 total, if a loss) then $100 ($200 total) and so on. You would have Six 12's 400-200-100-50-25-12.50 before you would be reduced to betting either one or the other, but of course, you rack it up to $30,000 faster.
https://wizardofvegas.com/forum/off-topic/gripes/11182-pet-peeves/120/#post815219
Mission146
Mission146
  • Threads: 142
  • Posts: 16832
Joined: May 15, 2012
October 13th, 2012 at 8:46:47 PM permalink
Quote: PensiveGerbil



Indeed, I should have said the strategy "NEARLY doubles" the player's expected loss.

I think casinos may be wise to penalize players using this strategy (i.e., offer less comps) for a couple of reasons. First, I think this strategy may detract from the fun party atmosphere and camraderie they hope to achieve at a craps table. Secondly, it should alert them to the possibility that this person is not playing craps to have fun or win money; he may have other goals. He may be an advantage gambler who expects to receive freebies worth more than the theoretical loss of his action. Of course, if this doey-don't player starts making high HE prop bets, it would be a different story!



I don't know. The HE is way lower, but it seems to me that it is about the same thing as slapping the same amount of money down on Black/Red, Even/Odd, 1-18/19-36 on Roulette, or any combination of those bets. The House just kicks back and waits for a zero. I suppose if what the House was offering in Comps was GREATER than the player's theoretical loss, then I would be inclined to agree with you, but how often does that happen? If it were the case, though, then I suppose (particularly with Craps) that you would just count one of the bets for Comp purposes.
https://wizardofvegas.com/forum/off-topic/gripes/11182-pet-peeves/120/#post815219
PensiveGerbil
PensiveGerbil
  • Threads: 2
  • Posts: 29
Joined: Oct 13, 2012
October 13th, 2012 at 10:02:05 PM permalink
Quote: Mission146


The Minimum Bet is $10.00 and you must win $30,000 before you go bust. If you slap $10 on DP + PL, then you lose only $10 of your $800 bankroll everytime a Twelve comes up. You would make a total of 30,000/20 = 1,500 such bets in order to complete the objective.

Of the 1,500 bets, 1/36 should result in a CO of 12.

1,500 * 1/36 = 41.66666667 times one should expect to see Midnight. The Expected Loss on that is -$416.67.

You'd have to see Midnight 79 times in the 1500 rolls (It wouldn't be 80, because after 79 you no longer have enough to make both bets. Specifically, it would be 79 Midnights and then hope for a win on DP or PL)

You'd have to see Midnight, then, nearly twice as often as you should. Specifically, at an absolute minimum (to fail) you would have to see Midnight 79 times in 1,499 rolls (and still lose the last one) which would be 1,499/79 = 1:18.975 rolls.



So I'm guessing now that "midnight" = 12 :)

Note that the hypothetical contest required making $30,000 worth of bets, not winning $30,000. The expected prize money value of your doey-don't strategy can be increased by taking/laying odds to reduce the overall house edge of your bets.
98Clubs
98Clubs
  • Threads: 52
  • Posts: 1728
Joined: Jun 3, 2010
October 13th, 2012 at 10:44:26 PM permalink
$10 Pass and $10 Field. (Field is 3x on the 12.)
4-9-10 uses Field win as odds, 2-Yo-12 banks.
Some people need to reimagine their thinking.
Mission146
Mission146
  • Threads: 142
  • Posts: 16832
Joined: May 15, 2012
October 13th, 2012 at 11:49:25 PM permalink
Quote: PensiveGerbil

So I'm guessing now that "midnight" = 12 :)

Note that the hypothetical contest required making $30,000 worth of bets, not winning $30,000. The expected prize money value of your doey-don't strategy can be increased by taking/laying odds to reduce the overall house edge of your bets.



I'm glad to hear that, because betting the Doey-Don't and taking/laying Odds on both or not taking/laying odds means that I can't possibly win either way, all I can do is lose. I'm going to tell you this, though, that's a bullshit play even for a hypothetical!!! I'm going to get away with running the Doey-Don't while simultaneously Taking/Laying Max Odds and that's all going to count towards the $30,000 bet? No way!

OK.

