How Many Come Bets is Too Many?

This actually started out on the thread that should not be named, and I felt that further discussion on this topic should be moved to its own thread.

Quote: RaleighCraps

Okay, it has been shown that the bet with the lowest HA is the DP w odds, but since most of us prefer to play the 'right' side, the second best bet becomes the PL w odds.
Now, given that I am not inclined to just sit there and wait for that point to be decided, I would assume that my next best play is to make a Come bet and take the odds, which provides the same HA as the PL.
I now have 2 bets working, but again, I am now bored, so I place a 3rd come bet with odds. Most books I have read indicate that one PL bet and 2 come bets are about as high as you should go, after that you get diminishing returns. Is there math to support that limit of 2 come bets? Or, does the reason have more to do with the extreme volatility that is introduced with so many come bets in play, and the massive loss from the 7 out?

I have recently been playing to establish 3 come bets. My thinking is that with the PL and 3 come bets I have an average of 16 rolls that win vs 6 that lose. There are two very subjective reasons I even considered this: 1) the number of times I established 2 come bets, and the shooter had a long roll but never hit the points I was on; with 3 come bets I cover more than half the possible "good" rolls. 2) the frequency of long rolls I observed when playing the don't, and getting my bets wiped out.

It hurts when that 7 rolls before any of the numbers hit, but it seems rare that I actually establish the four points without repeats and then 7 out before any of the numbers have hit. Somewhere I tried to calculate the probability for this - maybe I'll track that down tomorrow. I feel like I have had more winning sessions playing this way, but this is a very subjective feeling and is not based on a very large number of sessions. I may have just been lucky!!

The volatility definitely seems higher, but I limit my initial stake so that I don't lose big (to me) on any given session, but if I happen to hit even a moderately long roll I come out ahead. I have not yet encountered the inevitable streak of consecutive losing sessions that will make me wonder what the heck I was thinking.

All those come bets are subject to the same 1.4% HA, so in the long run, it does not matter how many come bets are made. So far I have been happy and comfortable with the +/- effect on my bankroll.

edit: Thread title should be "How Many Come Bets are Too Many" -- my English teachers are cringing. :)

My MO is playing $10-25 with 5x on the Pass Line and then $10 Coming with Full Odds (5x) every roll. It can be very frustrating to get the "Bases Loaded" and then 7 out, but I have had nice long streaks of just getting paid on every roll (except 2,3,12) and raking in the chips. Usually I place the 6/8 for $30 as well, then if it hits, that pays for the Odds and the next Come. A lot of action, a low HA, and a lot of fun for me. I don't like to stop betting the Come, because then I end up with uncovered numbers.

seattledice,
Thanks for creating the new thread.
Here is my thinking on the issue (for simplicity sake Come bets means PL and/or Come):
The total amount one is willing to put into play is the biggest factor. Let's assume $100 is the bet limit. Are you better off with 2 Come bets with 40 odds each, or 5 come bets with 10 odds?
As I understand it, the HA is 1.41% for the Come bet. When you add odds to the bet, you effectively lower the HA a bit. So 5 come bets has a higher HA than 2 come bets with higher odds. However, I would think you have a better chance of losing less with the 5 come bets, than with the 2, since you have more numbers in play.

If we expand this to Come vs Place bets, I think an overlooked factor is the comp factor. Odds bets are monies at risk to a 7 out, but, you get no comp credit for that bet, since there is no HA. On a 3-4-5x table, a PL bet with full odds pays the same amount as a place bet. Yes, the HA is higher on the place bet, due to the loss of the come out roll advantage. But the comps that are available for making Place/Buy bets can add up quite nicely, and thus should be factored in to get the true EV.

So now the decision is 1 PL bet, plus x number of Come bets, or Place bets instead of, or in conjunction with, Come bets

Gordy

Quote: cclub79

My MO is playing $10-25 with 5x on the Pass Line and then $10 Coming with Full Odds (5x) every roll. It can be very frustrating to get the "Bases Loaded" and then 7 out, but I have had nice long streaks of just getting paid on every roll (except 2,3,12) and raking in the chips. Usually I place the 6/8 for $30 as well, then if it hits, that pays for the Odds and the next Come. A lot of action, a low HA, and a lot of fun for me. I don't like to stop betting the Come, because then I end up with uncovered numbers.

A Vegas friend, who is as close as you can get to a "professional" dice player (tongue-in-cheek, of course; I know they don't exist), plays this way. He always chides me for putting out one or two come bets per roll. He explained it once; something about having to hit too many numbers to get back to even my way. I didn't really understand the math behind it. Maybe someone here does?

I'd never chide anyone playing ANY Come Bets. If I see someone playing the Come, I know that at the least they have an above average understanding of the best bets on the table. (actually, EDIT, I never chide anyone at all...) Again, I play this way because it maximizes my enjoyment and action at one of the statistically best values in the casino (.326% or even .184% if I play at Bally's or Wild Wild West, where they offer 10x Odds 24/7...and 1.52% on the couple of place bets).

Your expected loss is the same with a $20 PL bet as it is with a $10 Pl bet and a $10 come bet. For every $100 you wager on either line, you expect to lose $1.41. As long as you're taking max odds, taking more come bets is a decent alternative to increasing your pass line wager.

I would tend to agree about the risk. A come bet with odds is as good as the pass line with odds. People get hung up with the fact that the come bet comes after the come out roll and you are apt to lose more this way. But just like the pass bet, you have an 11.111% on the come bet when it is placed.

Many are superstitious about come bets and even pass line bet as you have to hit the same number twice to win.

For those on a limited bankroll, you have to be careful with come bets and odds because you can easily put lots of money out there.

In the end, "too" many come bets is a function of your tolerance to volatility. If you can handle putting up continuous come bets with max odds ($280 loss at a $10 3-4-5 table), then go for it.

For myself, my volatility at a table is limited to about 6 times the minimum which limits me with two come bets with double odds or a pass line bet with triple odds and a 6 - 8 to place. How I play is based on my mood. I know that the HA on my bet is at the most 1.52% and I am comfortable with that.

It seems to me that there are several factors to consider. The first, and probably most important, is how much money are you willing to risk? Obviously, the more you bet the more you can lose if you have bad luck, regardless of where the money is bet.

Secondly, consider the "tax" that you are going to play on your bets. The come bet costs 1.41% of the bet, just like the passline. Placing the 6/8 costs 1.52%, placing the 5/9 costs 4.0%, buy it may be cheaper to buy those, depending on bet level and casinos rules, buying the 4/10 costs 4.76%, or substantially less, here again depending on the bet level and casino rules. OTOH, money you bet on odds bets behind your pass/come bets is not taxed, although it also does not count for comps purposes. So, from a purely expected-loss point of view, the more money you bet on odds bets, the lower the expected loss.

Third, what about volatility? Obviously, the more you bet the more volatility, but different bets have different degrees of variance, the flat part of the pass and come having the least variance (standard deviation almost exactly equal to the bet amount), while place and buy bets have more and odds bets slightly more again. Here:

bet SD as % of bet
pass/come 99.9%
place 6/8 107.9%
place 5/9 117.6%
buy 4/10 139.1%
odds 6/8 109.5%
odds 5/9 122.5%
odds 4/10 141.4%
odds avg. 122.5%

So, if you want volatility, you get by far the most "bang for your buck" with odds bets, which have the highest variance at no cost. Of course, volatility works both ways, so you can get hammered more, but really not much more than with place bets.

