docsjs
docsjs
  • Threads: 22
  • Posts: 62
Joined: Oct 31, 2009
February 22nd, 2010 at 12:27:21 PM permalink
I am a place better. What is a good progression for increasing your place bets to take advantage as a roll gets longer?
DJTeddyBear
DJTeddyBear
  • Threads: 210
  • Posts: 11060
Joined: Nov 2, 2009
February 22nd, 2010 at 12:42:08 PM permalink
I'm a place bettor too.

Typically, I start out only placing the 6 & 8. I'd bet the minimum, let it hit a few times, then, once I'm profitable, start pressing it one unit at a time, almost every time they hit. When they get to $24 or $30, I might play the 5 & 9 rather than press, and get them going the same way. I also pause when they get to $42. There's something sexy about dropping a white and getting two greens back!


I've been toying with a different launch idea. My concept revolves around the belief / hope that a typical player will hit the other numbers once before a seven-out.

Bet more than the minimum, then DROP to the minimum after he hits, THEN start pressing if he hits it again. It brings you into profit sooner, so you can start pressing sooner. I.E. Bet $18 each on the 6 & 8 When one hits, drop them both to $6, and you're already ahead $9.

Yeah, I know often a player establishes a point and hits the 7 on the next roll, but, that's the concept.
I invented a few casino games. Info: http://www.DaveMillerGaming.com/ ————————————————————————————————————— Superstitions are silly, childish, irrational rituals, born out of fear of the unknown. But how much does it cost to knock on wood? 😁
docsjs
docsjs
  • Threads: 22
  • Posts: 62
Joined: Oct 31, 2009
February 22nd, 2010 at 1:17:52 PM permalink
It is not my style to regress. I will stop at a level for one hit to collect, but never decrease the bet. An example is 6 or 8 place 18 - 30 -42 - 42 - 60 - 60 -90 -90 etc..

5 or 9 place 15 -20-25 - 35 - 50 - 50 -75 etc.
goatcabin
goatcabin
  • Threads: 4
  • Posts: 665
Joined: Feb 13, 2010
February 22nd, 2010 at 5:00:56 PM permalink
Quote: docsjs

I am a place better. What is a good progression for increasing your place bets to take advantage as a roll gets longer?



What's "good"? When you progress bets, you risk losing back what you've already won, in return for a chance to win more if you win again. The same principle applies after every win. Do I want to risk what I've won in order to win more if I win again? Progressive betting results in more losing sessions, but bigger wins, compared to flat betting. Why? Because you have to win bets in a row. In flat betting, the only thing that determines your net win/loss is how many bets you win, i.e. your win-loss record; the order of the wins and losses doesn't matter. When you progress your bets, however, a WLWLWLWLWL pattern causes you to lose instead of breaking even (or winning a bit if you're getting 7 to 6, like place 6/8. The more you progress your bets, the more dependent you become on streaks of wins. Take a pure parlay, for example; if you keep doubling your bet, then you always lose the whole amount when you lose, until and unless you get up to the table maximum and cannot double your bet. Of course, nobody bets that way, but the principle is the same, just a difference of degree.

The more you progress your bets, the greater the "skew" of the results becomes, because you have fewer, but bigger wins. Try doing something like this:

flat $6 bet on 6
win two, then up one unit
up one unit every time
$18 6, then regress to $6, win two then up one

Figure out the net result for different numbers of wins, from zero to five or so, before losing the next bet. See which method does best at each number of wins. Then figure out the probability of each number of wins, because that's an important factor.

wins flat 2 then press press regress
0 -6 -6 -6 -18
1 +1 +1 -5 +15
2 +8 +2 +3 +22
etc., etc.

Probability of losing first bet .5454, probability of WL = .4545 * .5454 = .2479, etc., etc.

The more wins in a row before losing, the better an aggressive strategy will do, but the lower the probability of that happening. Overall and in the long run, the more you bet the more you are expected to lose, and the expected value is always edge * action, but the distributions of your results will be quite different.