I'm going to Establish a point 24/36 times and break even otherwise 11/36 times. I'm going to put $61 on each end of the Doey-Don't. My bet is $122 and I am going to always be Laying $366 in Odds and Taking up to $305 for a total of up to $793.

2: 1/36 = $122 bet

3: 2/36 = $122 bet

4: 3/36 = $671 bet

5: 4/36 = $732 bet

6: 5/36 = $793 bet

7: 6/36 = $122 bet

8: 5/36 = $793 bet

9: 4/36 = $732 bet

10: 3/36 = $671 bet

11: 2/36 = $122 bet

12: 1/36 = $122 bet ($61 lost)

Bet Totals: 12/36 = $122....Expected Bet $40.67

6/36 = $671....Expected Bet $111.83

8/36 = $732....Expected Bet $162.67

10/36 = $793....Expected Bet $220.28

Overall Expected Bets = $535.45 (Per CO)

Expected Loss 1/36 * 61 = $1.69 (Per CO)

---If you lose a bet, then you have $739 left. The next step down would be $56 on each with $112 bet Laying $336 Taking $280 (Max) for a total of up to $728.
https://wizardofvegas.com/forum/off-topic/gripes/11182-pet-peeves/120/#post815219
MathExtremist
MathExtremist
  • Threads: 88
  • Posts: 6526
Joined: Aug 31, 2010
October 14th, 2012 at 12:10:27 AM permalink
Quote: PensiveGerbil

To be honest, I did not consider the 12 hedge because I did not know such a bet existed. I am a craps newbie and I only just glanced at the "house edge" column in the wizard's "Proposition Bets" table before concluding I would not make any of those bets so no need to learn them.


That's normally wise, but you've injected a very juicy prize into the game that is strongly positive. The question is how to get at it.
Quote:

So the question now is whether the 12 hedge bet is worth the cost. Two $60 bets and one $2 bet adds up to $122, and 45 x $122 = $5490. Since another $24510 is needed to hit 30k, your projection assumes that around 82% of your bets following this strategy will be on the odds. Based on the craps simulation results recently posted by 7craps here: https://wizardofvegas.com/forum/gambling/craps/11507-high-variance-right-vs-darkside/ , I think the true proportion is closer to 76% Excluding the $2 hedge bets, your average bet size will be $120 plus $380 in odds for a total of $500, requiring about 60 bets to reach 30K at a cost of about $101


No, you take *and* lay odds. Point is 6, you take 250 and lay 300. They wash, but it's an extra 550 in action plus your original 120 is 670. Now, Mission already pointed out that no casino is going to give you this kind of deal, where odds action counts as handle for promo purposes, so this is all academic. The real answer to your question is "don't play for comps". A typical casino will evaluate your play and estimate your worth ... then comp you about 30% of that. If you just stick to the pass line and odds, you'll be rated a tight player and not get much. Your $10 line bet is only worth 14c to the casino and they know it. An hour of play at a full table making just pass bets, regardless of odds, is worth about half a bet to the house ($5 at a $10 table) so what do you think you'll get back? Not much, if the house is paying attention.

Edit: I meant 300 and 360, but you get the point.
"In my own case, when it seemed to me after a long illness that death was close at hand, I found no little solace in playing constantly at dice." -- Girolamo Cardano, 1563
PensiveGerbil
PensiveGerbil
  • Threads: 2
  • Posts: 29
Joined: Oct 13, 2012
October 14th, 2012 at 12:22:36 AM permalink
Quote: Mission146

I'm glad to hear that, because betting the Doey-Don't and taking/laying Odds on both or not taking/laying odds means that I can't possibly win either way, all I can do is lose. I'm going to tell you this, though, that's a bullshit play even for a hypothetical!!! I'm going to get away with running the Doey-Don't while simultaneously Taking/Laying Max Odds and that's all going to count towards the $30,000 bet? No way!



