Comparative Strategy Analysis

[DeadCat]

Looking back at this thread I wonder if it would be better for comparing to other strategies my Put bets should be half the size they are now, making them each $10 flat/$100 odds although this makes the relative vig a little more.

Still, any time you can put the great bulk of your money in Odds you are better off.


-DC

[Ahigh]

It's interesting to note that across is holding it's own in the #2 spot behind 10x odds in the second chart above. The thing that is keeping this random sample of 10,000 outcomes better than the 1.41% downslope is the non-working bets on the comeout. My 32 across turns the place bets off on the comeout rolls.

Otherwise the divergence between these two groups would not be there.

There are lots of colors but look at the purple that's the 2nd place winner in the second chart. That's 32 across losing only $300-400 bucks in 10,000 randomly generated rolls. It's harder to tell in the first chart if the winner is passline with no odds or 32 across. But I'm guessing it's do,0x (passline no odds).

[Ahigh]

These are my 300 august 2012 rolls with all the strategies. It's busy, but you can see strategies by the shape without having to use the legend.

The strategy that looks almost as good as MP204 on my last 300 rolls is a strategy that the boxman gave me: Fibonacci in the field.

[Ahigh]

Hitting pay dirt already.

[heavy]

Much easier to read when you BOLD the graphs. It would be interesting to consider a combination of your Four and Ten strategy and the Fibo Field based on the game at the Santa Ana Star in Albuquerque. They have no vig ever on the four and ten with a $500 max bet - and pay triple in the Field on the two AND the twelve, which effectively makes the Field a no-vig bet. Of course, this is about a 16 foot table that's a tad bouncy so it might be a challenge to toss on with any effectiveness.

[Ahigh]

Yeah, I tried MP204 and Fib today, and they both sucked hard and fast for me. I had more details, but let's just leave it at that.

MY CONCLUSION: strategies that depend on ANY bet with an edge over 2% are not something I would endorse or try again. EVER!

If you want to make a longshot bet like 204 across, AT LEAST leave out the 5 and 9. And absolutely be aware that you're one roll away from losing all those bets .. each time the dice get tossed. Don't kid yourself about that by focusing on winning and taking it down.

[slayyyyerfan]

Other DIs that you aren't familiar with their abilities, not chicken feeders.

[Ahigh]

The MP204 I bet on a random shooter.

http://diceinstitute.maxforum.org/2011/ ... ing/#post9

I suppose my mistake was that this betting strategy was intended for specific shooters.

But let me tell you this: the strategy is a terrible strategy for random rollers.

Certainly that's something I can assert without too much arguments from anyone.

[Ahigh]

I will admit that I had thought to have read the explanation of this strategy to be one that you could use when you weren't shooting. When I went back to read it, apparently the strategy is one to use when you were shooting but someone who doesn't suck major donkey balls is shooting.

But still I can see the graph, and I caught one of those down drafts right out of the chute. So timing is everything, but not a good taste in my mouth, let me tell you.

[Ahigh]

To make my point, let me suggest this:

Just for sattire, let me invent the AH274

$2 aces
$4 ace-deuce
$32 4
$45 5
$54 6
$54 8
$45 9
$32 10
$4 yo
$2 hi

win:
2: 50
3: 52
4: 50
5: 51
6: 51
8: 51
9: 51
10: 50
11: 52
12: 50

Two hits, and we'll just go with going across for $60 pocketing the $40 - $42 as locked up profit, heavy on the 5 and 9. Go ahead and tip the extra $1 or $2 every time because that doesn't really matter anyway either. It's only a couple bucks.

Then use these press schedules:

4/10 -> ( 5, 10, 15, 20, 25, 50, 75, 100, 150, 200, 250 , 500 )
5/9 -> ( 25, 50, 75, 100, 150, 200, 250, 300, 400, 500 )
6/8 -> ( 5, 10, 15, 35, 65, 95, 100, 200, 300, 400, 500 )

Let's say the subtitle to this betting strategy is

"screw the edge .. my srr is in another universe, bro"

I'll code this up, and we can compare the two strategies. But a lot of what I hear is along these lines, I'm just exxagerating it a bit. Who knows?!? Maybe this is the secret. Screw the M.F. edge?