What is "good" boils down to how much you want to win and how much you are willing to risk to win it.
Cheers,
Alan Shank
Cheers, Alan Shank "How's that for a squabble, Pugh?" Peter Boyle as Mister Moon in "Yellowbeard"
Phosphorous
Phosphorous
  • Threads: 4
  • Posts: 14
Joined: Oct 31, 2009
March 11th, 2010 at 8:05:06 PM permalink
I typically start with $18 on 6 & 8, then progress to 24, 30, 42, 60, 90, 120, 150, 180, 210... I limit the increments to $30 max to recap max profits when the table it hot.

For the other numbers it's typically 15, 25, 35, 50, 75, 100... max increment $25.

I know it's agressive... but I'm playing for the opportunities where the shooter rolls for 30 minutes or more. And often times when they roll that long, there's one number they hit over and over again, so it really pays off then. There's nothing better than getting black chips on ever roll of the dice...
goatcabin
goatcabin
  • Threads: 4
  • Posts: 665
Joined: Feb 13, 2010
March 14th, 2010 at 7:14:00 PM permalink
Quote: Phosphorous

I typically start with $18 on 6 & 8, then progress to 24, 30, 42, 60, 90, 120, 150, 180, 210... I limit the increments to $30 max to recap max profits when the table it hot.

For the other numbers it's typically 15, 25, 35, 50, 75, 100... max increment $25.

I know it's agressive... but I'm playing for the opportunities where the shooter rolls for 30 minutes or more. And often times when they roll that long, there's one number they hit over and over again, so it really pays off then. There's nothing better than getting black chips on ever roll of the dice...



To illustrate the "continuum of progressiveness", so to speak, I wrote a computer program to compare some different betting strategies for placing the 6. We will call them "the four guys". Everyone starts at $18. Guy #1 then regresses twice before going back up, then regresses again. Guy #2 increases his bets a unit at a time, but only every other bet. Guy #3 progresses a unit at a time every bet. Guy #4 doubles first and then adds $18 each bet. So, their bet series are:
Guy #1 18,12,6,12,18,24,18,12,6,12,18
Guy #2 18,18,24,24,30,30,36,36,42,42,48
Guy #3 18,24,30,36,42,48,54,60,66,72,78
Guy #4 18,36,54,72,90,108,126,144,162,180,198

My program examines the net outcome for each of these "Guys" after different numbers of hits, always assuming that the next bet is lost. So, they all lose $18 if they lose the first bet. If they win the first, then lose, only Guy #1 has a profit, since he locked up his initial win. Here is a chart of all four Guys' net results for zero through 10 hits, along with the probabilities of each number of hits:

num hitsguy 1guy 2guy 3guy 4prob
0-18-18-18-180.545455
193-3-150.247934
229181990.112697
3304648540.051226
43868841200.023285
5531031272070.010584
6871321773150.004811
71141742344440.002187
81342102985940.000994
91352593697650.000452
101433024479570.000205


"Mr. Conservative" is still ahead after two hits, but then falls behind fast. It only takes Guy #4 three hits to take the overall lead, but he doesn't net a profit until two hits, so he's losing money almost 80% of the time.

The program also figured the weighted net outcome for each guy for zero through ten hits:

Guy 0 weighted net is -0.473997
Guy 1 weighted net is -0.674662
Guy 2 weighted net is -0.835687
Guy 3 weighted net is -1.349305

This simply reflects the fact that the expectation is always edge * action; it also favors the lower bettors because it does not consider streaks longer than ten, which can occur.

Within the next few days, I will also run a few thousand two-hour sessions for each "Guy" and examine the differences in variance among them.
Cheers,
Alan Shank
Cheers, Alan Shank "How's that for a squabble, Pugh?" Peter Boyle as Mister Moon in "Yellowbeard"
RaleighCraps
RaleighCraps
  • Threads: 79
  • Posts: 2501
Joined: Feb 20, 2010
March 15th, 2010 at 6:06:33 AM permalink
I like to try and minimize against the early 7 out, but get to a decent bet size as fast as possible.

