A request for Heavy

[Ugocat]

Hello Heavy, would you do me a favor and go over to John Patrick's forum and take a look at my post titled Progressions and Regressions. Tell me what you think. I post over there as Ugo. If you prefer you could just email me personally. Thank you very much for your time. Ugocat

[heavy]

Absoltely, but it will be later tonight before I can get on John's board. The computers here at the office have all those nasty old gambing websites blocked. I'm lucky to be able to get on my own sites.

[heavy]

I am not going to cut and paste from John Patrick's site to give you the full flavor of what went before this post. Suffice to say that Ugocat has inquired about the validity of playing a regression strategy versus progression strategies, etc. In thirty words or less - you play a regression strategy in order to minimize short-run bankroll volatility. You play a progression strategy in order to maximize your wins over the long run.

Here's the (much longer) post I made on Patrick's site (pending moderator approval):

I have managed to be fairly successful at craps through the years. And I can tell you that regressions - combined with sensible progressions - AND being smart enough to take what the table gives me when it's giving it up (and that includes playing the don'ts) with simple things like loss limits and win objectives.

Here's an example from some of my archival trip reports. I was playing my way up the Mississippi River about ten years ago. In St. Louis I strayed over to the Missouri River and got in an evening of play in the Harrah's property there. Missouri had laws at the time that limited the amount you could buy in for to something like $500 every two hours. On this particular trip I was playing essentially the same strategy every session. $66 inside for one hit, then regress to $22 inside. Collect the next two hits, then press the third. I would continue to press every third hit, regardless of what number rolled, until the ugly number showed. Unfortunately, in this session the ugly number showed way too often. Point-seven-out. Point-seven-out. Point-seven-out. In short order I was down $200. At that point I SHOULD have pulled the plus and called it a night. After all "lose three in a row and it's time to go" has always been one of my rules. On this trip, though, I decided to play through and attempt a recovery. When my losses reached $300 in about 20 minutes I decided to change my tact. I switched to the Don'ts. My Don't play in those days was right from John's Advanced Craps - $15 Don't Pass followed by a $10 Don't Come. I allowed myself $25 action on every shooter. It took me about three hours to grind out a $1 profit for the session. Once I'd gotten my $$$ back I colored up and called it a night - grateful to have gotten out with my bankroll intact. Would I recommend that play to everyone? Probably not. But it worked for me.

My biggest wins have pretty much all come on hands where I've taken an early steep regression followed by a fairly aggressive press schedule. In my Craps Boot Camp class I teach a regression that begins with $320 across on known dice influencers. But the strategy treats these DI's like any other shooter. It assumes they're hand is going to last an average of eight tosses. Then I plot a regression point on each shooter (this is a team play strategy) based on that shooter's individual sevens-to-rolls ratio. I regress somewhere around the halfway point on their average hand - usually after no more than four tosses. The regression is to $64 across. Bear in mind that on those four tosses the shooter's probably scored no more than three box numbers, so you've locked up $210 - perhaps more if the four and/or ten have rolled. But even with $210 in the rack you still have $110 exposed. Regressing to $64 across guarantees a $146 profit for the shooter's hand - no matter what else happens.

John would have you execute a first-win regression, and you could do that with this play and still come out ahead. First hit pays $70 or more, depending on what rolls. Regressing to $64 across that that point gives you a $6 profit and you're ready to press. Nothing wrong with that at all.

The real problem comes when your well-trained shooter fails to make it happen and you get killed by a Point Seven Out. Of course, this particular strategy protects you from that because the "rules" of the play prohibit placing any action on the shooter until he's established a point and thrown at least one paying number. Sure, you could get whacked by a "point-point-seven-out" but I find it happens much less than you might think.

How do I press? Most of the time I either press every third number that hits - or every other number that hits. With that said, there are times when I play more aggressively and press the six and eight in pairs. For example, if I have $18 each on the six and eight and the eight is tossed, I'll drop $3 from my rack on top of the $21 pay off and tell the dealer to "make my six and eight look like $30 each." On the next hit I'll drop $1 on top of the $25 pay off and make the six and eight look like $48 each. Then, at the $48 level, I'll start a take and press routine only on the number that rolls.

Another way to play that is very aggressive is to wager something like $110 even numbers. If I get a hit on the six or eight I lock it up. If I get a hit on the four or ten I full press it a couple of times. A $25 four pays $50 and gets pressed to $75. The $75 four pays $150 and gets pressed to $225. A $225 four gets pressed to $500 and you lock up $175. Now I'm set up to collect $1000 every time the four rolls. Does that happen very often? Probably more than you think, but it is a very high risk play. And I would never consider playing that strategy on a tight bankroll.

And now I have strayed completely off topic . . .

The "math guys" have told me for years that whenever I (as a skilled shooter) take a regression I am giving away part of my edge on the game. They tell me I should bet a complete Kelly (which is a bet sizing strategy where your bet varies depending on the size of your bankroll and your validated edge over the game). Yet virtually none of these math guys - almost all of whom are professional card counters - play a complete Kelly. Why? Because in many cases the volatility of the game causes wider swings in bankroll than they want to be exposed to. Craps is really no difference.

Picture a sine wave with peaks and valleys. The break-even point runs right down the center of the sine wave. You are going to have losses (below the line) and winners (above the line) and in a 50/50 (or close) game you're going to have some fairly large wins, some fairly large losses, and probably come out close to even over the long run. The problem comes when players do two things - overbetting their bankroll - and failing to maintain discipline. By discipline I mean you must learn how to have win goals and loss limits and quit when you are ahead. Even a $1 win like the one in St. Louis that I mentioned is a win at the end of the day. Small wins add up into big ones.

Going back to the late 90's and early 2000's when I finally woke up and started following many of John's principles I started string together hundreds of small wins. I remember one year where I played over 300 sessions with an average session win of around $37. Thirty-Seven Bucks! Doesn't sound like much until you multiply it times 300. Suddenly those $37 wins turned into over $11,000. That is not chump change. That can make your car payment, and in some cases your mortgage payment. Want to win more? Take that average $66 bet and put a zero behind it. Play $660 in side and regress to $220 inside. Then you can add a zero onto the average $37 win and make it $370. At 300 sessions a year you're talking $110K. It does add up.

If you program all of the millions of decisions I've played through the years into a computer and played out various strategies - Pass Line with Odds only, Don't Pass with Odds only, Placing only the six and eight, hit and down or hit and regress and down, Continuous Come betting, pressing, power pressing, or whatever strategy you can dream up - will any of those strategies out-perform the regression move. Probably. But I can tell you that you'll experience wide swings in bankroll along the way and you'll have far more sessions where you tap out than you will where you'll walk away with a win. Personally, the older I've gotten the less risk I'm willing to take. I got ride of my motorcycle. I no longer go off in the woods during deer season and get drunk with a bunch of guys who have enough firepower to take out a small nuclear production facility in Iran. I'm not going to be found hanging out at a west Texas rattlesnake round-up. Regress, my friend, and live to play another day.

[Blackcloud]

UNHH!!Ugocat, even though the math boyz swear that progressions don't work, they do quite well when the progression matches the flow of the game; but when the winning stops so does the progression. Regressions are used to lock up wins and then move on; you don't wanna lock up wins ?
UNHH!!Then don't use em

[heavy]

The math boys will tell you that progressions don't work - but then they tell you to bet the Kelly - which is a progression. When you take them to task on it you get the old "wellllll, that's different." No it's not.