I'm going to be having some tournaments at my house, so this will be one of the strategies for that tournament, and maybe this is how people throw, I don't know!

[Ahigh]

As far as clarification, I lost $600 at the tables on these two strategies.

[slayyyyerfan]

I will admit that I had thought to have read the explanation of this strategy to be one that you could use when you weren't shooting. When I went back to read it, apparently the strategy is one to use when you were shooting but someone who doesn't suck major donkey balls is shooting.

But still I can see the graph, and I caught one of those down drafts right out of the chute. So timing is everything, but not a good taste in my mouth, let me tell you.

My previous post got messed up. I meant it to say the MP204 was meant to be used on other shooters that are DIs that you aren't familiar with their strongest numbers or tendencies, nbut not random shooters.

[Ahigh]

As far as endorsing strategies, I absolutely do not endorse strategies. In fact, I don't think I have ever told anybody what to bet. I have only told people what _not_ to bet and even then it's only based on facts. And the only facts that I know about these bets is the edge.

[DeadCat]

Ahigh, I hope your "total 7exposure" for these varied strategies is the same for each. By that I mean, there,s no point in comparing $204 Across with $22 Inside and $5 starting Fibonacci on the Field. All risk sizes should be at least close to each other in dollars.

I think what you charts are showing is what a lot of us expected from the start. That is betting methods alone won't overcome house advantage. Maybe for a while, but ultimately all will fall to the math. The functional question with systems/methods then becomes, "Will my term of application be short enough to get out with a win?" As I said if we only played 1 shooter, for 1 series (hand) most of us would be able to win with big buy-in, a Martingale and high limits but you'd be nuts to use it day on, day out on random games.

As far as the higher house advantage making some bets worthless, that too depends on how much you play. I won a tournament betting primarily in the Horn. We've all seen people have slot-jackpot sized wins betting Horn bets but the more you play them the more detrimental to bankroll to bankroll they are.

In another thread you mentioned that craps tournaments are like lotteries, pick some numbers and hope they come up. In a way, so are very short random sessions.


-DC

[Ahigh]

If there were a one-roll bet called "don't roll a seven" the true-odds pay would be 1/5th the bet. For 274, that would be 54.8.

The total coming down on a one-roll bet would be 328.8 (274 + 54.8) -- if it were a free bet.

The combined edge for engineering that one-roll bet as above even where you toss in the extra dollar or two to the dealers is: 1.45% per roll. Because you're just looking for 50, and leaving the 4.8 for the house/dealers (makes the math easier anyway).

4.8 / 328.8 = 1.45985%.

So as far as one-roll bets go, it's not a bad bet for a one-roll bet when you look at the combined edge. It's half the edge of the field bet anyway.

So I guess what I'm saying is that if you can really not roll sevens, then just stick with this bet over and over.

The whole hypothesis is based on a SRR ratio. The dealers would love getting tipped on every 3, 5, 6, 8, 9, and yo. You still have a low house edge, and if you are really good at not rolling sevens, my betting strategy is WAY SIMPLER.

But it's still intended to be a joke. But who knows, maybe the joke is on me if there's all these guys that I don't know that have SRR's that are in another universe, this should work GREAT.

[Ahigh]

I will code up these "no seven" strategies for "off on the comeout" and see what they look like. I mean I know they go up linearly, and fall off a cliff every six rolls on average. But it would be interesting to see anyway.

[Mad Professor]

Hi Ahigh,

Given your propensity for detail, I am surprised that you missed the hundreds upon hundreds of times where I emphasized (as much as humanly possible without making it look like a Surgeon General’s cigarette-package warning) that the MP $204-Across Regression was NOT designed for random-betting at all, EVER.

I know you want to ‘discover’ all of these things entirely by yourself and on your own, but shouldn’t any of the common-sense “CAUTION: Contents of this coffee cup are HOT!!!” warnings be given at least a tiny fractional little bit of consideration, or are you immune even to those?

Add in the fact that you weren’t aware of bets being subject to volatility (meaning that any given hand, especially a negative-expectation randomly-rolled one could lose); and you have a one-way ticket to a one-session-trial DISASTER.

I don’t know if your programming skills are capable of doing this or not; but if you programmed one of your sims to run at the SRR-6.8 level (instead of the guaranteed-failure for EVERY betting-approach random-rate); then I think you’ll find that the MP $204-Across fares quite nicely.

