goatcabin
Joined: Feb 13, 2010
• Posts: 665
March 17th, 2010 at 3:28:11 PM permalink
I had an interesting idea for a betting strategy to implement at my local casino, Cache Creek. It involves just a don't come bet, laying different odds multiples depending on the state of my bankroll. On the one hand, it increases the odds bet a couple of times at different amounts of net winnings; on the other hand, it also increases the odds bet a couple of times at different amounts of net loss. So, it has elements of two main categories of betting systems: increasing on wins and increasing on losses. So, it has the "back half" of the D'Alembert system, but increases on wins rather than decreases, as does the D'Alembert.

My strategy is not based on wins or losses, but rather bankroll benchmarks.
Here's how it works:

when a point is established, make one \$5 don't come bet
if it goes to a point number, lay single odds (\$10 4/10, \$9 5/9, \$6 6/8)

if/whenever the bankroll gets over \$250 or under \$150, lay double odds (\$20, \$15, \$12)
if/whenever the bankroll gets over \$300 or under \$100, lay 3, 4, 5X odds (\$30 for all)

So, if I'm winning more than \$50, I increase the variance with more odds, hoping to win even more. If I lose back to under \$250, I go back to single odds.

If I'm losing more than \$50, I increase the variance, hoping to come back. Of course, this increases the probability of busting, but it also increases the probability of coming back. Yes, this is what Sam Grafstein calls "chasing losses", but I am only talking about \$200, which I am fully prepared to lose. If I come back up to over \$150, I go back to single odds.

I created a WinCraps auto-bet file to test this out. The parameters:

starting bankroll \$200
sessions end when:
1) bankroll less than \$35 (can't make \$5 + \$30)
2) 200 rolls and last bet resolved
10,000 sessions

avg. num. rolls 197
avg. num. bets 53.5

avg. bet handle \$523

avg. net result -\$3.62

median net result -\$3

mode of net result +\$1 to +\$7

standard deviation \$90

avg. house advantage 0.69%

winning sessions 4844

breakeven sessions 47

losing sessions 5109

number of busts 701

lost more than \$100 1228

lost more than \$50 2892

won more than \$50 2574

won more than \$100 935

won more than \$150 477

won more than \$200 242

biggest win \$438

So, it provides almost a 49% chance to break even or better, well under a 10% likelihood of busting and a 1-in-4 chance of winning at least 1/4 of my bankroll. Of course, things can go south pretty quickly - one session last only 42 rolls - but that's true of almost any strategy. There's always a sequence of rolls out there that will hammer any strategy, as well as a sequence that will suit it to a tea.

I tried it out last night. (Next post)
Cheers,
Alan Shank
Cheers, Alan Shank "How's that for a squabble, Pugh?" Peter Boyle as Mister Moon in "Yellowbeard"
goatcabin
Joined: Feb 13, 2010
• Posts: 665
March 17th, 2010 at 3:29:23 PM permalink
Went out to Cache Creek last night, waited for the current shooter to finish, bought in for \$200 and tried my don't come, variable odds strategy. Fortunately, I was able to squeeze in next to the dealer, with good access to the "Don't Come Bar 12" box. See initial post in thread for details.

The Midnight Skulker gave me the idea of playing don't come instead of don't pass to avoid betting directly against the shooter. Also, the DC box is pretty obscure, and some players probably aren't even aware of it.

After the comeout established a point, I put my \$5 down on DC, and a nine showed. The dealer moved my cheque behind the 9 and I gave him two reds and said, "Nine dollars odds don't nine". He changed a red, heeled my \$9 and gave me a yellow back. "Seven out!" He put out two reds, stacked them up with the winnings on top, and set them in front of me. He also opened his hands a couple of times, like they do when they go off shift or move to another position. So I had two reds, four yellows and two more reds in the stack. This happened twice more on 5 or 9, so I was up \$30. When the next DC point was six, I gave him \$6 and he bridged the yellow across the two reds. I won that one, too, and on my next DC bet I threw out a "two-way 12". I was thinking that if a 12 showed I would win my DC and the dealers and I would each get \$31. It didn't show. I won a couple more bets before I finally lost one, then won a couple more. I was now over \$50 ahead, so I moved to double odds, laying \$20, \$15 and \$12 to win \$10. I made another two-way 12 that didn't win. Finally, I got a 12 on my DC comeout; the dealer didn't pay me, so I pointed to it. He just pointed to the very clearly displayed "Bar 12". Duh!! I felt like an idiot! I've been playing craps since about 1978, but I don't think I've ever played the don'ts before.