I agree!!!! That is why in my third post of this thread (replying to MathExtremist), I added the condition of prohibiting simultaneous Pass/Don't Pass bets and simultaneous Come/Don't Come bets. Maybe I should have added this condition to my original post as well, but I think the problem as originally stated is a good starting point for understanding the basis of a more useful reduced variance betting strategy. I hope it was an interesting brain teaser. :)
PensiveGerbil
PensiveGerbil
  • Threads: 2
  • Posts: 29
Joined: Oct 13, 2012
October 14th, 2012 at 12:28:30 AM permalink
Quote: MathExtremist


No, you take *and* lay odds. Point is 6, you take 250 and lay 300. They wash, but it's an extra 550 in action plus your original 120 is 670.



I believe you may be failing to consider the times that your $120 come-out roll bet is resolved in one roll, providing no opportunity to take or lay odds. These $120 bets will bring down your average bet size.
AxiomOfChoice
AxiomOfChoice
  • Threads: 32
  • Posts: 5761
Joined: Sep 12, 2012
October 14th, 2012 at 12:38:56 AM permalink
Bet the do and the don't for the table minimum each time. When you establish a point, place max odds on both sides. Stop once you have bet a total of $30k.

One out of 36 times, you will lose $10.

Each time you are betting:

$20 when the bet is resolved on the come-out (12 times in 36)
$110 when the point is a 4 or 10 (6 times in 36)
$120 when the points is a 5 or 9 (8 times in 36)
$130 when the points is a 6 or 8 (10 times in 36)

So, on average, every 36 resolutions, you are wagering $3160 and losing $10.

So your EV here is over $900... that is, your expected loss before hitting $30k in bets is under $100.

Edit: And now I see, of course, that I should have read the whole thread before replying. It's definitely the correct answer to the question as posted, though :)
PensiveGerbil
PensiveGerbil
  • Threads: 2
  • Posts: 29
Joined: Oct 13, 2012
October 14th, 2012 at 12:45:54 AM permalink
Quote: Ahigh

The lowest variance strategy for craps is minimum on the pass or don't pass and nothing else.

All I'm saying, and you can figure it out however you want, is you can increase and decrease the subsequent bets on the come bet and still have a relatively low edge strategy where each bet has a good chance to win and therefore your variance is relatively minimized.

This also allows you to play a game where you try to guess how long the shooter is going to last, and try to bet the come bet right before he sevens out without looking like a player who is betting against the shooter.

I hope this makes sense. There's lot of variations on combinations of come bets and DC bets after a $5 line bet that are very low edge and low variance.

Good luck experimenting and have fun. If you keep your bet sizes down, and don't think too much just randomly pick the pass line, the don't pass line, the come bet, or the DC, you will be just fine. Don't think too much and try not to win too much and you will be fine.



I will have fun in Vegas next month while staying in a nice comped room on the strip that I earned by playing craps (for the first time) last month for about three hours. Since playing craps is not among the activities I most enjoy, however, I do not really want to extend the time needed to reach my playthrough goal by keeping my bets small.

I have been experimenting with combining Don't Pass and Come bets (and visa versa) so that I can take and lay the odds at the same time (but not on the same number). I can adjust the amount of my odds bets so that a rolled "7" will cause me to break even or have a small net win. One thing I like about this strategy is it would enable me to root for the shooter (to roll my Come number) even while the bulk of my bets are against the shooter.

I would like to hear people's opinions about this general strategy and how it may be perceived by other players and the casino staff. What potential pitfalls may I be overlooking?
Mission146
Mission146
  • Threads: 142
  • Posts: 16832
Joined: May 15, 2012
October 14th, 2012 at 12:47:00 AM permalink
Quote: PensiveGerbil

I believe you may be failing to consider the times that your $120 come-out roll bet is resolved in one roll, providing no opportunity to take or lay odds. These $120 bets will bring down your average bet size.



ME is basing his Odds off of a $50 bet, not a $60 one, and you can see my last post on the previous page for Expected Total Bet given perfect distribution of the dice, though that was based on a $61 bet for $122 total. The $100 total is 81.97%, so the overall expected bet should be my $535.45 * .8197 for $438.91.

That's still going to rack up pretty quickly.
https://wizardofvegas.com/forum/off-topic/gripes/11182-pet-peeves/120/#post815219
PensiveGerbil
PensiveGerbil
  • Threads: 2
  • Posts: 29
Joined: Oct 13, 2012
October 14th, 2012 at 12:48:03 AM permalink
Quote: AxiomOfChoice

Bet the do and the don't for the table minimum each time. When you establish a point, place max odds on both sides. Stop once you have bet a total of $30k.