The reason I point to the SRR-6.8 rate is because that was the combined-average de-randomizing ability-level of the three-dozen skilled-shooters that I tested the MP $204-Across against for ~1700 hands; and I’m wondering if your findings (using your sim at the SRR-6.8 mark if you are capable of programming it that way) would mirror my +$20,000 results…or if it would fall as flat as your one-session trial?

By the way, your “Tried it once, didn’t like it” decision reminds me of an old joke:

A man walked into a bar on a slow night and sat down. After a few
minutes, the bartender asked him if he wanted a drink. He replied,
"No thanks. I don't drink. I tried it once, but I didn't like it."

So the bartender said, "Well, would you like a cigarette?"

But the man said, "No thanks. I don't smoke. I tried it once,
but I didn't like it."

The bartender asked him if he'd like to play a game of pool, and
again the man said, "No thanks. I don't like pool. I tried it once,
but I didn't like it. As a matter of fact, I wouldn't be here at
all, but I'm waiting for my son."

The bartender said, "Let me guess, your only son?"

In any case, if you’ve got the programming chops to run an SRR-6.8 SRR-rate on your sim; I would still be interested to see if your war-gaming results come anywhere close to replicating the +$20,000 roll-stat results that those three-dozen skilled-shooters produced when I applied the MP $204 Across Regression to their rolls...or if it is doomed for failure because of its 'so-called' ( ) high house-edge.


MP

[Ahigh]

Yes, I made the mistake of using this strategy on a random roller. The strategy packing all the risk up front. It works great about 25/36th of the time, or maybe slight more frequently on a better shooter. The rest of the time, you get punched in the gut and seeking another 204,408,612 bucks and hoping to press your way back with $44 on a couple of hits. That $44 inside is awesome when you're in the plus, but if you got punched for $600 before you get to that point, I can imagine it's gotta feel pretty shitty hoping to come back on that. Even if they have a 6.8 SSR and say something like "goddamn it" when they roll a seven while you're up there for a couple of hits, still it has to suck.

To me this is still some heavy gambling against something that happens a little more often than 1 out of 4 times. Even on a good shooter.

I don't really need the sim to tell me how likely it is for this strategy to fail, and with all the risk packed against a single roll, it's got a lot of gambling elements in it all there up front.

It takes a lot of rolls to exploit the edge that a good shooter has, and I just don't think MP204 is the best way to make money off any bias. If it has credibility, it has to hold up against random rolls equally well to other strategies over the longer haul. And I just don't really think that's the case.

Any strategy that is looking to milk a players edge has to have an edge on the strategy low enough that the player can overcome it.

I will code up the SRR and we can look more in general at what you can expect and/or what SRR is required for MP204 to have a favorable risk/reward ratio. But that's a lot of risk!

[Ahigh]

I will go another step further and say that to _really_ exploit a shooter, you want to have more knowledge of not just what they have rolled in the past, but what you expect they are going to roll right now.

I will tell you that I have often put up a big come bet and rolled a seven to clear people off the table. It's easier to clear the table without hurting yourself than it is to win money, because plenty of other people leave as soon as the next seven rolls with their strategies. My playing techniques (I won't even say strategies) generally keep me at the table a very long time even when things are not going my way.

[heavy]

I will tell you that I have often put up a big come bet and rolled a seven to clear people off the table. It's easier to clear the table without hurting yourself than it is to win money, because plenty of other people leave as soon as the next seven rolls with their strategies.

I've done the same thing with Lay bets.

You are making a good case for charting tables and shooters - which is something I do when playing lone wolf style. I want to see every shooter at the table handle the dice at least once before I bet with or against him. I have no qualms about switching from the right side to the wrong side based on a shooter's past performances.

One of the benefits of playing in a specific casino or casinos on a regular basis is that you get an opportunity to see the local regulars hitting the tables. For example, there's one doctor who plays in Shreveport/Bossier that I've never seen more than a 3 number hand. I play the don'ts on him every time. Other guys I've seen have decent tosses and hit a lot of repeating numbers. One or two have "signature" numbers that I automatically bet when they have the dice. I think charting has a place in a players overall scheme. As Dylanfreake used to say - the dice don't have a memory but they do have a history and history tends to repeat itself.