Just one time, I made a DC bet while I already had one behind a point. Naturally, the next roll was a seven, so I won the \$20 but lost the \$5. I got up as much as \$80, so I put my original stake in my pocket and played with the rest. The table had gotten very quiet and several people had left as the seven-outs piled up. I switched to a two-way 2 next DC, so I would win both bets, but still no dice on that. I tried a two-way trey, giving us a little better chance to win, but still no dice. In almost two hours' play, I only won a single DC comeout, on a 3, and I lost only a couple of times to a 7 or 11. After shooters knocked off a couple of my points, I fell back below +\$50 so went back to single odds. Some people who had left came back to play some more. I had been declining to roll the dice, but finally when they came to me I decided to switch to pass. I tossed a seven, an eleven, then set a point, took single odds and sevened out.

I went back to DC, hovering around +\$50, but I started to get bored. When I had \$30 plus five yellows left I tossed the yellows and a red to the dealers and left with \$25 plus my original. So, I won \$40, left a \$10 toke and another \$5 in failed dealer bets.

Nobody said a thing to me about betting the "wrong" way.

It was a pretty interesting experience, so maybe I'll try it again next time.
Cheers,
Alan Shank
Cheers, Alan Shank "How's that for a squabble, Pugh?" Peter Boyle as Mister Moon in "Yellowbeard"
teddys
Joined: Nov 14, 2009
• Posts: 5522
March 17th, 2010 at 3:45:31 PM permalink
Interesting, thanks for sharing your results. I'm also a don'ts player. I make either a \$5 DP with max odds, and one DC with max odds (5x), or a \$5 DP with two \$5 DCs with \$30 backing up each line/come bet. (\$30 works as lay odds for every point number).

I find the variance to be tolerable, but I have only played two sessions! The first time I won \$250. The second time I lost \$130. I occasionally shoot from the don'ts. I usually will throw a couple bucks on the field for the dealers. That way they win if I roll craps or a outside number (except 5 of course), which is what I'm going for.

My bankroll is \$500 for every session. That is not my total bankroll.
"Dice, verily, are armed with goads and driving-hooks, deceiving and tormenting, causing grievous woe." -Rig Veda 10.34.4
boymimbo
Joined: Nov 12, 2009
• Posts: 5994
March 17th, 2010 at 4:40:22 PM permalink
Congrats on your winning session with Don't Come. Too funny about the 12.

I hardly ever play the dark side, but every good craps table should have DC and DP as part of their repertoire. Certainly the table will not cheer you when you win, and some people will downright just leave the table if you win alot.

At the casino I play at, there was a young guy (a regular) who just challenged me when I threw, proudly throwing his money on DP and Martingaling. I was quite happy when I hit him up for a few hundred before he got his last one, and I was okay with that.
----- You want the truth! You can't handle the truth!
goatcabin
Joined: Feb 13, 2010
• Posts: 665
March 18th, 2010 at 2:07:40 PM permalink
Quote: teddys

Interesting, thanks for sharing your results. I'm also a don'ts player. I make either a \$5 DP with max odds, and one DC with max odds (5x), or a \$5 DP with two \$5 DCs with \$30 backing up each line/come bet. (\$30 works as lay odds for every point number).

My bankroll is \$500 for every session. That is not my total bankroll.

Well, of the two I would prefer the former, as you minimize the \$\$ exposed to the comeout disadvantage. I was quite lucky on my comeouts, only a couple of 7/11; in fact, I had bloody few comeout decisions at all, but did have two bar 12s.

For a \$5 DP with 3,4,5 (\$30) odds, the ev for 60 bets is -\$4.21 with a SD of \$193. When you are making one DC bet, there will be a DC bet about 80% of the time and two of them .8^2 = .64, so you have about -\$10.25 ev with the additional variance yielding something in the neighborhood of \$301 SD (estimated as SD of one bet times sqrt(2.44*60) ). For the one DP, one DC with 5X odds, you have about -\$7.50 ev, roughly \$303 SD. Your ev/sd ratios are .0340 and .0247. With the 3 bets, 3,4,5X odds, you're paying more for a tiny bit less variance.

As a shortcut way of simulating this, I just took 60 * 1.8 = 108 and ran 50,000 sessions of 108 \$5 DP with 5X odds. I used a \$500 bankroll and a \$250 extendable win goal.

parameter
mean bet handle \$2713
mean net result -\$4.77
standard deviation \$275.40
mean house advantage 0.17%
winning sessions 26836
breakeven sessions 54
losing sessions 23110
number of busts 4272 (8.4%)
reached win goal 19170 (38.3%)
reached 108 bets 26558 (53.1%)
biggest win \$1357

Cheers,
Alan Shank
Cheers, Alan Shank "How's that for a squabble, Pugh?" Peter Boyle as Mister Moon in "Yellowbeard"
teddys
Joined: Nov 14, 2009
• Posts: 5522
March 18th, 2010 at 2:54:54 PM permalink
Wow, that is awesome! Certainly, thank you for doing that. It's interesting that the mean net result is \$-4.77, when the E.V. should be -\$7.50. (I am correct in saying -E.V. for a session=number of pass/come bets*avg. bet*1.4%?) I assume that is because of the stop loss and win goals. Or is this really a genuine method to lose less money?