Fourth, how much "action" do you want, not in terms of total money bet, but in terms of having lots of numbers covered, so you get more bets resolved. It can be very frustrating to be on the pass line with 5X or 10X odds and have a shooter roll every point number except THE point, especially when you are the shooter, and you're making money for everybody at the table except yourself. Of course, there's the opposite situation where you make your point, winning the flat and the odds, while the place bettors get nothing. It a matter of patterns; the pass/odds bettor is chasing a pattern of making points, not caring how many others numbers are thrown; the "$nn inside" or "$nn across" bettors are chasing a pattern of lots of point numbers rolled before a seven. They require maybe 4 hits before the seven in order to pocket a win for that shooter. If you have the inside numbers covered, here are the probabilities of different numbers of hits, assuming winning bets are replaced:

0 hits .25 cum
1 hit .1875 .4375
2 hits .1406 .5781
3 hits .1055 .6836
4 hits .0791
5 hits .0593
6 hits .0445
etc., etc.
So, if you need 4 hits to be ahead, almost 70% of the time that's not going to happen.

Another way to get more numbers covered is to make come bets, and you can also take odds on those bets for more "free variance". You can make a come bet on every non-comeout roll. People think you can quickly cover all the numbers that way, but it's not so, because you have 1/3 of come bets resolved on the first roll. Then, you may have a come bet go to the six, you make another come bet and the six shows. You win the first one and still have the six covered, but not an additional number. I ran a quick simulation with a $10 pass and $10 come bet on every non-comeout roll. The number of come bets resolved was only 2.39 times the number of passline decisions. Of course, that means less money bet than placing all the inside numbers, and less expected loss, but also less "action".

If you make come bets, then the question comes up of whether to "work" your odds on the passline comeout roll. It's revenue-neutral, of course, in terms of expectation, but if you want "action" and variance, then you get less of it if you don't have them working. Of course, the dealers will love you. >:-) Actually, one mark of a good dealer, in my view, is the ability to "absorb" a player's (or multiple players') strategy and not have to be reminded all the time. I used to tell the dealers, "My come odds ALWAYS work", and I'd expect him/her to say, "You got it!", and then put that "On" button on my odds every time.

Let's compare a $10 come bet, taking double odds, to a $20 buy 4, $1 vig collected only on a win:

$10 come $20 buy 4
avg. bet: $23.33 $20
ev -$.14 -$1
SD $28.57 $27.81

For 60 bets:
ev -$8.48 -$20
SD $221 $215

The buy 4 resolves at a somewhat slower rate, 4.0 rolls to 3.375, on average.
Many people overlook the comeout roll when comparing come bets to place bets.

If you can't wait to get your money out there on lots of numbers, the place bets allow you to do that, but you pay for it in expected loss.

I used to play pass w/two come bets, taking whatever odds I felt I could afford, but I stopped, for two reasons:
1) Casinos started allowing higher odds multiples than double (when I started playing some casinos offered only single odds), and I understood that I was paying 1.4% for the flat part of those come bets.
2) I didn't like the "mixed feelings" I had on the passline comeout roll, when the seven would win the pass but lose any come points, including their odds. I decided I wanted to "root" for the seven on the pass comeout and against it when there was a point, period.

So, how many come bets are too many? I would say, more than you can afford! Every player should decide for him/herself his/her own comfort level, preferably in the light of knowledge of the risks and advantages.
Cheers,
Alan Shank

Quote: boymimbo

Many are superstitious about come bets and even pass line bet as you have to hit the same number twice to win.

I have to call out your language here (but not your premise). You do not have to hit the same number twice. The first number only establishes the point. It has no effect on your bet (unless a natural or craps is rolled, but that is the same as the line bet). After that, the odds are whatever the odds against making the point is: 6-5, 3-2, or 2-1. The odds are not calculated based on hitting a certain number "twice."

Edit: goatcabin, great analysis. What is your bettting pattern now, just one pass line with max odds? Why not just have your odds turned off on the comeout roll, then you can "root for the seven?" My dice-playing friend does this. He hates rolling the seven, though, because all his flat bets come down, and like you said, it's a lot of work to get those numbers up there.

Quote: teddys

Quote: boymimbo

Many are superstitious about come bets and even pass line bet as you have to hit the same number twice to win.

I have to call out your language here (but not your premise). You do not have to hit the same number twice. The first number only establishes the point. It has no effect on your bet (unless a natural or craps is rolled, but that is the same as the line bet). After that, the odds are whatever the odds against making the point is: 6-5, 3-2, or 2-1. The odds are not calculated based on hitting a certain number "twice."

Agreed. I think my point is that there are some who believe that once a point has been rolled, the odds are heavily in the favor of rolling other numbers before that number again, because dice apparently have memory.

I think it is possible it drives the result towards the ultimate expected value to place extra come bets. My analogy is playing the daily number. Generally the legal daily number of a state pays $500 while you can bet on 1000 numbers. That stinks of course, but one analogy can be illustrated: if you bet on all 1000 numbers, you will win exactly $500. Each time exactly! Now there is a way of betting with no variance! You might as well hand over $1000 and just have the retailer give you $500 on the spot [this I have heard from someone who held people who bet on more than one number in contempt]

So there is a question in the back of my mind, do multiple bets all hanging on the same roll of the dice actually help guarantee a losing day at the table?

I can't resist making and building multiple simultaneous come bets nonetheless; to have just two going, though, seems desirable for the most part. One simple way to look at the drawback of it is that the more you bet, the more likely you are to have an outcome close to the expected value. The house sure counts on that.

Alan,
Once again, thank you for that excellent analysis on the SD of a PL/odds and a Buy 4 bet. That is the sort of data I was looking for. Between your data analysis, and the results from the simulations I am running on WinCraps, I am reevaluating my craps play strategy. However, at the same time, I have been very successful with what I have been doing the past 2 years, (obviously thanks to dumb luck), so I am reluctant to change too much. I am certainly a much more educated gambler now, which is a great thing. Thanks to all of the posters in this forum, I have a greater understanding of the game, beyond the standard HA for the PL/Place/Buy bets.
My successes and failures have seemed to have been caused more by poor bankroll management than anything else. Sticking to the smarter bets on the table helps give me a better chance of winning, but poor bankroll management when playing on house money, has let me give a lot of it back. Hmmm, this subject would make a good thread....
Gordy

Quote: teddys

[goatcabin, great analysis. What is your bettting pattern now, just one pass line with max odds? Why not just have your odds turned off on the comeout roll, then you can "root for the seven?" My dice-playing friend does this. He hates rolling the seven, though, because all his flat bets come down, and like you said, it's a lot of work to get those numbers up there.

Just testing here. My prior response to this post did not show up for some reason.
Cheers,
Alan Shank

That was my stupid mistake; I hit "preview", but then didn't hit "post". Duh!

Quote: teddys

goatcabin, great analysis. What is your bettting pattern now, just one pass line with max odds? Why not just have your odds turned off on the comeout roll, then you can "root for the seven?" My dice-playing friend does this. He hates rolling the seven, though, because all his flat bets come down, and like you said, it's a lot of work to get those numbers up there.

Thanks.
Well, you answered your own question, there. Even if the odds are "off", you still lose any established come bets on the PL comeout seven, so I can't really "root" for it.

My current betting pattern is restricted by the fact that I have retired and no longer have much income. I have a conservative pattern, like this:

minimum PL, no odds
if I get a few units ahead, I take single odds
a few more, I go to double, etc. etc., up to 3,4,5X

I'm not taking odds unless I'm ahead, IOW.

I used to start at double odds and go up.

OTOH, the last time I went to Cache Creek, I played much more aggressively, playing minimum PL with single odds and placing the rest of the inside numbers, also at minimum, so I would have four or five (if point was 4/10) numbers working for me. I played this a couple of sessions with WinCraps and did very well, which, of course, is meaningless but fun. It did not work out well the other night, however; I started out ahead but after a couple of hours I was down 1/2 my bankroll.

If I were going to play seriously, i.e. I had a substantial bankroll, I would find a casino allowing at least 10X odds, play minimum PL with double odds and progress the odds up to the max before raising the flat bet. There was a guy who used to post on rec.gambling.craps who called his system "Tough Craps"; it was a progression on odds, but a very aggressive one, much more aggressive than mine.