One out of 36 times, you will lose $10.

Each time you are betting:

$20 when the bet is resolved on the come-out (12 times in 36)
$110 when the point is a 4 or 10 (6 times in 36)
$120 when the points is a 5 or 9 (8 times in 36)
$130 when the points is a 6 or 8 (10 times in 36)

So, on average, every 36 resolutions, you are wagering $3160 and losing $10.

So your EV here is over $900... that is, your expected loss before hitting $30k in bets is under $100.

Edit: And now I see, of course, that I should have read the whole thread before replying. It's definitely the correct answer to the question as posted, though :)



Haha! You get an A on the quiz. ;)
MathExtremist
MathExtremist
  • Threads: 88
  • Posts: 6526
Joined: Aug 31, 2010
October 14th, 2012 at 12:50:54 AM permalink
Quote: PensiveGerbil

I believe you may be failing to consider the times that your $120 come-out roll bet is resolved in one roll, providing no opportunity to take or lay odds. These $120 bets will bring down your average bet size.


Sure, but that doesn't matter for the problem at hand. You just care about making that promo. But now we know you're actually just interested in playing for comps, which again is just a bad idea. If you must, craps is the wrong game to play for comps. Learn to count BJ well enough to play even and get comps that way. Or play VP on multiple points days (which almost never happens with tables).
"In my own case, when it seemed to me after a long illness that death was close at hand, I found no little solace in playing constantly at dice." -- Girolamo Cardano, 1563
PensiveGerbil
PensiveGerbil
  • Threads: 2
  • Posts: 29
Joined: Oct 13, 2012
October 14th, 2012 at 1:18:05 AM permalink
Quote: MathExtremist

... But now we know you're actually just interested in playing for comps, which again is just a bad idea. If you must, craps is the wrong game to play for comps. Learn to count BJ well enough to play even and get comps that way. Or play VP on multiple points days (which almost never happens with tables).



The fact that VP and BJ require considerably more skill and effort to play well is a relevant factor when deciding which games may be best for comps. Depending on your betting strategy, I suspect that craps can also provide a lower variance pathway to comps. This would be a good topic for another thread. If my craps play ceases to gain the desired prizes, I will consider other options. :)
PensiveGerbil
PensiveGerbil
  • Threads: 2
  • Posts: 29
Joined: Oct 13, 2012
October 14th, 2012 at 2:24:04 AM permalink
Quote: MathExtremist

Sure, but that doesn't matter for the problem at hand. You just care about making that promo....



For the purpose of closure, I will say what I think is the optimal answer to the hypothetical problem posed in my original post (before I added the condition of no simultaneous Pass/Don't Pass bets). The player should start with the strategy of making the minimum Pass and Don't Pass bet on come-out rolls and taking/laying maximum odds when a point is established. Then if his bankroll drops below a certain threshold point due to an extraordinarily unlikely run of Box Cars (12's) on the dice, he should transition into a strategy that adds the 12 prop bet to hedge against more 12's being rolled to guarantee he will not go bust. Details like what the transitional bankroll threshold should be could be mathematically calculated, but I don't think it is worth the effort. ;)
MathExtremist
MathExtremist
  • Threads: 88
  • Posts: 6526
Joined: Aug 31, 2010
October 14th, 2012 at 8:39:28 AM permalink
Quote: PensiveGerbil

The fact that VP and BJ require considerably more skill and effort to play well is a relevant factor when deciding which games may be best for comps. Depending on your betting strategy, I suspect that craps can also provide a lower variance pathway to comps. This would be a good topic for another thread. If my craps play ceases to gain the desired prizes, I will consider other options. :)


If you want skill-free comps, play baccarat. Many casinos don't rate odds play in dice, and the edge on bacc is lower than the pass line. If your casino rates $25 pass + 3/4/5x odds as $100 avg bet (or even $90) that's one thing. But if you're rated at $25, the edge on both baccarat and European roulette (outside bets) is lower. Now, if you have more fun playing craps than those other games, you should play craps. But that's why I keep saying not to play for comps. You have a better EV by going to work, earning money, and spending it on stuff than you do trying to get the casino to give you the same stuff for "free".
"In my own case, when it seemed to me after a long illness that death was close at hand, I found no little solace in playing constantly at dice." -- Girolamo Cardano, 1563
AxiomOfChoice
AxiomOfChoice
  • Threads: 32
  • Posts: 5761
Joined: Sep 12, 2012
October 14th, 2012 at 3:35:38 PM permalink
Quote: MathExtremist