BTW, I am thinking of tracking come-out rolls and point decisions next trip. I plan on using a baccarat card and one of those pens to track how many times the shooter passes, doesn't pass, comes, doesn't come, etc. Basically keeping track of how many line/come bets I make. I will share this info.
"Dice, verily, are armed with goads and driving-hooks, deceiving and tormenting, causing grievous woe." -Rig Veda 10.34.4
goatcabin
Joined: Feb 13, 2010
• Posts: 665
March 18th, 2010 at 5:24:02 PM permalink
Quote: teddys

Wow, that is awesome! Certainly, thank you for doing that. It's interesting that the mean net result is \$-4.77, when the E.V. should be -\$7.50. (I am correct in saying -E.V. for a session=number of pass/come bets*avg. bet*1.4%?) I assume that is because of the stop loss and win goals. Or is this really a genuine method to lose less money?

There are a couple of reasons for the lesser average loss: 1) when a session ends with a win goal, it is almost always shorter than the 108 bets, which reduces the bet handle and hence the expected loss. 2) this particular sample showed an HA slightly lower than expected. You could call that a "good seed". >:-)

By far the greater effect is #1. The \$7.50 was based on calculation, assuming always 108 bets, averaging out the different points.

As far as keeping track of all that stuff you mentioned, it's interesting but I caution you about drawing any conclusions from it. I think its value is going to be in comparing the results to expectation and then looking at your own results. IOW, did the rolls come in a pattern that favored your strategy or not. Here again, looking at the pattern after the fact should not tell you, "I should have switched at this point". How would you have known?
Cheers,
Alan Shank
Cheers, Alan Shank "How's that for a squabble, Pugh?" Peter Boyle as Mister Moon in "Yellowbeard"
teddys
Joined: Nov 14, 2009
• Posts: 5522
March 22nd, 2010 at 5:20:43 PM permalink
Update: I played again this week, and charted my bets. I played for one hour, laying one \$5 don't pass with \$30 odds and two don't comes with \$30 odds. I made 66 line/come bets in total, so my expected loss was \$4.60 (They moved the dice pretty slowly). I won \$85. I made one \$5 (losing) field bet for the dealers.
"Dice, verily, are armed with goads and driving-hooks, deceiving and tormenting, causing grievous woe." -Rig Veda 10.34.4
goatcabin
Joined: Feb 13, 2010
• Posts: 665
March 23rd, 2010 at 11:34:38 AM permalink
Quote: teddys

Update: I played again this week, and charted my bets. I played for one hour, laying one \$5 don't pass with \$30 odds and two don't comes with \$30 odds. I made 66 line/come bets in total, so my expected loss was \$4.60 (They moved the dice pretty slowly). I won \$85. I made one \$5 (losing) field bet for the dealers.

Congrats!
I may or may not visit Cache Creek this week. I was thinking of perhaps laying double odds from the git-go and walking if I reach +\$50.
Cheers,
Alan Shank
Cheers, Alan Shank "How's that for a squabble, Pugh?" Peter Boyle as Mister Moon in "Yellowbeard"
goatcabin
Joined: Feb 13, 2010
• Posts: 665
March 23rd, 2010 at 10:49:25 PM permalink
Quote: goatcabin

Congrats!
I may or may not visit Cache Creek this week. I was thinking of perhaps laying double odds from the git-go and walking if I reach +\$50.
Cheers,
Alan Shank

I just ran a WinCraps simulation of this, making one DC bet after a point is established, laying double odds and quitting if I got more than \$50 ahead. It yielded 69.6% winning sessions, 21.1% busts (\$200 bankroll), 68.0% reached the \$50+ win goal. Overall mean outcome was -\$3.23, median +\$60, mode +59 to + 61. Of course, these results depend on actually quitting if you get more than \$50 ahead, which I don't know whether I would do. Several of the sessions lasted just four bets, winning \$15 each time and stopping. I don't know whether I would walk away with \$60 after just a few minutes of play.

I ran the sim again using "equal to or greater" than \$50 as a stop condition, and the percentage of \$50+ nets was 70%, with an under 20% bust rate.

Here's another possiblity: if I get ahead \$75, pocket my original stake plus \$50 and play with \$25, enough for one DC with double odds. If I lose, walk; else keep going until reaching +\$100 or losing the \$25.
Well, there are lots of options when you get ahead.
Cheers,
Alan Shank
Cheers, Alan Shank "How's that for a squabble, Pugh?" Peter Boyle as Mister Moon in "Yellowbeard"