Here is one scheme I have used. Assume the point is 6.
I will bet $80 inside ($25 on 5&9, $30 on 8).
If one of the numbers hits, I put $45 down and say press all the insides.
I now have $50--point--$60--$50 (5--6--8--9)
If one of them hits again, I put $10 on the table and say press all the insides.
I now have $75--point--$90--$75.
My total outlay so far is $80+$45+$10 = $135, and 2 inside numbers have hit.
I will get paid $105 if any number rolls again, so 2 more hits and I am plus $75.
-----------------------------------------------------------------------------------

This is another variation on a similar scheme.
Bet $18 on 6 & 8. (You have $36 invested.)
If one of them hits for $21, add $3, and press them both to $30. (You have $39 invested.)
If one of them hits, take the $35. (You now have $4 invested.)
If one hits for $35, press each to $42. (you get $11 back, you are plus $7.)
If one hits for $49, press each to $54. (you get $25 back, you are plus $32.)
4 numbers have rolled so far. You will get paid $63 for each new hit.

I will add the 5 & 9 as well, doing the same press with the 5/9 as a pair (press in $10 units)
------------------------------------------------------------------------------------------

Both of these are pretty aggressive press strategies. They require 4 hits to get back to even, but if a roll gets 5 or 6 hits, you are getting paid a decent amount. On a cold table, both of these pressing routines will deplete your bank in a hurry.

Goat's probabilities chart above shows a 2.3% chance of 4 numbers hitting. I won't argue with the number, but it sure 'seems' to me that at least 4 numbers are rolled more often than that, otherwise I would never last more than an hour on a table...... ;-)
Always borrow money from a pessimist; They don't expect to get paid back ! Be yourself and speak your thoughts. Those who matter won't mind, and those that mind, don't matter!
seattledice
seattledice
  • Threads: 18
  • Posts: 171
Joined: Dec 3, 2009
March 15th, 2010 at 6:30:11 AM permalink
Quote: RaleighCraps

I like to try and minimize against the early 7 out, but get to a decent bet size as fast as possible.

Here is one scheme I have used. Assume the point is 6.
I will bet $80 inside ($25 on 5&9, $30 on 8).
If one of the numbers hits, I put $45 down and say press all the insides.
I now have $50--point--$60--$50 (5--6--8--9)
If one of them hits again, I put $10 on the table and say press all the insides.
I now have $75--point--$90--$75.
My total outlay so far is $80+$45+$10 = $135, and 2 inside numbers have hit.
I will get paid $105 if any number rolls again, so 2 more hits and I am plus $75.
-----------------------------------------------------------------------------------

This is another variation on a similar scheme.
Bet $18 on 6 & 8. (You have $36 invested.)
If one of them hits for $21, add $3, and press them both to $30. (You have $39 invested.)
If one hits for $35, press each to $42. (you get $11 back, you have $28 invested.)
If one hits for $49, press each to $54. (you get $25 back, you have $3 invested.)
3 numbers have rolled so far. You will get paid $63 for each new hit, so you will be plus $60 if a 4th 6 or 8 is rolled.

I will add the 5 & 9 as well, doing the same press with the 5/9 as a pair (press in $10 units)
------------------------------------------------------------------------------------------

Both of these are pretty aggressive press strategies. They require 4 hits to get back to even, but if a roll gets 5 or 6 hits, you are getting paid a decent amount. On a cold table, both of these pressing routines will deplete your bank in a hurry.

Goat's probabilities chart above shows a 2.3% chance of 4 numbers hitting. I won't argue with the number, but it sure 'seems' to me that at least 4 numbers are rolled more often than that, otherwise I would never last more than an hour on a table...... ;-)



Alan's table shows the probability of just one number, the 6 or 8, repeating exactly N times -- i.e. N repeats followed by a 7. So on the row for 4 hits, the probability is 5/11 * 5/11 * 5/11 * 5/11 * 6/11, or more compactly (5/11)^4 * 6/11.

The inside numbers have 18 combinations. For this situation, you only care about those 18 and the 6 combinations for the 7. The probability of exactly 4 hits is (18/24)^4 * 6/24 = 0.079 or 7.9%. There is a 31.6% probability of at least 4 hits -- this should be more in line with your perception.
DJTeddyBear
DJTeddyBear
  • Threads: 210
  • Posts: 11060
Joined: Nov 2, 2009
March 15th, 2010 at 6:44:23 AM permalink
Quote: goatcabin

Within the next few days, I will also run a few thousand two-hour sessions for each "Guy" and examine the differences in variance among them.

I hope you were simply wording that badly. You made it sound like you're going to have each guy run independent simulations.

Have all 4 players betting on the same 'shooter', at the same time.

And, if I may suggest an additional strategy:
18, 12, 12, 18, 18, 24, 24, 30, 30.... Press 6 every other roll thereafter.
I invented a few casino games. Info: http://www.DaveMillerGaming.com/ ————————————————————————————————————— Superstitions are silly, childish, irrational rituals, born out of fear of the unknown. But how much does it cost to knock on wood? 😁
boymimbo
boymimbo
  • Threads: 17
  • Posts: 5994
Joined: Nov 12, 2009
March 15th, 2010 at 7:16:57 AM permalink
On six and eight the distribution possibilities are:

Lose initial bet: 6/16 = 37.50%
Hit once = 23.4375%
Hit twice = 14.6484%
Hit three = 9.1553%
Hit four = 5.7220%
Hit five = 3.5763%
Hit six = 2.2352%
Hit seven = 1.397%
Hit eight = 0.8731%
Hit nine = 0.5457%
Hit ten = 0.3411%
Hit more than ten = 0.5684%

So you need to look at your betting progression on six and eight only that way. You can start big and regress or you can bet small and progress. A $12 flat strategy ends up with a HA of 2.778% and an expected value of -$.66667. A $12 press $12 strategy ($12 -> $18 -> $24 has an EV of -$1.22174. In the end, the EV is the betting value x the HA, pure and simple.
----- You want the truth! You can't handle the truth!
RaleighCraps
RaleighCraps
  • Threads: 79
  • Posts: 2501
Joined: Feb 20, 2010
March 15th, 2010 at 7:58:32 AM permalink
Quote: seattledice


The inside numbers have 18 combinations. For this situation, you only care about those 18 and the 6 combinations for the 7. The probability of exactly 4 hits is (18/24)^4 * 6/24 = 0.079 or 7.9%. There is a 31.6% probability of at least 4 hits -- this should be more in line with your perception.



Thanks Seattle, for correcting my misreading of the posted information.
Can you show me the math to get to the 31.6% probability for 4 inside numbers? I'm sure it has been shown before, but this time I am going to print it out, so I have it for instant referral.
And yes, 31% of the time to get 4 numbers is more of a number I can agree with.
Always borrow money from a pessimist; They don't expect to get paid back ! Be yourself and speak your thoughts. Those who matter won't mind, and those that mind, don't matter!
boymimbo
boymimbo
  • Threads: 17
  • Posts: 5994
Joined: Nov 12, 2009
March 15th, 2010 at 8:23:21 AM permalink
0 inside numbers = 6/24 = 25%
1 or more inside number = 18/24 = 75%. Exactly one number = 75% - 56.25% = 18.75%
2 or more inside number = (18/24)^2 = 56.25%. Exactly two numbers = 14.06%
3 or more inside numbers = (18/24)^3 = 42.19%. Exactly three numbers = 10.55%
4 or more inside numbers = (18/24)^4 = 31.64%.

There are 18 ways to make an inside number, 6 ways to big red. Therefore the odds of hitting an inside number before big red is 18/24.

Expected value of $44 inside, infinite rolls, no pressure, -$2.00, 4.55% HA
----- You want the truth! You can't handle the truth!
RaleighCraps
RaleighCraps
  • Threads: 79
  • Posts: 2501
Joined: Feb 20, 2010
March 15th, 2010 at 8:31:24 AM permalink
Quote: boymimbo

0 inside numbers = 6/24 = 25%
1 or more inside number = 18/24 = 75%. Exactly one number = 75% - 56.25% = 18.75%
2 or more inside number = (18/24)^2 = 56.25%. Exactly two numbers = 14.06%
3 or more inside numbers = (18/24)^3 = 42.19%. Exactly three numbers = 10.55%
4 or more inside numbers = (18/24)^4 = 31.64%.

There are 18 ways to make an inside number, 6 ways to big red. Therefore the odds of hitting an inside number before big red is 18/24.

Expected value of $44 inside, infinite rolls, no pressure, -$2.00, 4.55% HA



Thanks boy. It really is simple when you put it down, I'm not sure why I can't get this set enough in my mind to be able to calc it myself. Hoping osmosis will step up..........

If the HA on the 6/8 is 1.52%
and the HA on the 5/9 is 4.00%

How can the HA on the 5,6,8,9 be 4.55%? I would expect it to fall roughly midway between the 1.52 and the 4.00, meaning around 2.75%.
Always borrow money from a pessimist; They don't expect to get paid back ! Be yourself and speak your thoughts. Those who matter won't mind, and those that mind, don't matter!
goatcabin
goatcabin
  • Threads: 4
  • Posts: 665
Joined: Feb 13, 2010
March 15th, 2010 at 10:58:24 AM permalink
Quote: DJTeddyBear

Quote: goatcabin

Within the next few days, I will also run a few thousand two-hour sessions for each "Guy" and examine the differences in variance among them.

I hope you were simply wording that badly. You made it sound like you're going to have each guy run independent simulations.

Have all 4 players betting on the same 'shooter', at the same time.

And, if I may suggest an additional strategy:
18, 12, 12, 18, 18, 24, 24, 30, 30.... Press 6 every other roll thereafter.



Yes, of course, they will all experience the same decisions. I have enough scenarios here, I think. This is not really a matter of specific scenarios, but rather a way to get a picture of what progression and regression do to the shape of the graph of possible outcomes.

BTW, I extended the program to 20 consecutive hits, which essentially removes any bias toward the more conservative strategies. Note the probabilities get below 10 to the minus 6th power.

num hitsguy 1guy 2guy 3guy 4prob
0-18-18-18-180.545455
193-3-150.247934
229181990.112697
3304648540.051226
43868841200.023285
5531031272070.010584
6871321773150.004811
71141742344440.002187
81342102985940.000994
91352593697650.000452
101433024479570.000205
1115835853211700.000093
1218040862414040.000042
1322147172316590.000019
1425552882919350.000009
1528259894222320.000004
16302662106225500.000002
17303739118928890.000001
18311810132332490.000000
19326894146436300.000000
20348972161240320.000000

Guy 0 weighted net is -0.413625
Guy 1 weighted net is -0.543476
Guy 2 weighted net is -0.639012
Guy 3 weighted net is -0.916979

Cheers,
Alan Shank
Cheers, Alan Shank "How's that for a squabble, Pugh?" Peter Boyle as Mister Moon in "Yellowbeard"
boymimbo
boymimbo
  • Threads: 17
  • Posts: 5994
Joined: Nov 12, 2009
March 15th, 2010 at 11:17:53 AM permalink
Quote: RaleighCraps

Quote: boymimbo

0 inside numbers = 6/24 = 25%
1 or more inside number = 18/24 = 75%. Exactly one number = 75% - 56.25% = 18.75%
2 or more inside number = (18/24)^2 = 56.25%. Exactly two numbers = 14.06%
3 or more inside numbers = (18/24)^3 = 42.19%. Exactly three numbers = 10.55%
4 or more inside numbers = (18/24)^4 = 31.64%.

There are 18 ways to make an inside number, 6 ways to big red. Therefore the odds of hitting an inside number before big red is 18/24.

Expected value of $44 inside, infinite rolls, no pressure, -$2.00, 4.55% HA



Thanks boy. It really is simple when you put it down, I'm not sure why I can't get this set enough in my mind to be able to calc it myself. Hoping osmosis will step up..........

If the HA on the 6/8 is 1.52%
and the HA on the 5/9 is 4.00%

How can the HA on the 5,6,8,9 be 4.55%? I would expect it to fall roughly midway between the 1.52 and the 4.00, meaning around 2.75%.



