Three Point Dolly

I think Vic's play is a real word example why players are better off making table minimum place bets and then pressing if they start hitting repeaters.

Making come bets with big odds is just risking too much money. Worse is that you need to roll a number twice to collect.

Come betting works only when you have a long roll and my crystal ball was smashed in the Northridge earthquake.

Quote: Ace2

Tuttigym,

The numbers are:

Average units bet per session = 30.313926
Average loss per session = .08315 units

.08315 / 30.313926 is an average loss of 0.27%, which is the overall house edge when betting don't pass with 3-4-5 odds. So the expected loss on $1000 wagered is $2.70

This is actually 5th grade level arithmetic so don't even bother trying to understand
link to original post

I want to assure you that I won't "bother to understand" how one can lose $2.70/$1,000 wagered because it absolutely CANNOT happen at the tables. I also know that you or anyone else cannot PROVE or PERFORM the precision (1 millionth of 1%) "average loss per session." Especially since the calculations are a result of computer SIMULATIONS which are "only done when something is IMPOSSIBLE or TOO DIFFICULT to calculate directly." (Ace2 5-10-22; Three Point Mollie pg.24)

Do you suppose Mr. W. will answer my questions directly or will your post satisfy?

tuttigym

If you were supposed to win on the 12 on the Don't line instead of pushing, the HA is taking your win right there.

Quote: Ace2

This is actually 5th grade level arithmetic so don't even bother trying to understand
link to original post

That was a low blow. Warning issued for making a personal insult.

My post upstream was not well worded and probably caused his confusion.

tuttigym,

Below is a genuine response. Please don’t read it as condescending or snide in any way, as that honestly is not my intention.

You are correct that one cannot be down by exactly $2.70 after any amount of betting, as this amount would be smaller than any payout, and thus impossible. This ($2.70/$1000) is a ratio and isn’t intended to be taken as a literal outcome. However, since it is a ratio, one could inflate both numbers equally until one sees numbers that make sense, such as losing roughly $270.00 (yes, I rounded) for every $100,000.00 in total bets, or losing roughly $2,700.00 for every $1,000,000.00 in bets. Can you see how I arrived at those numbers? First, I multiplied both figures by 100, and next, I multiplied both by 1,000. I could use any number that makes the smaller number a multiple of the bet or payout amounts, and as long as I apply I that same figure both numbers, the ratio remains intact. Furthermore, this is a valid way to look at such numbers, as the bigger they both get, the closer one’s real life results should get (Law of Large Numbers from the field of statistics).

The reason that we use the smaller numbers is that we are trying to view long term results in way that demonstrates how they average out during a series of shorter sessions. You may still find these figures impractical or difficult to understand. That doesn’t make them useless to the majority of the rest of us. If you would like help understanding figures presented like the “$2.70 for every $1,000” please just ask. I think you’ll get a better response than shooting it down and demanding a different analysis.

Sincerely,
camapl

Quote: tuttigym

I want to assure you that I won't "bother to understand" how one can lose $2.70/$1,000 wagered because it absolutely CANNOT happen at the tables. I also know that you or anyone else cannot PROVE or PERFORM the precision (1 millionth of 1%) "average loss per session." Especially since the calculations are a result of computer SIMULATIONS which are "only done when something is IMPOSSIBLE or TOO DIFFICULT to calculate directly." (Ace2 5-10-22; Three Point Mollie pg.24)

Do you suppose Mr. W. will answer my questions directly or will your post satisfy?

tuttigym
link to original post

I just read that a US family has an average of 1.93 children. Sounds reasonable, but according to your logic that figure must be wrong since it's impossible to have 1.93 children. Averages are averages

We don't blindly rely on simulations. Usually we have a good idea of what the answer should be (reasonable range) but might use a simulation to confirm a number or get a higher level of precision. For the 3 point molly, I easily calculated the increase in bet volume relative to a single PL bet (increases by a factor of 2.4). However I did use a simulation to get the increase in standard deviation (increases by a factor of 1.91) since I could not come up with a way to directly calculate that number. You could estimate the increase as 2.4^.5 = 1.55, but the actual increase is higher since the results of the 3PM bets are correlated, especially when there's a seven-out with three numbers covered. Since the increase in SD could not be more than 2.4, we have a range of 1.55 - 2.4, so 1.91 seems quite reasonable.

How many trials are needed in a simulation varies. Yes, it does sometimes take millions of trials (or more) before the answer converges to three significant digits, which is usually enough. Not sure why you continue referencing my prior statement about simulations... it's like you think it invalidates my comments

Quote: AlanMendelson

Making come bets with big odds is just risking too much money. Worse is that you need to roll a number twice to collect.
link to original post

I never understood this.

If you compare someone who always covers all the numbers with place bets to a someone who always makes come bets, the latter will take much longer to get multiple numbers covered but his average loss will be much less when the seven-out happens. And the latter will come out far ahead of the former since the edge on come bets is much lower than the edge on place bets (and even much lower when 3-4-5 odds are taken)

Quote: Ace2

Quote: AlanMendelson

Making come bets with big odds is just risking too much money. Worse is that you need to roll a number twice to collect.
link to original post

I never understood this.

If you compare someone who always covers all the numbers with place bets to a someone who always makes come bets, the latter will take much longer to get multiple numbers covered but his average loss will be much less when the seven-out happens. And the latter will come out far ahead of the former since the edge on come bets is much lower than the edge on place bets (and even much lower when 3-4-5 odds are taken)
link to original post

other players and the dealers constantly make the comment that the problem is that you have to roll the number twice, and they do so sincerely

Quote: Ace2

Quote: AlanMendelson

Making come bets with big odds is just risking too much money. Worse is that you need to roll a number twice to collect.
link to original post

I never understood this.

If you compare someone who always covers all the numbers with place bets to a someone who always makes come bets, the latter will take much longer to get multiple numbers covered but his average loss will be much less when the seven-out happens. And the latter will come out far ahead of the former since the edge on come bets is much lower than the edge on place bets (and even much lower when 3-4-5 odds are taken)
link to original post

No comps for odds bets, man.

Quote: odiousgambit

Quote: Ace2

Quote: AlanMendelson

Making come bets with big odds is just risking too much money. Worse is that you need to roll a number twice to collect.
link to original post

I never understood this.

If you compare someone who always covers all the numbers with place bets to a someone who always makes come bets, the latter will take much longer to get multiple numbers covered but his average loss will be much less when the seven-out happens. And the latter will come out far ahead of the former since the edge on come bets is much lower than the edge on place bets (and even much lower when 3-4-5 odds are taken)
link to original post

other players and the dealers constantly make the comment that the problem is that you have to roll the number twice, and they do so sincerely
link to original post

Thing is, you have to roll a number and a seven to lose that number with a PL/come bet. But with a place/buy bet it only takes a seven to lose it. So I guess you could say you win twice as often with a place/buy bet, but you also lose twice as often...