In either case, the idea is to buy more and more variance, with won money, without increasing the expected loss.
Cheers,
Alan Shank

Quote: odiousgambit

...
So there is a question in the back of my mind, do multiple bets all hanging on the same roll of the dice actually help guarantee a losing day at the table?
...

And that, in one short sentence, is my reluctance to accept PL/odds & CB/odds as the smartest play.
I do not dispute the mathematical arguments, and if I were to play a million rolls, that would be my best play. But then again, if I were to play one million rolls, I would certainly land near the EV, which means I LOSE. Why play?

So now that we know we are dealing with short time frames, anything can happen, so it can be expected we will have rolls that deviate greatly from the expected outcome. Therefore, IMO, I believe playing the game based solely on the HA of the various bets is not taking everything into account. But these other factors are not science or math, and since we can't quantify them, they confound most of us, me included.

This is not license to play stupidly though. Playing hardways as a way to win money is highly unlikely, although I hit the hard 4 when the guy 2 from me had $500 sitting on it. Netted him $3,500 and got me a $100 tip :) Still, I suspect he loses much more than he wins doing that.

Quote: RaleighCraps

Quote: odiousgambit

...
So there is a question in the back of my mind, do multiple bets all hanging on the same roll of the dice actually help guarantee a losing day at the table?
...

And that, in one short sentence, is my reluctance to accept PL/odds & CB/odds as the smartest play.

Not sure what you would conclude is the "smartest play", then. If you do the "Iron Cross" or place the inside or across, you still have "multiple bets all hanging on the same roll of the dice". Or, are you arguing for sticking to PL w/odds?

Quote: RaleighCraps

I do not dispute the mathematical arguments, and if I were to play a million rolls, that would be my best play. But then again, if I were to play one million rolls, I would certainly land near the EV, which means I LOSE. Why play?

So now that we know we are dealing with short time frames, anything can happen, so it can be expected we will have rolls that deviate greatly from the expected outcome. Therefore, IMO, I believe playing the game based solely on the HA of the various bets is not taking everything into account. But these other factors are not science or math, and since we can't quantify them, they confound most of us, me included.

On the contrary, the other factors, chiefly variance, but also skew, are treatable mathematically and/or via simulation. You can derive, for any consistent strategy, the distribution of possible outcomes for sessions of any length. This is very similar to knowing the probabilities of different dice combinations, the probabilities of winning/losing individual bets, etc.
Cheers,
Alan Shank

Quote: goatcabin

... are you arguing for sticking to PL w/odds?

Speaking for myself, pretty much that or limiting yourself to one or two unresolved come bets, not more. Certainly not the Iron Cross.

In an earlier thread the Wizard indicated building Come bets is not hedging. I guess the example of buying all 1000 numbers simultaneously on the "daily number" is the ultimate in hedging, and after thinking about it, I withdraw the assertion that it is similar. I stick with the assertion that it is simply gambling more against a house edge. It wound up being a thought experiment. I guess getting them all resolved simultaneously on a roll of '7' gives it a certain feel that is misleading.

Quote: goatcabin

Not sure what you would conclude is the "smartest play", then. If you do the "Iron Cross" or place the inside or across, you still have "multiple bets all hanging on the same roll of the dice". Or, are you arguing for sticking to PL w/odds?

Cheers,
Alan Shank

A few sessions with WinCraps has shown the Iron Cross to be the quickest way to give the casino all of my bankroll. It rarely, if ever, has a winning session. That has turned out to be an experiment that will never make it to the table. LOVING this WinCraps tool. Best $20 I ever spent on craps.

The following thought process is incorrect. Since I posted it, I will leave it out here, and you can all snicker at my expense. Please do so behind my back though :-)


If you have a PL and a come bet on the table, the HA for the next come bet cannot be 1.41%, because if a 7 rolls to give you a win, it will lose 2 other bets. Let's use $30 bets, so later we can place the same amount. I tried to use an old post to create the math, but I would get it wrong..... For simplicity, we will say no odds in play for now.

PL bet -- $30
8 ways to win $30
4 ways to lose $30

PL bet + this come bet (CB) -- $30
6 ways to push - win the CB, lose the PL
2 ways (11) to win $30
4 ways to lose $30

PL + CB + new CB -- $30
2 ways to win $30
10 ways to lose $30 (7 wins 30 but loses 60, and 4 craps ways)

PL + CB + CB + new CB -- $30
2 ways to win $30
4 ways to lose $30
6 ways to lose $60

So, it appears to me that the 2-1 advantage that the come out roll is supposed to have, gets negated because of previous bets that are on the table. I will leave it to you math gurus to tell me the HA. Once we have the new HA, we can then compare that to the place HA.

Now to my second area. Let's assume we will cover all the numbers.
PL + 5 place/buy bets
$30 on PL (point 5) and $30 on rest of numbers.

I hate to use a roll sequence, but it gets too hard to explain without it, so I will try to create 3 scenarios, 2 on each extreme, and 1 in middle ground.

Point is 5. Roll sequence is 6,6,10,10,8,9
Place bets win 35,35,59,59,35,42 = $265 >> $180 still at risk on table
PL + CB bets win --,30,--,30,--,-- = $60 >> $90 is on table (5,8, and 9)
----------------------------------------------------------------------------
Point is 5. Roll sequence is 6,8,9,10,6,5
place bets win 35,35,42,59,35,30 = $236 >> 150 still in play (PL is down)
PL+CB bets win --,--,--,--,30,30 = $60 >> 120 still in play (8,9,10,CB(5))
----------------------------------------------------------------------------
Point is 5. Roll is 6,8,7
place bets win 35,35,-180(4,5,6,8,9,10) = -110
PL+CB bets win --,--,-90(5,6,8) = -90
----------------------------------------------------------------------------

We can add the odds into the scenarios, but that just means bigger losses for the PL+CB and we need to make bigger bets on the place bets to equal the total amounts being bet.

My point is, because the come out roll advantage is negated with multiple come bets, and if we are playing on a 3-4-5 odds table so the place payout is the same as on a PL/FO, then by placing the bet, I get paid on the first and second roll of the number, whereas the come bet only gets paid on the second occurrence. Yes I have more money at risk in the event of a 7 out on rolls 1 or 2, but by the 3rd roll, we have evened out, assuming the same number has not rolled every time.

Raleigh: I don't want to try to go through all the various outcomes, but I think you are leaving out some of the possibilities. For example, when you have one pass line bet in play and place your first come bet, I don't think your list of possible results include winning the pass bet and having the come bet still in play on the same point number.

I think the same house advantage applies and adding come bets is equivalent to speeding up the game.

Quote: Doc

Raleigh: I don't want to try to go through all the various outcomes, but I think you are leaving out some of the possibilities. For example, when you have one pass line bet in play and place your first come bet, I don't think your list of possible results include winning the pass bet and having the come bet still in play on the same point number.

I think the same house advantage applies and adding come bets is equivalent to speeding up the game.

Doc, You are right, I did not include that, but that is because it is a neutral outcome. We both have a bet on the PL, so we both win. The CB is now locked on that same number, paying even money. I can place my $30 on the same number, and get place odds.
I don't think the HA applies, because doesn't the HA include the 8 ways to immediately win the bet? With a PL bet already in play, you only push on 6 of those 8 ways.

Perhaps I am not following you correctly. This is the way I would think about it: Suppose there were six friends standing at the crap table. Friend #1 makes a pass bet while the others watch. Once a point is set, friend #2 places a come bet. If/when both of them simultaneously have points working, friend #3 places a come bet. This continues, even to the point that is possible that all six of the friends have different points working. You can consider them to be placing anywhere from no odds to max odds, but it's probably easier to think about if they are all handling odds in the same manner.

Wouldn't you expect each of these friends individually to be experiencing the usual house advantage of a pass or come bet with whatever the odds strategy is? If so, suppose they pooled their bankrolls for the evening. Wouldn't the group as a whole be experiencing the usual house advantage on the totality of their wagers? Isn't this the same as one person placing all of these bets from his own bankroll? I think it also turns out to be equivalent to an individual just placing pass line bets with odds (no come bets) and playing long enough for the same total number of wagers to have been placed. It probably would turn out differently in an individual experiment because of the sequence of rolls that actually occur, but in terms of expectations and long term averages, I think it is the same.

Quote: Doc

Perhaps I am not following you correctly. This is the way I would think about it: Suppose there were six friends standing at the crap table. Friend #1 makes a pass bet while the others watch. Once a point is set, friend #2 places a come bet. If/when both of them simultaneously have points working, friend #3 places a come bet. This continues, even to the point that is possible that all six of the friends have different points working. You can consider them to be placing anywhere from no odds to max odds, but it's probably easier to think about if they are all handling odds in the same manner.

Wouldn't you expect each of these friends individually to be experiencing the usual house advantage of a pass or come bet with whatever the odds strategy is? If so, suppose they pooled their bankrolls for the evening. Wouldn't the group as a whole be experiencing the usual house advantage on the totality of their wagers? Isn't this the same as one person placing all of these bets from his own bankroll? I think it also turns out to be equivalent to an individual just placing pass line bets with odds (no come bets) and playing long enough for the same total number of wagers to have been placed. It probably would turn out differently in an individual experiment because of the sequence of rolls that actually occur, but in terms of expectations and long term averages, I think it is the same.

Thanks Doc. As I sat down to work out the scenario for you, I realized that what I was saying about the come bet was right, BUT, I miscalculated the effect on the place bets. Depending on how I play the place bets, I will always lose the same amount as you on the 7, but you will win the $10 Come Bet.

Quote: RaleighCraps

Thanks Doc. As I sat down to work out the scenario for you, I realized that what I was saying about the come bet was right, BUT, I miscalculated the effect on the place bets. Depending on how I play the place bets, I will always lose the same amount as you on the 7, but you will win the $10 Come Bet.

No, that is not correct. The HA on the come bet is still 1.414%. All bets are independent. Even when bets tend to offset each other, like making an Any Craps bet to hedge the comeout roll, the expected loss is ALWAYS, each HA times each bet amount. You can prove that to yourself with WinCraps. Or, try this:

bet $5 pass and $5 don't pass, so they completely negate each other except for the 12:

According to the "perfect 1980", we have:

1925 outcomes that net zero
55 outcomes that net -5

55 * -5 = -$275

The total bet handle is (1980 * 5) + (1925 * 5) = $19525
-275 / 19525 = -.0140845, which is halfway between 1.414 and 1.402, the HAs, respectively, of the pass and don't pass bets.

There is no way around this, RaleighCraps.
Cheers,
Alan Shank

Quote: goatcabin

Quote: RaleighCraps

Thanks Doc. As I sat down to work out the scenario for you, I realized that what I was saying about the come bet was right, BUT, I miscalculated the effect on the place bets. Depending on how I play the place bets, I will always lose the same amount as you on the 7, but you will win the $10 Come Bet.

No, that is not correct. The HA on the come bet is still 1.414%. All bets are independent. Even when bets tend to offset each other, like making an Any Craps bet to hedge the comeout roll, the expected loss is ALWAYS, each HA times each bet amount. You can prove that to yourself with WinCraps. Or, try this:

bet $5 pass and $5 don't pass, so they completely negate each other except for the 12:

According to the "perfect 1980", we have:

1925 outcomes that net zero
55 outcomes that net -5

55 * -5 = -$275

The total bet handle is (1980 * 5) + (1925 * 5) = $19525
-275 / 19525 = -.0140845, which is halfway between 1.414 and 1.402, the HAs, respectively, of the pass and don't pass bets.

There is no way around this, RaleighCraps.
Cheers,
Alan Shank

So what you are trying to prove to me is, each bet stands on its own merit?
IOW, the fact that CB #3 wins with a 7 is enough to say that the HA on that bet is 1.414%, and that any other bet that is in play is irrelevant?

Perhaps I am incorrectly referring to HA, when I am talking about something else.

$30 PL immediately wins 8 ways, and loses 4 ways.
You have a 8/36= 22.2% chance of winning $30.
You have a 4/36= 11.1% chance of losing $30.

But, if you add a $30 Come Bet
You have 2/36 = 5.5% chance to win $30 (11)
You have 6/36 = 16.7% chance of a push (win CB, lose the PL)
You have 4/36 = 11.1% chance of losing $30

So, how can the CB be the same as the PL bet, if the total payout affect it has is different?
Or am I not allowed to look at previous bets, since they have already been placed and are immaterial?

RaleighCraps: It is HAMMERTIME!! One CB is TOO MANY!!

A very simple illustration:

$10 table with 5X odds:

The point is 9; I Place $60 on the 6

Goat bets $10 Come; the roll is a 6: I win $70; Goat's $10 goes to the 6 and he FO bet $50 on top
and again bets $10 Come

Next roll is a 6; I win another $70; Goat wins his first $70 and the $10 Come goes to the 6 again
and so on. Anytime there is a 7 out, I will always be ahead of goat on that one number after the first hit. The bad news for me is when there is an immediate 7 out. I lose $60 and goat wins $10 but be comforted in knowing that goat's PL/FO bet was a loser with that 7 out.

tuttigym

The previous post is another example of poor analysis based on the belief that one specific sequence of possible rolls can determine what the relative merits of two betting strategies might be. As a couple of examples of alternate, specific sequences that give opposing results, consider:

(1) A series of rolls of 11, with Goat collecting $10 each roll and tutti getting nothing.
(2) A series of rolls of 9, with Goat collecting $85 each time after the first one, with tutti still getting nothing.

In the post two items above, Raleigh makes a similar if less extreme omission. In his example, he might be right if the world came to an end after the second roll, but his analysis seems to have disregarded all of the outcomes where both the original pass bet and the come bet (with its own point) are still in play, plus the possibility that the second roll matches the first one, giving a pass line winner and a come bet point still in play.

I don’t think you can draw any conclusions when only considering a portion of the possible outcomes. Or maybe I am missing something in the logic of the previous posts.

On another topic: Raleigh, I did send that off-forum email that I promised you last night. If you didn’t receive it, let me know here in the forum, and I will try to figure out what went awry.

Quote: tuttigym

RaleighCraps: It is HAMMERTIME!! One CB is TOO MANY!!

A very simple illustration:

$10 table with 5X odds:

The point is 9; I Place $60 on the 6

Goat bets $10 Come; the roll is a 6: I win $70; Goat's $10 goes to the 6 and he FO bet $50 on top
and again bets $10 Come

Next roll is a 6; I win another $70; Goat wins his first $70 and the $10 Come goes to the 6 again
and so on. Anytime there is a 7 out, I will always be ahead of goat on that one number after the first hit. The bad news for me is when there is an immediate 7 out. I lose $60 and goat wins $10 but be comforted in knowing that goat's PL/FO bet was a loser with that 7 out.

tuttigym

Not quite right tuttigym. Because goat's $10 come bet can end up on any number, in order to be equal, you have to have place money on every number.Now it is an even scenario.
So, yes, every time a number rolls, you get paid the same as goat, BUT, if a seven comes out early, you lose ALL of your place bets. Goat only loses unresolved Come bets. If Goat got to bases loaded, then you win, because, as you noted, you have wins in the bank.
I bet if you work through the numbers, you will find that as soon as 4 place bets are hit, that the place bets work out to be better than the Come Bets.

Quote: RaleighCraps

I bet if you work through the numbers, you will find that as soon as 4 place bets are hit, that the place bets work out to be better than the Come Bets.

Just be careful here about focusing on this case as a basis for saying place bets are better. In another thread I showed that you will hit 4 or fewer place bets about 2/3 of the time.

Quote: Doc

The previous post is another example of poor analysis based on the belief that one specific sequence of possible rolls can determine what the relative merits of two betting strategies might be. As a couple of examples of alternate, specific sequences that give opposing results, consider:

(1) A series of rolls of 11, with Goat collecting $10 each roll and tutti getting nothing.
(2) A series of rolls of 9, with Goat collecting $85 each time after the first one, with tutti still getting nothing.

In the post two items above, Raleigh makes a similar if less extreme omission. In his example, he might be right if the world came to an end after the second roll, but his analysis seems to have disregarded all of the outcomes where both the original pass bet and the come bet (with its own point) are still in play, plus the possibility that the second roll matches the first one, giving a pass line winner and a come bet point still in play.

I don’t think you can draw any conclusions when only considering a portion of the possible outcomes. Or maybe I am missing something in the logic of the previous posts.

On another topic: Raleigh, I did send that off-forum email that I promised you last night. If you didn’t receive it, let me know here in the forum, and I will try to figure out what went awry.

I did not get the note. I am sure you know to replace the at and the dot. I have not received any mail all day though, so perhaps my mail server is down.

Up to this point I have taken pains to not compare the Come Bet to Place Bets. I can't argue for either one yet, until I am certain I understand the merits of both bets, and at this point I don't. What I am trying to do is understand the mechanics and math behind the come bet.

Quote: RaleighCraps

$30 PL immediately wins 8 ways, and loses 4 ways.
You have a 8/36= 22.2% chance of winning $30.
You have a 4/36= 11.1% chance of losing $30.

But, if you add a $30 Come Bet
You have 2/36 = 5.5% chance to win $30 (11)
You have 6/36 = 16.7% chance of a push (win CB, lose the PL)
You have 4/36 = 11.1% chance of losing $30

What I am asking is, When a Come Bet is ADDED in addition to a PL bet, isn't the Come Bet advantage somewhat diluted because a 7 costs you a previous bet.
I am not saying anything about Place bets here.

Quote: RaleighCraps

What I am asking is, When a Come Bet is ADDED in addition to a PL bet, isn't the Come Bet advantage somewhat diluted because a 7 costs you a previous bet.
I am not saying anything about Place bets here.

I understand that you are not the one discussing place bets.

I do not think that the come bet house advantage is changed in any way by what happens to the pass bet. All of the losses associated with the pass bet and an eventual seven out are reflected in the house advantage on the pass bet -- nothing to do with the come bet.

There is, however, one thing that I am not certain about. I am convinced that a combination of a pass bet and any additional come bets has the same house advantage (as a percentage) as each individual wager. The more you wager, in one bet or a combination, the more you should expect to lose, on average. The part I am not clear on is whether combining bets on the same rolls of the dice has an effect on the variance of the outcome, compared to placing the same number of wagers on completely separate rolls of the dice.

The reason I am not clear on this (without attempting the calculation) is that I know that multiple wagers at roulette do indeed affect the variance: placing single-number wagers for 38 consecutive spins of the roulette wheel give an expected loss of 2 units, with some variance, while placing wagers on each of the 38 numbers on the same spin gives the same expected loss of 2 units but with no uncertainty. I know that a combination of the pass/come bets is not the same thing, because it is possible to win both the pass bet and the come bet whereas you can't win but one of the individual numbers on a single spin of the roulette wheel, but it seems that the possibility of losing both the pass and come bets on the same 7 out might result in a similar impact on the variance (but not affecting the expected value or house advantage).

Can someone help clear up my thinking on this?

And Raleigh: I resent the email and copied myself this time. It came through to me, but for some reason it went into my spam folder. Kind of ironic that my computer thinks that email from me might be spam! I have no idea why it did that. If you don't get a copy of either message, check your spam box.

Quote: seattledice

Quote: RaleighCraps

I bet if you work through the numbers, you will find that as soon as 4 place bets are hit, that the place bets work out to be better than the Come Bets.

Just be careful here about focusing on this case as a basis for saying place bets are better. In another thread I showed that you will hit 4 or fewer place bets about 2/3 of the time.

Correct, and I agree with you, which is why I won't argue that place bets are better than Come Bets.
They CAN be better, but you have to get more than 4 hits, which is hard to do on a repeatable basis.
However, does that mean that one should declare the PL bet and Come Bet as the BEST strategy?
At this point, I'm not sure that I agree.

Let me make a roll independent scenario, and demonstrate tutti's point:
Table is $10 , 3-4-5x , Place bets only, no buys.
We both make $10 PL bets. Point is 6 (point can be anything, btw)
(I will ignore the PL from here on out. As we are both on it, our results will always be equal)
We both put $50 odds on the PL. We will get paid $70 on a 6.
I will make Place bets that will pay the same as your CB/ full odds.
Place 4/10 for $40
Place 5/9 for $50
Place 8 for $60
-You make a $10 Come Bet.
-If an 11 rolls, you WIN $10 (this will happen 5.5% of the time)
-If a (2,3,12) rolls, you LOSE $10 (this will happen 11.1% of the time)
-If a 7 rolls, you WIN $10, I LOSE $240 (this will happen 16.7% of the time)
-If any number rolls, you have $10+(30,40,or 50 odds) on the number.
-I get paid $70.
--You make a $10 Come Bet.
--Same as above. 5.5% you win $10, 11.1% you lose $10.
--If a 7 rolls, I LOSE $240-$70(my 1 win)=$170, You LOSE (40 to 60)-10 depending on the Come point.
--If a number rolled, you have $10+(30,40,or 50) on the number
--There is also a (8.3% to 13.8%) chance you hit your previous number so you got paid $70
--I get paid $70
---You make a $10 Come Bet.
---Same as above. 5.5% you win $10, 11.1% you lose $10.
---If a 7 rolls, I LOSE $240-$140(my 2 win)=$100, You LOSE ((40 to 60)-10)x1 or x2 depending on whether or not the Come point was a repeat.
If the Come Bet did not repeat, then your loss is now slightly less than the Place Bets, or slightly higher than the Place Bets, and we are only on Come Bet #3.
If the Come Bet did repeat, then your loss is still less than half of the Place Bet.
---If the same number repeated a third time (need help with this %) you won again, you lucky stiff, so you are still positive.

I will stop here, but if you take it to the next roll, and another number hits, the place bet loss is now down to $30, which is less than ANY come bet that you will have on the table, unless you just hit the same come bet for a 4th time.

As tutti pointed out, the first time the come number is rolled, you put out $40 to $60, but we collect $70. The Price we pay for that is $240 exposed right from the start. If the shooter rolls 2 different come numbers, I have collected $140 and you now have about $100 on the table, which means we both stand to lose $100 if the 7 rolls.

-If any number rolls, you have $10+(30,40,or 50 odds) on the number.
-I get paid $70.
--You make a $10 Come Bet.
--Same as above. 5.5% you win $10, 11.1% you lose $10.
--If a 7 rolls, I LOSE $240-$70(my 1 win)=$170, You LOSE (40 to 60)-10 depending on the Come point.
--If a number rolled, you have $10+(30,40,or 50) on the number

Raleigh

A $10 bet with max odds at a 3x4x5x table always pays $70 with a win no matter what the number is, that is the power of the free odds. I am having trouble understanding your post but it seems that you are calculating your $70 win at more than the come bet win when they will be always be the same.

Quote: Doc

The previous post is another example of poor analysis based on the belief that one specific sequence of possible rolls can determine what the relative merits of two betting strategies might be. As a couple of examples of alternate, specific sequences that give opposing results, consider:

(1) A series of rolls of 11, with Goat collecting $10 each roll and tutti getting nothing.
(2) A series of rolls of 9, with Goat collecting $85 each time after the first one, with tutti still getting nothing.

In the post two items above, Raleigh makes a similar if less extreme omission. In his example, he might be right if the world came to an end after the second roll, but his analysis seems to have disregarded all of the outcomes where both the original pass bet and the come bet (with its own point) are still in play, plus the possibility that the second roll matches the first one, giving a pass line winner and a come bet point still in play.

I don’t think you can draw any conclusions when only considering a portion of the possible outcomes. Or maybe I am missing something in the logic of the previous posts.

On another topic: Raleigh, I did send that off-forum email that I promised you last night. If you didn’t receive it, let me know here in the forum, and I will try to figure out what went awry.

The scenario given was only used to show that in order to collect on ANY CB/FO, the number wagered on must be thrown twice in order to collect. It should be obvious to Raleigh and anyone else that there are dozens of possibilities, and some of them are actually good.

Why not comment on the actual validity on the two hits before collecting?

tuttigym

Quote: RaleighCraps

Quote: seattledice

Quote: RaleighCraps

I bet if you work through the numbers, you will find that as soon as 4 place bets are hit, that the place bets work out to be better than the Come Bets.

Just be careful here about focusing on this case as a basis for saying place bets are better. In another thread I showed that you will hit 4 or fewer place bets about 2/3 of the time.

Correct, and I agree with you, which is why I won't argue that place bets are better than Come Bets.
They CAN be better, but you have to get more than 4 hits, which is hard to do on a repeatable basis.
However, does that mean that one should declare the PL bet and Come Bet as the BEST strategy?
At this point, I'm not sure that I agree.

Let me make a roll independent scenario, and demonstrate tutti's point:
Table is $10 , 3-4-5x , Place bets only, no buys.
We both make $10 PL bets. Point is 6 (point can be anything, btw)
(I will ignore the PL from here on out. As we are both on it, our results will always be equal)
We both put $50 odds on the PL. We will get paid $70 on a 6.
I will make Place bets that will pay the same as your CB/ full odds.
Place 4/10 for $40
Place 5/9 for $50
Place 8 for $60
-You make a $10 Come Bet.
-If an 11 rolls, you WIN $10 (this will happen 5.5% of the time)
-If a (2,3,12) rolls, you LOSE $10 (this will happen 11.1% of the time)
-If a 7 rolls, you WIN $10, I LOSE $240 (this will happen 16.7% of the time)
-If any number rolls, you have $10+(30,40,or 50 odds) on the number.
-I get paid $70.
--You make a $10 Come Bet.
--Same as above. 5.5% you win $10, 11.1% you lose $10.
--If a 7 rolls, I LOSE $240-$70(my 1 win)=$170, You LOSE (40 to 60)-10 depending on the Come point.
--If a number rolled, you have $10+(30,40,or 50) on the number
--There is also a (8.3% to 13.8%) chance you hit your previous number so you got paid $70
--I get paid $70
---You make a $10 Come Bet.
---Same as above. 5.5% you win $10, 11.1% you lose $10.
---If a 7 rolls, I LOSE $240-$140(my 2 win)=$100, You LOSE ((40 to 60)-10)x1 or x2 depending on whether or not the Come point was a repeat.
If the Come Bet did not repeat, then your loss is now slightly less than the Place Bets, or slightly higher than the Place Bets, and we are only on Come Bet #3.
If the Come Bet did repeat, then your loss is still less than half of the Place Bet.
---If the same number repeated a third time (need help with this %) you won again, you lucky stiff, so you are still positive.

I will stop here, but if you take it to the next roll, and another number hits, the place bet loss is now down to $30, which is less than ANY come bet that you will have on the table, unless you just hit the same come bet for a 4th time.

As tutti pointed out, the first time the come number is rolled, you put out $40 to $60, but we collect $70. The Price we pay for that is $240 exposed right from the start. If the shooter rolls 2 different come numbers, I have collected $140 and you now have about $100 on the table, which means we both stand to lose $100 if the 7 rolls.

Short rolls are more prevalent than those that might create lots of CB/FO wins. Players rarely bring down or even reduce the FO bets. That would be counterproductive to the intent or goals of the FO bets. However, winning a couple of $60 Place bets would afford a player the luxury of reducing in half his Place bets to run out the point so that even an immediate 7 would not create a total loss. But players do not usually do that either because of their greed and short-sightedness. For me, winning half is alot better than winning only 10-15%.

tuttigym

Quote: kenarman

-If any number rolls, you have $10+(30,40,or 50 odds) on the number.
-I get paid $70.
--You make a $10 Come Bet.
--Same as above. 5.5% you win $10, 11.1% you lose $10.
--If a 7 rolls, I LOSE $240-$70(my 1 win)=$170, You LOSE (40 to 60)-10 depending on the Come point.
--If a number rolled, you have $10+(30,40,or 50) on the number

Raleigh

A $10 bet with max odds at a 3x4x5x table always pays $70 with a win no matter what the number is, that is the power of the free odds. I am having trouble understanding your post but it seems that you are calculating your $70 win at more than the come bet win when they will be always be the same.

We are in agreement on the payouts. It is always $70. Actually the place 4/10 pays $72, but close enough for what I am demonstrating.

Each time a number is rolled for the FIRST time, the come bet puts between $40 and $60 into play on the board (4/10=$40, 5/9=$50, 6/8=$60), whereas the Place bet will get paid $70, since we are already on that number.
The Place bet always gets paid $70. Any repeat rolls on a number will be a wash, as we both get paid $70.

Assuming a new Come Bet is put up every time, a 7 will lose the come bet that is in play, but will win $10 from the new Come Bet. I was trying to show that (40 to 60) -10 = $30 to $50 loss, depending on the come bet point number that was lost.

Win Payout is $70.
I have $240 (if point is 6/8) to $260 (if point is 4/10) to cover the board.
If only 1 new number is rolled, and then a 7 out, the come bet comes out way ahead, as they have less money exposed for a loss.
If 2 new numbers roll, I have been paid $140, so my loss exposure is now
($240 or $260)-$140 = $100 to $120
For the Come Bets, 2 new numbers means a low of $40(4/10), up to $60(6/8) is now on the board.
Since there is no win for putting the come bet up, your loss exposure is now as low as $80 (if 4&10 were rolled) to as high as $120 (if 6&8 were rolled).
So 2 new numbers means the place bets and the Come bets have almost the same loss exposure.
At this point, we have 3 of the possible 6 numbers in play (PL, CB1, CB2)

The next new number come bet that goes into play, pays the Place Bets another $70.
The Place bet loss exposure is now
($240 to $260)-($70x3 wins)= $30 to $50
The come bets now have 3 points in play ($40 to $60)x 3 bets, which is greater than the place ($30 to $50) exposure.

Again, this is valid, because, on the FIRST roll of the number, The Place Bets get paid, Come Bets have to put odds money on the table.

What if come points got hit along the way? They become a wash, because we both get paid $70.

Quote: tuttigym

Short rolls are more prevalent than those that might create lots of CB/FO wins. Players rarely bring down or even reduce the FO bets. That would be counterproductive to the intent or goals of the FO bets. However, winning a couple of $60 Place bets would afford a player the luxury of reducing in half his Place bets to run out the point so that even an immediate 7 would not create a total loss. But players do not usually do that either because of their greed and short-sightedness. For me, winning half is alot better than winning only 10-15%.

tuttigym

We are all different and we all play with different approaches and take different risks based on what we feel comforatble doing and what our expected reward will be. Therefore, what one person feels is the best strategy for him may not be the best strategy for others. That doesn't mean it's wrong - just different.

I think the key is to completely understand how the particular betting strategy one employs will work in ALL possible outcomes. You shouldn't consider only the scenarios that work, but what will happen if the dice do exactly the worst thing for that strategy.

A lot of the feedback given on this and other threads has been to try to point out where a poster has ignored possible outcomes.

----

As I re-read your post, this just caught my eye, "For me, winning half is alot better than winning only 10-15%." Half is 50%, so, I'd have to agree with you there, but I don't think that is what you meant. You were talking about reducing place bets after a couple of wins, so I assume you meant winning half as much if the place bets hit again, but I'm not sure what the 10-15% is.

Agree. If you like to place come bets with odds, what is the most you are willing to lose? For me, it's $50. So on a $10 table, that's one come bet. On a $5 table, it's two. If the bets start to hit, I add more come bets.

Quote: RaleighCraps

....

What if come points got hit along the way? They become a wash, because we both get paid $70.

Except your calculation of the come bettor's exposure is off. If there is one repeat, then there is one less CB at risk for the same win. This type of calculation can get complicated very quickly -- for X CB, there will be repeats Y% of the time, and new CB Z% of the time, which then changes the number of CB and the % of repeats changes ... aargh. Other than raw calculations of HA and variance (which others have done) you can run simulations which, if long enough, will exercise all possible outcomes. (Actually, I guess I shouldn't say "all", but pretty darn close. :-)

Raleigh Thank You for your explanation and I understand your post now and don't disagree with that scenerio. What I think makes up the difference and brings the odds back into line for the come better (which I am) is all the times the shooter sevens out when you have all the numbers placed and the come bettor has less than the 5 other numbers working. The come bettor takes a lot smaller hit in those scenerios.

Quote: seattledice

Quote: RaleighCraps

....

What if come points got hit along the way? They become a wash, because we both get paid $70.

Except your calculation of the come bettor's exposure is off. If there is one repeat, then there is one less CB at risk for the same win. This type of calculation can get complicated very quickly -- for X CB, there will be repeats Y% of the time, and new CB Z% of the time, which then changes the number of CB and the % of repeats changes ... aargh. Other than raw calculations of HA and variance (which others have done) you can run simulations which, if long enough, will exercise all possible outcomes. (Actually, I guess I shouldn't say "all", but pretty darn close. :-)

Yes, it does get very confusing. I was trying to make your point above by indicating NEW come bets, ie not a repeat point.
I won't repeat all the numbers again, but if you work through a few roll sequences, you will see that when the 3rd NEW come bet number has been added on the board, the come bet player now has more to lose when the 7 comes out, then the place bettor does.
This is true, no matter how many come bets were paid off by repeat throws.

This is an example of what I mean:
In the following, risk of loss refers to money that is on the table when a 7 out is rolled, where we have not collected anything for that particular bet.
I also assume a new CB is being put up every roll.

PL point is 6, CB1 is on 8, CB2 is on 5, CB3 is on 10.
All repeat CBs that have been PAID prior to this are a wash, because we both got paid $70 each and every time.

In the case above the CB now has at risk of loss ($60(PL)+$60(8)+$50(5)+$40(10))=$210
Since the point is 6, The place bettor spread $240 across ($40,$50,-,$60,$50,$40)
The Place Bettor has the at risk of loss ($60(PL)+$240-$212)=$88
I deduct the ($70(8)+$70(5)+$72(10))=$212, because that is the amount the Place Bettor got paid when the Come Bettor was having to put odds out there on the first roll of the number.That number is now resolved for the place bettor, but not yet for the Come bettor.

Does this mean the Place bet is the better strategy? I'm not saying that.
In order to get this advantage after 3 CB are in play, I had to put at least $240 at risk of an early 7 out. With a 7 out occurring early over 20% of the time, the Place bet will lose more than the PL + CB strategy, because as you say, I can't overcome math.
However, every time the CB player gets to the point where their 3rd CB is in play on the felt, they have reached the point where they would have been better off if they had placed the bets.
They had less at risk early on, but they will end up with less of a win than the place bettor, once 3 CB are in play on the felt.

Rayleigh without getting into your math, after a pre-dinner martini and a bottle of wine with dinner, which would be much make it too imposing right now I think you could be right after 3 new numbers have hit. I am usually much too impatient at the table to wait for all the come numbers too hit and almost always place the 6 and 8 often even have them working for the come out so will have 3 numbers working immediately if the point isn't a 6 or 8. Which probably makes your point even stronger.

Quote: RaleighCraps

Quote: seattledice

Quote: RaleighCraps

....

What if come points got hit along the way? They become a wash, because we both get paid $70.

Except your calculation of the come bettor's exposure is off. If there is one repeat, then there is one less CB at risk for the same win. This type of calculation can get complicated very quickly -- for X CB, there will be repeats Y% of the time, and new CB Z% of the time, which then changes the number of CB and the % of repeats changes ... aargh. Other than raw calculations of HA and variance (which others have done) you can run simulations which, if long enough, will exercise all possible outcomes. (Actually, I guess I shouldn't say "all", but pretty darn close. :-)

Yes, it does get very confusing. I was trying to make your point above by indicating NEW come bets, ie not a repeat point.
I won't repeat all the numbers again, but if you work through a few roll sequences, you will see that when the 3rd NEW come bet number has been added on the board, the come bet player now has more to lose when the 7 comes out, then the place bettor does.
This is true, no matter how many come bets were paid off by repeat throws.

....

Does this mean the Place bet is the better strategy? I'm not saying that.
In order to get this advantage after 3 CB are in play, I had to put at least $240 at risk of an early 7 out. With a 7 out occurring early over 20% of the time, the Place bet will lose more than the PL + CB strategy, because as you say, I can't overcome math.
However, every time the CB player gets to the point where their 3rd CB is in play on the felt, they have reached the point where they would have been better off if they had placed the bets.
They had less at risk early on, but they will end up with less of a win than the place bettor, once 3 CB are in play on the felt.

OK - I see what you are getting at. I'm also not saying that putting up new come bets on every throw is better either -- they are different strategies with different HA, variance, and probably bank roll requirements. You just have to really know what you are getting yourself into, and I believe that with the calculations you are making and the simulations you are running, you really do get it.

Just for fun, I created a WinCraps autobet file that takes as input the maximum number of come bets to allow at a time. I set it up to stop betting at 200 rolls, wait for any bets to be resolved and end the session. I ran 2000 sessions for each input 1-6. I had discovered before that, although a come bet is made only 2/3 of the time, when the passline comeout establishes a point, the number of come bets resolved is more like 80% of the number of passline bets. This is because sometimes the passline bet gets resolved and there's still a come bet outstanding. On the next pass comeout, the come bet may be resolved and another come bet put up.

max comes	1	2	3	4	5	6
come/pass	.82	1.45	1.91	2.21	2.38	2.44

The numbers are fairly close to .8 + .8^2 + .8^3 + ... .8^n.

So, basically, if you make a come bet on every non-comeout roll, your average total of flat bets will be close to 3 1/2 times your basic amount. So, if you are a $10 passline bettor, and you make 60 bets, your total bet handle would be close to $2060, yielding an expected loss of about $29. That's without taking any odds, of course. If you took odds, the proportions would remain the same, assuming you always worked your come odds. Taking the odds would not increase the expected loss, just the volatility, risk of ruin, chance of coming out ahead and chance of a big win.

I will run a sim comparing these strategies:

Guy #1
wait for point
buy 4 and 10 for $30 each (best possible conditions)
buy 5 and 9 for $30 each (best possible conditions)
place 6 and 8 for $30 each

The sessions will be 200 rolls (i.e. no more bets after 200 rolls, resolve outstanding bets). Here is what we expect:

buy 4/10 resolved every 4 rolls, so 50 * $60 = $3000 * -.0111 = -$33.33
buy 5/9 resolved every 3.6 rolls, so 55.55 * $60 = $3333 * -.0133 = -$44.33
place 6/8 resolved every 3.27 rolls, so 61.1 * $60 = $3667 * -.015 = -$55.05
Total approximate bet handle is $10,000, expected loss $133.

Guy #2
bets $20 pass, double odds and $20 come, double odds every chance he gets.
Here is what we expect:

3.44 * 20 * 60 = $4128 * -.01414 = -$58.36
3.44 * 40 * 40 = $5504 * 0 = $0
Total approximate bet handle is $9632, expected loss $58.

I will use the same seed for both. All bets are left up until resolved.
Cheers,
Alan Shank

In my prior post I overestimated the number of resolved place/buy bets, because I overlooked them not working on the comeout, so the amount bet in the $180 across scenario is going to be considerably less. That's OK.

I went ahead and set up two WinCraps autobet files. I set the bankrolls at $1500, and considered it a bust if the player did not have at least $180 left, ending the session regardless of the number of rolls completed. Winning and losing bets were replaced.

To review, the "Two Guys" bet this way:

Guy #1
wait for point
buy 4 and 10 for $30 each (best possible conditions)
buy 5 and 9 for $30 each (best possible conditions)
place 6 and 8 for $30 each

Guy #2
bets $20 pass, double odds and $20 come, double odds every chance he gets.

No more betting after 200 rolls. Guy #1 quits on next comeout roll, taking any outstanding bets down. Guy #2 waits for all bets to be resolved; odds work at all times.

parameter	Guy #1	Guy #2
avg. num. rolls	189	193
avg. num. bets	222	319
avg. bet handle	6695	8931
mean net result	-$96	-$55
median net result	-$107	-$136
mode of net result	-$1401 to -$1351	-$1372 to -$1312
standard deviation	$941	$1037
avg. house advantage	1.43%	0.62%
winning sessions	4604	4507
breakeven sessions	0	12
losing sessions	5396	5481
number of busts	1854	1810
lost more than $1000	2209	2280
lost more than $500	3530	3780
won more than $500	2703	2879
won more than $1000	1316	1635
won more than $1500	512	813
won more than $2000	161	343
biggest win	$3908	$5408

So, despite a bet handle of a third more than Guy #1, Guy #2 loses less than 60% of what Guy #1 does. Of course, the reason is that only about 43% of Guy #2's bet handle is subject to a house edge, while all of Guy #1's is. Due to having more money out there, up to $360 at a time, Guy #2 is somewhat more likely to lose $500 or more, although Guy #1 had a few more busts. Guy #2 did have a few more winning sessions, 97 more of 10,000. As the amount won went up, Guy #2 had more and more of those.

Just for more fun, I ran Guy #2 again just taking single odds. He didn't have as many winning sessions, 4453, his mean bet handle was $6740, his mean loss was $55, median $90 and mode was -$272 to -$232, SD $668. His bust rate, no surprisingly, was much lower, just 261 sessions, but his wins were smaller, as well. None of this should be surprising.
Cheers,
Alan Shank

I missed your post Alan, as I was in Biloxi getting the snot kicked out of my place bets.
$1500 bankroll, betting $180 across, is too little. From my experience playing with WinCraps, $3000 is the bankroll required to avoid busting out so many times. If you still have the seed it would be interesting to see what the change is with a higher bankroll.
I guess numbers repeat during a roll more often than I realize, although when I think about the number of place bets that get pressed, there are quite a few repeats.

Quote: RaleighCraps

I missed your post Alan, as I was in Biloxi getting the snot kicked out of my place bets.
$1500 bankroll, betting $180 across, is too little. From my experience playing with WinCraps, $3000 is the bankroll required to avoid busting out so many times. If you still have the seed it would be interesting to see what the change is with a higher bankroll.
I guess numbers repeat during a roll more often than I realize, although when I think about the number of place bets that get pressed, there are quite a few repeats.

Per Raleigh's request I set the bankrolls at $3000, and considered it a bust if the player did not have at least $180 left, ending the session regardless of the number of rolls completed. Winning and losing bets were replaced.

To review, the "Two Guys" bet this way:

Guy #1
wait for point
buy 4 and 10 for $30 each (best possible conditions)
buy 5 and 9 for $30 each (best possible conditions)
place 6 and 8 for $30 each

Guy #2
bets $20 pass, double odds and $20 come, double odds every chance he gets.

No more betting after 200 rolls. Guy #1 quits on next comeout roll, taking any outstanding bets down. Guy #2 waits for all bets to be resolved; odds work at all times.

I also added a third "Guy", who limited himself to three come bets at a time, which brought his bet handle down closer to Guy #1's. Here's how they came out:

parameter	Guy #1	Guy #2	Guy #3
avg. num. rolls	202	205	205
avg. num. bets	237	338	286
avg. bet handle	7177	9476	8014
mean net result	-$101	-$56	-$48
median net result	-$95	-$116	-$72
mode of net result	+104 to +$166	-$328 to -$256	-$164 to -$108
standard deviation	$981	$1086	$925
avg. house advantage	1.41%	0.59%	0.60%
winning sessions	4650	4538	4678
breakeven sessions	4	20	14
losing sessions	5346	5442	5308
number of busts	59	18	6
lost more than $1000	1784	1919	1507
lost more than $500	3420	3589	3185
won more than $500	2734	2868	2717
won more than $1000	1322	1593	1307
won more than $1500	483	843	516
won more than $2000	142	387	175
biggest win	$3646	$5248	$3468

You pays your money and you takes your choice! As far as I'm concerned, the key issue is that Guy #1 basically starts out about $50 behind the other Guys, the difference in the mean outcomes.
Cheers,
Alan Shank

Quote: goatcabin

You pays your money and you takes your choice! As far as I'm concerned, the key issue is that Guy #1 basically starts out about $50 behind the other Guys, the difference in the mean outcomes.
Cheers,
Alan Shank

Yeah, it is pretty hard to argue against those results.
I'm still stuck trying to reason out the results though. With a PL and 2 come bets, both guy 1 and guy 3 are getting the same amount of bet money onto the table, although guy 1 is more at risk of the early 7 out. Also, guy 3 will get paid way better at $20w40odds than the $30 place. I guess that is where the difference lies. I'm also surprised that guy 2 doesn't take a bit more of a beating, come betting every roll.

I'm pretty well convinced that the come bet strategy can provide superior results to a PL w Place Bet strategy.
The last piece of the puzzle now is the rating/comps factor. Since odds don't factor into your avg bet/hour rating, the Come bet strategy will result in a significantly lower rating. Guy 3 would be rated on the $60 in PL/Come bets, whereas, Guy 1 gets rated on the full $180. I don't think the $60 is going to get the two flights on the junkets, room, and most of the time the meals, all of which adds up to a decent value. Of course, if I was having to pay for the flights and room I would not be going at all, so I'm not sure how to really factor that in. One could argue that the Come bet strategy will keep your bankroll intact longer, meaning longer craps play, but I suspect that both strategies will die equally fast on 'cold' tables, so that really doesn't help for those trips where you get wiped out early.

I guess the best strategy may be to start playing w Place bets to get that initial rating at $180 or so, and then sneak in a few rounds of come bets, and then back to place bets. Of course, since the box is doing the rating, you're probably not going to fool them too much alternating back and forth. I guess it all depends on how accurate the box feels like making the numbers.

So what is the perfect answer? Three point Molly?

Quote: DeMango

So what is the perfect answer? Three point Molly?

Perfect answer for whom? I don't believe in any "one-size-fits-all" answer. For me, the biggest factor is getting variance for free by taking/laying odds. Any money you don't bet on the flat bets, but do bet on the odds, saves you expected loss while buying you a bit more variance. (The SD of passline money is .9999 of the bet, while the weighted SD on the odds is 1.22 of the bet.) Another factor to consider is that, for a given bet handle, you get more variance by making fewer, larger bets than more, smaller bets. See my blog for today for an expansion on that.
Cheers,
Alan Shank

Wouldn't you be better off placing the numbers instead of wasting a roll to get your come bet up there?