It's because the HA is based on "per decision" where as this "HA" is based on decision to leave the bet up until it dies.

-On a per decision basis for the 6/8, the HA is (7/6*10/16 - 6/16*2) = 1.042% with a EV of ($.25) on $24.
-On a per decision basis for the inside bet, the HA is (18/24*14 - 44*6/24)/44 = 1.136% with an EV of ($.50) on $44.
-On a $64 across bet, on a per decision basis, the HA is (18 x 6/30 + 14 x 18/30 - 64 x 6/30)/64 = 1.25% with an EV of ($.80).

The reason why the HA is lower is because there are alot more decisions. But the EV does indeed go up with the more money bet.




The HA is actually lower on a per decision basis but the EV is higher.
----- You want the truth! You can't handle the truth!
RaleighCraps
RaleighCraps
  • Threads: 79
  • Posts: 2501
Joined: Feb 20, 2010
March 15th, 2010 at 2:04:35 PM permalink
Thank you. I think it is actually starting to sink in.
With 4 bets active, the HA is lower than it would be with just one single bet, but the lower HA is offset by the act that there is now more money at risk (4 bets), so the house is expected to collect more of my money overall.
Always borrow money from a pessimist; They don't expect to get paid back ! Be yourself and speak your thoughts. Those who matter won't mind, and those that mind, don't matter!
goatcabin
goatcabin
  • Threads: 4
  • Posts: 665
Joined: Feb 13, 2010
March 15th, 2010 at 4:16:45 PM permalink
Quote: goatcabin

Quote: DJTeddyBear

Quote: goatcabin

Within the next few days, I will also run a few thousand two-hour sessions for each "Guy" and examine the differences in variance among them.

I hope you were simply wording that badly. You made it sound like you're going to have each guy run independent simulations.

Have all 4 players betting on the same 'shooter', at the same time.

And, if I may suggest an additional strategy:
18, 12, 12, 18, 18, 24, 24, 30, 30.... Press 6 every other roll thereafter.



Yes, of course, they will all experience the same decisions. I have enough scenarios here, I think. This is not really a matter of specific scenarios, but rather a way to get a picture of what progression and regression do to the shape of the graph of possible outcomes.



So, I finished writing the program and ran 10,000 sessions of 60 bets each. The overall W-L percentage was .453978, so the sample was slightly unfavorable. Each player experienced exactly the same bet results, since they were standing right next to each other, making the same bet but each in his own amount. It takes about 200 rolls to resolve 60 place bets on the six, so these sessions would take somewhere in the neighborhood of two hours.

parameter Guy 1 Guy 2 Guy 3 Guy 4
mean net result -$14.95 -$19.23 -$22.24 -$31.35
median net result -$14 -$26 -$37 -$81
standard deviation $130 $166 $200 $323
winning sessions 4599 4342 4223 3965
breakeven sessions 6 47 19 12
losing sessions 5395 5611 5758 6023
lost more than $300 154 384 659 2051
lost more than $500 3 5 17 257
won more than $300 63 357 618 1424
won more than $500 0 22 128 614
biggest win $416 $641 $894 $2112
biggest loss -$509 -$579 -$644 -$786
skew -.06 .25 .50 1.14


The idea here is to isolate the effect of progressing/regressing bets by looking at just one bet at a time.
Cheers,
Alan Shank
Cheers, Alan Shank "How's that for a squabble, Pugh?" Peter Boyle as Mister Moon in "Yellowbeard"
RaleighCraps
RaleighCraps
  • Threads: 79
  • Posts: 2501
Joined: Feb 20, 2010
March 15th, 2010 at 5:11:54 PM permalink
Quote: goatcabin


To illustrate the "continuum of progressiveness", so to speak, I wrote a computer program to compare some different betting strategies for placing the 6. We will call them "the four guys". Everyone starts at $18. Guy #1 then regresses twice before going back up, then regresses again. Guy #2 increases his bets a unit at a time, but only every other bet. Guy #3 progresses a unit at a time every bet. Guy #4 doubles first and then adds $18 each bet. So, their bet series are:
Guy #1 18,12,6,12,18,24,18,12,6,12,18
Guy #2 18,18,24,24,30,30,36,36,42,42,48
Guy #3 18,24,30,36,42,48,54,60,66,72,78
Guy #4 18,36,54,72,90,108,126,144,162,180,198
Cheers,
Alan Shank



I have quoted the original post, to bring the betting patterns closer to the new results.
Awesome analysis Alan. There is a lot of insight that can be gleaned from studying these results.