If you want skill-free comps, play baccarat. Many casinos don't rate odds play in dice, and the edge on bacc is lower than the pass line. If your casino rates $25 pass + 3/4/5x odds as $100 avg bet (or even $90) that's one thing. But if you're rated at $25, the edge on both baccarat and European roulette (outside bets) is lower. Now, if you have more fun playing craps than those other games, you should play craps. But that's why I keep saying not to play for comps. You have a better EV by going to work, earning money, and spending it on stuff than you do trying to get the casino to give you the same stuff for "free".



I'm not sure that I agree with you about baccarat. It depends what table you're at, and how they rate it. If you're at a mini table (no squeeze), forget it. The HE per hand may be lower, but it is a lot higher per hour (the hands just fly out there). If you're at a midi squeeze game, that goes a lot slower, so maybe? Also, some casinos don't expect you to bet on every hand in baccarat, and will rate you at double your bet if you do, so that might be a point in its favor. I'm not sure how widespread this practice is.

One of my favorite games for comps is UTH. First, I really enjoy it (most important thing). The variance is high, and you have the edge when you have the big bets out. If you play properly (get Grossjean's strategy card... I got it from the link on the Wizard's site; best $6 I ever spent) the house edge is 2.3% of your ante (don't play the trips). 2.3% is kind of high, but there are two points working in its favor:

1. It is a VERY slow game if the table is full. Most dealers have trouble getting 25 hands per hour out. So the house edge per hour is only about 1/2 an ante or so.

2. Most people play the game terribly, and mistakes are extremely expensive, so the house actually makes a lot more money than the 2.3% that they should. I suspect that most players are giving up a double-digit edge to the house. Since they make so much money from it, they rate the game VERY well. I know that at MGM casinos they rate you at 3x your ante bet (plus 1x your trips bet, if you make it, which you shouldn't). So, at a full table, you are being rated at about 6x your hourly loss. For example, if you play green chips, your expected loss is about $13 an hour or so, but you are being rated at $75/hand. This is better than playing basic strategy at all but the very best (usually high-limit) blackjack games.

There is one other thing about this game -- at many casinos, the dealers are not good. The misread the board and make mispays. This can be good for you if you don't speak up when they make a mistake in your favor, but very bad for you if you don't know how to read the board yourself. The most common board-reading errors are kicker errors (not realizing which kickers play and which don't). The most common mispay error is not paying the blind bet in non-obvious situations. Eg, the dealer has a low pair; you have a high pair and make a straight with your other card. The high pair is more obviously seen than the straight (which you don't "need" in order to win the hand) and so they don't pay your blind at 1-1 for your straight. I have seen a dealer pay everyone's ante instead of pushing it when she didn't qualify, but there are definitely more mispays in favor of the casino than the player, so pay attention.
teddys
teddys
  • Threads: 150
  • Posts: 5527
Joined: Nov 14, 2009
October 14th, 2012 at 3:48:29 PM permalink
Quote: AxiomOfChoice

One of my favorite games for comps is UTH. First, I really enjoy it (most important thing). The variance is high, and you have the edge when you have the big bets out. If you play properly (get Grossjean's strategy card... I got it from the link on the Wizard's site; best $6 I ever spent) the house edge is 2.3% of your ante (don't play the trips). 2.3% is kind of high, but there are two points working in its favor:

It's like that at a lot of casinos. One casino in Michigan would give me back about 2-2.5x my expected loss per hour in comp dollars. And you could convert those comp dollars to freeplay. Don't bother playing at Harrah's casinos, though. They never rate table games players worth anything. Good info on the MGM though.
"Dice, verily, are armed with goads and driving-hooks, deceiving and tormenting, causing grievous woe." -Rig Veda 10.34.4
  • Jump to: