dalsanto3
Joined: Oct 13, 2018
• Posts: 27
January 30th, 2019 at 8:43:49 PM permalink
I made a craps odds calculator designed to calculate expected loss per hour of play. It is based on 100 rolls per hour.

I've opened up editing to anyone if people want to put in other bets in there or download it or whatever.

It will not allow me to post the link because I am a new member: if someone could repost this as a hyperlink by removing the spaces below I would very much appreciate it.

I thought this might be helpful to some here, no idea if something similar has been done before. I did every numbers bet, passline, odds, and a couple of representative inside bets.

Any feedback is welcome, especially if someone thinks I'm grossly miscalculating something. Hope it helps someone out.

EDIT: After working on this for a little bit, I am second guessing my initial numbers on resolved bets per hour.

I started with 30.6 for 6/8, 27.8 for 5/9 and 25 for 4/10. This is just a divisor for the dice combinations (11/36; 10/36; 9/36). For example, there are 5 sixes and 6 sevens that resolve the six place bet; that's 11 out of 36 combinations to resolve the bet, so that's .306 resolved bets and .694 pushes. But then I realized that you need to incorporate the comeout roll. Clearly, over any given hour, a certain percentage of the time dice are coming out and numbers bets are off. Say, for example, you're coming out on average 5% of the time, for 6/8 you'd multiply .95 * .306 which would actually in 29 resolved bets per hour instead of 31.

At the least, the dice are coming out 6/36 rolls, the number of combinations of 7. During that next roll, no numbers bets resolve. Which means that on average numbers bets are in play in 30/36 rolls, or 5/6th of the time. I would compound the resolution of the numbers bets with the amount of times the numbers bets are actually in play for total resolved bets per hour. So 6/8 would actually be (11/36)*(30/36) = 25.5 resolutions per hour; 5/9 are (10/36)*(30/36) resolutions per hour= 23.1; 4/10 are (9/36)*(30/36) = 20.8 resolutions per hour.

I think those numbers are closer to the correct numbers, but I don't believe I'm properly accounting for the chance that on a comeout roll, there's 12 combinations that force a second comeout roll and 24 combinations that are pushes. If anyone is better at this than me - which I'm sure someone is - I would appreciate an attempt to compound these probabilities properly.

The second thing that may not be properly accounted for is the resolved bets per hour on pass / don't pass. I started with 30, which I just copied from another site, but I don't think that's right. A point is made on 24/36 of every comeout roll with a 4, 5, 6, 8, 9, 10; the bet resolves 12/36 of the time on the comeout with a 7, 11, 2, 3, 12. When a point is established, depending on what point it is, there's a different amount of resolutions per hour. When a point is established, there's a 25% chance it's a 4/10, a 33% chance it's a 5/9, and a 42% chance that the point is 6/8. If, for example, the point 6/8, then it will resolve 30.6% of the time as I discussed above, but that would change for the other points. The question is out of 100 rolls how many times does a pass line bet resolve? So after 12/36 rolls (a previous 7, 11, 2, 3, or 12), the passline bet resolves on 12/36 combinations once again and is a push on the remaining 24 combinations. After 10/36 rolls (6/8), the passline bet resolves 16/36 times (6/7/8); after 8/36 rolls (5/9), the passline bet resolves 14/36 times (5/7/9); after 6/36 rolls (4/10), the passline resolves 12/36 times (4/7/10). So (12/36)*(12/36)+(10/36)*(16/36)+(8/36)*(14/36)+(6/36)*(12/36) = 1/9+(5/18)(4/9)+(1/4)(7/18)+(1/6)(1/3) = .11111 + .12346 + .09722 + .055556 = 38.73 percent of rolls verses the 30 that I have in my spreadsheet.

This seems consistent with experience playing the game. Obviously, if you have \$20 on the passline for an hour and \$60 on an odds bet for an hour, that odds bet is going to resolve many fewer times over the course of that hour than the passline bet.

Does anyone have any thoughts on this or have some experience compounding probabilities? Or are these numbers out there and known?
Last edited by: dalsanto3 on Jan 31, 2019
mayoroftitycity
Joined: Jan 31, 2019