The stat I find the most interesting is the biggest loss. Guy 1 lost $509, while Guy 4 lost only slightly more at $786.
However, on the biggest win, Guy 1 was $416, compared to Guy 4 at $2112.
So, Guy 4 got 5x the money on the biggest win, and only lost 1.5x as much on the worst loss?
Of course, Guy 4 also had a much greater number of losses over $500
Always borrow money from a pessimist; They don't expect to get paid back ! Be yourself and speak your thoughts. Those who matter won't mind, and those that mind, don't matter!
goatcabin
goatcabin
  • Threads: 4
  • Posts: 665
Joined: Feb 13, 2010
March 15th, 2010 at 7:16:33 PM permalink
Quote: RaleighCraps

Quote: goatcabin


To illustrate the "continuum of progressiveness", so to speak, I wrote a computer program to compare some different betting strategies for placing the 6. We will call them "the four guys". Everyone starts at $18. Guy #1 then regresses twice before going back up, then regresses again. Guy #2 increases his bets a unit at a time, but only every other bet. Guy #3 progresses a unit at a time every bet. Guy #4 doubles first and then adds $18 each bet. So, their bet series are:
Guy #1 18,12,6,12,18,24,18,12,6,12,18
Guy #2 18,18,24,24,30,30,36,36,42,42,48
Guy #3 18,24,30,36,42,48,54,60,66,72,78
Guy #4 18,36,54,72,90,108,126,144,162,180,198
Cheers,
Alan Shank



I have quoted the original post, to bring the betting patterns closer to the new results.
Awesome analysis Alan. There is a lot of insight that can be gleaned from studying these results.

The stat I find the most interesting is the biggest loss. Guy 1 lost $509, while Guy 4 lost only slightly more at $786.



Well, that's 54% more; I would hardly call that "slightly more".

Quote: RaleighCraps

However, on the biggest win, Guy 1 was $416, compared to Guy 4 at $2112.
So, Guy 4 got 5x the money on the biggest win, and only lost 1.5x as much on the worst loss?
Of course, Guy 4 also had a much greater number of losses over $500



Look at the standard deviation figures, Raleigh. That is the key. The reason the graph of Guy 4 is right skewed, not symmetrical, is that he progresses only on wins, returning to $18 after each loss, so he doesn't lose very many big bets. For the opposite reason, Guy 1's is slightly left-skewed due to the regressions. Note also that Guy 4 had a net loss in over 60% of his sessions.

BTW, I created formulae for each "Guy's" betting pattern so they could extend indefinitely.
Cheers,
A
Cheers, Alan Shank "How's that for a squabble, Pugh?" Peter Boyle as Mister Moon in "Yellowbeard"
RaleighCraps
RaleighCraps
  • Threads: 79
  • Posts: 2501
Joined: Feb 20, 2010
March 16th, 2010 at 6:02:15 AM permalink
Quote:

Well, that's 54% more; I would hardly call that "slightly more".

Quote: RaleighCraps

However, on the biggest win, Guy 1 was $416, compared to Guy 4 at $2112.
So, Guy 4 got 5x the money on the biggest win, and only lost 1.5x as much on the worst loss?
Of course, Guy 4 also had a much greater number of losses over $500



Look at the standard deviation figures, Raleigh. That is the key. The reason the graph of Guy 4 is right skewed, not symmetrical, is that he progresses only on wins, returning to $18 after each loss, so he doesn't lose very many big bets. For the opposite reason, Guy 1's is slightly left-skewed due to the regressions. Note also that Guy 4 had a net loss in over 60% of his sessions.

BTW, I created formulae for each "Guy's" betting pattern so they could extend indefinitely.
Cheers,
A


Yes, the loss is 54% greater, but it still was only $280 more. So $280 more at risk, and a $1700 greater profit on the biggest win. And of course, way more losing sessions overall.
But I think most of all this shows that to be happy with your craps sessions, you really need to decide what kind of results you are looking for.

++ Are you looking for a big killing (3x or 4x your buyin)?
Aggressively presss, play full odds, etc. Very strong likely hood you are going to end the session losing money.

++Are you a grinder, looking to play for a few hours, and end up close to even?
Look at using a variation of a regression system, and keep your odds bets low to reduce the variance.

I find this thread fascinating reading. It really is insightful into how different betting patterns will affect the outcome of a 2-3 hour session.
Always borrow money from a pessimist; They don't expect to get paid back ! Be yourself and speak your thoughts. Those who matter won't mind, and those that mind, don't matter!
goatcabin
goatcabin
  • Threads: 4
  • Posts: 665
Joined: Feb 13, 2010
March 16th, 2010 at 9:26:19 AM permalink
Quote: RaleighCraps


Yes, the loss is 54% greater, but it still was only $280 more. So $280 more at risk, and a $1700 greater profit on the biggest win. And of course, way more losing sessions overall.
But I think most of all this shows that to be happy with your craps sessions, you really need to decide what kind of results you are looking for.



Yes, that is exactly what, in my view, people should do, and running simulations like this is exactly the way to do it. It's an extension of the same idea as knowing the probabilities of the dice combinations, or the HA, variance, skew, etcc. of individual bets. If I play this way, I have a n% chance to bust my $1000 bankroll, I have a n% chance to come out ahead, I have a n% chance to win $500, etc. etc. You can see the "shape of possible futures", just like you can with the "perfect 36" pyramid or the "perfect 1980". When you go out and actually play, the dice select for you one of those possible outcomes.

Cheers,
Alan Shank
Cheers, Alan Shank "How's that for a squabble, Pugh?" Peter Boyle as Mister Moon in "Yellowbeard"
DeMango
DeMango
  • Threads: 36
  • Posts: 2958
Joined: Feb 2, 2010
March 18th, 2010 at 9:21:30 AM permalink
Alan;
I wonder if you would consider cross posting your material on some of the dice boards?
When a rock is thrown into a pack of dogs, the one that yells the loudest is the one who got hit.
goatcabin
goatcabin
  • Threads: 4
  • Posts: 665
Joined: Feb 13, 2010
March 18th, 2010 at 12:52:48 PM permalink
Quote: DeMango

Alan;
I wonder if you would consider cross posting your material on some of the dice boards?



What boards? I also post on crapsforum.com
Cheers,
Alan Shank
Cheers, Alan Shank "How's that for a squabble, Pugh?" Peter Boyle as Mister Moon in "Yellowbeard"
DeMango
DeMango
  • Threads: 36
  • Posts: 2958
Joined: Feb 2, 2010
March 19th, 2010 at 4:10:08 PM permalink
Just posted these on another thread:
http://axispower.proboards12.com

http://diceinstitute.proboards55.com

The latter especially deals in practical theory, the former in dice seminars.
When a rock is thrown into a pack of dogs, the one that yells the loudest is the one who got hit.
goatcabin
goatcabin
  • Threads: 4
  • Posts: 665
Joined: Feb 13, 2010
March 19th, 2010 at 6:30:01 PM permalink
Quote: DeMango

Just posted these on another thread:
http://axispower.proboards12.com

http://diceinstitute.proboards55.com

The latter especially deals in practical theory, the former in dice seminars.



I respectfully ask you not to post any of my stuff on those sites.
Thanks,
Alan Shank
Cheers, Alan Shank "How's that for a squabble, Pugh?" Peter Boyle as Mister Moon in "Yellowbeard"
DeMango
DeMango
  • Threads: 36
  • Posts: 2958
Joined: Feb 2, 2010
March 20th, 2010 at 8:30:51 AM permalink
No I would not. I wondered if you would.
When a rock is thrown into a pack of dogs, the one that yells the loudest is the one who got hit.
  • Jump to: