Realistic expectations for skilled shooters!!!

[heavy]

Well, close enough. Regarding Irish's statement - I'd say he's probably correct. Sort of. 80% of a skilled shooters tosses are probably not optimally influenced. How's that for a term? 80% of so-called DI's probably are not. I've been told by some old-time box people that the game is so strong that it's going to beat all of us over the long run. I don't necessarily believe that is correct - however it is tough to bet on every toss and win - even when you're talking about betting on DI's. The reason is not so much tied to dice influence - I think - as it is to poor betting strategies and lack of discipline. There are few losing sessions that could not have been winners at some point if we had the good sense to quit while we're ahead. What you need to do as far as your toss is concerned - is to become one of the 20 percenters - or identify some of the 20 percenters and bet with them. Play tight. Evaluate your profit situation after every toss of the dice, continually adjusting your strategy to guarantee a win. Be willing to walk with a small win - and if necessary - a smaller loss.

[bobr]

Or maybe...

If the shooter is throwing the dice randomly, there is a 1 in 6 chance that you will lose all your bets on the next roll. On the roll after that, the odds of losing all your bets is still 1 in 6. If the shooter rolls 20 numbers, the odds on the next roll is still 1 in 6. If this were not true, then in order for the overall probability to be one in six, the odds of a PSO would have to be a lot less than 1 in 6, and then the odds of a seven coming up would increase, and anyone with money on the table after six rolls would be a chump. Pressing bets kind of assumes the opposite, that a good roll will keep going because it's a Good Roll. Odds are still just 1 in 6 each time, though.

So maybe a good shooter can change that to 1 in 7. (I'm probably closer to 1 in 6.5...) In my experience, that's because I'll have a couple of good rolls, and a lot of random ones, but I don't get to know in advance which rolls are going to be good. I'm trying every time, so I assume that the distribution of my random rolls is...random. So if you're betting on that 1/7 shooter who can influence the dice I would guess that, at least for the first dozen rolls, your odds on each roll of losing all your money is 1 in 7. Chasing someone who is influencing the dice is a little different because you have things like fatigue, increasing chips on the table, etc. weighing against positives like the shooter getting into a nice groove and maybe improving their chances even more.

Still, though, I think that using statistics as a justification for increasing or decreasing your bet as the roll progresses looks more like gambling than science.

- Bob

[heavy]

The whole math of probability gets a little wonky sometimes. I've fought the battle with some of the top math guys in the gaming industry. You are absolutely correct - that the odds of the seven showing up on the next toss of the dice is always one in six. However, and this is the point the statistics folks always want to argue, there is also what I'll refer to as a rate of decay for every hand. That seven is lurking out there somewhere. And while there's always a one in six chance of the seven showing on the next roll - with every roll you're one roll closer to it. Now, it sounds like there are some fundamental conflicts there. There aren't. The odds just don't change.

DI's with a little skill will toss longer hands on average - over the long run. But they will also have PSO's and PPSO's. It is what it is.

[Mad Professor]

If an advantage-play dice-influencer doesn't understand the difference between single-event probability and cumulative-probability after reading the hundred or so articles that I've penned on this subject (and explained in excrutiatingly simple detail); then frankly, their continued losses (or continued under-earnings) are well deserved.

I'm not trying to be overly harsh, but there is an incredibly huge difference between the probability of a single-roll event, such as the 1-to-6 randomly-expected ratio of 7's-to-non-7's...versus the likelihood of rolling two consecutive non-7's in a row, three consecutive non-7's in a row, four consecutive non-7's in a row, and so on.

The information is here...free for the taking...but like H likes to say, "You can take 'em to Shorty Smalls, but you can't make 'em order the ribs!"



MP

[bobr]

Ok, MP, I'll try an experiment in my practice. I'll record how many consecutive non-sevens I can roll, and I'll test this 15 times. I expect the answer will be pretty close to random because of how my practice rig is set up. I'll be using the hardways set.

Then, I will count my non-seven rolls again, but after I get to two, I will start over. So what I will have is a recording of my rolls that are longer than two, and how far I could roll after two consecutive non-sevens. This will take a while, because not all rolls will qualify, but eventually I'll have 15 samples also.

So what should I find regarding my rolls? How much shorter will the average in second set of rolls be? I expect the results in the second set to be similar in duration, because at the point I start measuring the first two rolls are in the past and should be irrelevant to what happens next. This sample size is pretty small, but if the difference should be dramatic, like three rolls instead of five, then it should be noticeable.

I will also, in the interest of science, stop at Shorty Smalls for ribs the next time I'm in OKC.

- Bob

[Maddog]

... I'll record how many consecutive non-sevens I can roll...

BTW.. this is EXACTLY what the RBS tab of BoneTracker does.

[Mad Professor]

Hi Bob,

As Heavy was trying to explain, the PER-ROLL 7's-expectation-rate does not change no matter whether you are on your 1st toss of the hand or your 100th toss of the hand; but what does change is each hand's expected survival-rate. That's where cumulative-prpbability comes into play.

That is why, when a large enough sample size is taken; you have more hands last 5 rolls than last 10 rolls in duration...and why there are more 15-roll hands than there are 25-roll hands.

Per-roll expectation never changes, but cumulative-probablility always does.

Average point-cycle hand-duration is determined by each shooter's SRR-rate. Now, many here will tell you that SRR-rate is meaningless because it doesn't indicate the quality of its contents...but of course we have OTHER metrics for that.

Instead, our SRR-rate tells us, on average, how LONG we've have to work with the overall dice-influencing skills that we've GOT; and so, it allows us to structure our betting-regimen in a way (if we choose) to more profitably-optimize our point-cycle's average survival-rate.

That knowledge allows us to do MORE with what we've USUALLY got.

For example, if you tracked the in-casino multi-session results of a reasonably-sized group of SRR-7 dice-influencers; you would likely find the following average point-cycle hit-rate for the Inside-bet:

1 hit per hand.......78.2%

2 hits per hand.....61.2%

3 hits per hand.....47.8%

4 hits per hand.....37.4%...meaning that, on average, 2.7 hands out of 10 will contain at least 4 Inside-hits per point-cycle for a reasonably-sized group of SRR-7 shooters.

5 hits per hand.....29.2%

6 hits per hand.....22.9%

7 hits per hand.....17.9%

8 hits per hand.....14.0%

9 hits per hand.....10.9%

10 hits per hand...8.6%...meaning that, on average, 11.6 hands out of 100 will contain at least 10 Inside-hits per point-cycle for a reasonably-sized group of SRR-7 shooters.

This is just one tiny example (out of the hundreds of charts that I've posted); where cumulative-probability can be used.


MP

[heavy]

Thanks for the expanded post, MP. It's not rocket science - you just have to get your mind wrapped around it.

[bobr]

You guys have obviously given this a lot of thought, but I'll just throw out one more example:

The number of hands that will make eight hits is obviously more than the number of hands that will make ten hits, so that's easy. I would say, though, that if you limit yourself to the group of hands that contain at least two hits, then the portion of those hands that then go on to make eight more hits will be similar to the overall ratio of 8-hits to all hands. In other words, you aren't less likely to get another eight hits based on the past events. I think this could be simulated or analyzed from a large set of data pretty easily.

So because of this, when you make a betting decision based on something that has already happened, you're probabilities are basically starting over, even though you might be in the middle of a hand.

- Bob

[gargoil]

Maybe I missed something here. The topic is about "Realistic expectations for SKILLED SHOOTERS". To me if you are a skilled shooter then your expectations should be different than a "Wanna be" skilled shooter. Therefore the laws of rolls and averages may still apply but not at the same level. I may be all the way in left field by myself here but here is how I understand this.

The rule says a seven could happen one in every 6 rolls. To me this rule applies to random shooters and non skilled shooters. If you are a skilled shooter than it should state one in every ??? 10 rolls. If the rules and averages don't change as you are better skilled then why waste your time?

Am I missing something here?? Please be gentle

[bobr]

Right, so if the shooter is skilled you hope that you have an edge on some of the bets on the table. It seems like your question was how your edge changes over the course of the roll. I operate under the assumption that if the shooter's skills don't change (tired, sloppy, or for the better after getting in the zone), then the edge stays the same, even if he has already had ten successful rolls.

[Mad Professor]

Hi Bob,

You are absolutely correct on the following assertion:

~Hand-duration projections (insofar as how long a hand will last) can only be made AHEAD of time (before a hand actually begins).

~Once a hand has commenced, mid-hand projections CANNOT take into account how many rolls have already passed. That is, the dice don't 'know' whether it's their 1st roll in this hand or their 101st.

However, let's put cumulative-probability into a real-world context:

~If an SRR-7 shooter subscribes to the notion that he should wait for one point-cycle roll after the PL-Point is established because he doesn't want any of his Place-bets to fall to a PSO/7-Out; that's fine, as long as he knows what he is automatically giving up by doing so.

Let me illustrate:

~There is an 85.72% chance that an SRR-7 shooter will survive past his first point-cycle toss (meaning that, on average, there is an 8.5-out-of-10 chance he WON'T 7-Out on his first point-cycle toss).

~There is an 73.48% chance that he will survive past his 2nd point-cycle toss...meaning that an average of only 7.3-hands-out-of-10 will still be alive at this point. So, by foregoing that first point-cycle roll, he willingly surrenders 12.24% of point-cycle roll-duration 'survivability'.

~There is an 62.99% chance that he will survive past his 3rd point-cycle toss

~There is an 53.99% chance that he will survive past his 4th point-cycle toss

~There is an 46.28% chance that an SRR-7 shooter will survive past his fifth point-cycle toss (meaning that, on average, there is an 4.6-out-of-10 chance he WON'T 7-Out by that juncture in his point-cycle).

~There is an 39.67% chance that he will survive past his 6th point-cycle toss

~There is an 34.01% chance that he will survive past his 7th point-cycle toss

~There is an 29.15% chance that he will survive past his 8th point-cycle toss

~There is an 24.99% chance that he will survive past his 9th point-cycle toss

~There is an 21.42% chance that an SRR-7 shooter will survive past his tenth point-cycle toss (meaning that, on average, there is an 2.1-out-of-10 chance he WON'T 7-Out by that juncture in his point-cycle).

~There is an 18.36% chance that he will survive past his 11th point-cycle toss

~There is an 15.74% chance that he will survive past his 12th point-cycle toss

~There is an 13.49% chance that he will survive past his 13th point-cycle toss

~There is an 11.57% chance that an SRR-7 shooter will survive past his fourteenth point-cycle toss (meaning that, on average, there is an 1.2-out-of-10 chance he WON'T 7-Out by that juncture in his point-cycle).

Now let's look at the obverse:

~There is a 14.28% chance that an SRR-7 shooter will NOT survive past his first point-cycle toss (meaning that, on average, there is a 1.5-out-of-10 chance that his hand WON'T survive past his first point-cycle toss).

~There is an 26.52% chance that he won't survive past his 2nd point-cycle toss

~There is an 37.01% chance that he won't survive past his 3rd point-cycle toss

~There is an 46.01% chance that he won't survive past his 4th point-cycle toss

~There is an 53.72% chance that an SRR-7 shooter will NOT survive past his fifth point-cycle toss (meaning that, on average, there is a 5.4-out-of-10 chance that his hand WON'T survive his fifth point-cycle toss).

~There is an 60.33% chance that he won't survive past his 6th point-cycle toss

~There is an 65.99% chance that he won't survive past his 7th point-cycle toss

~There is an 70.85% chance that he won't survive past his 8th point-cycle toss

~There is an 75.01% chance that he won't survive past his 9th point-cycle toss

~There is an 78.58% chance that an SRR-7 shooter will NOT survive past his tenth point-cycle toss (meaning that, on average, there is a 7.9-out-of-10 chance that his hand WON'T survive his tenth point-cycle toss).

~There is an 81.64% chance that he won't survive past his 11th point-cycle toss

~There is an 84.26% chance that he won't survive past his 12th point-cycle toss

~There is an 86.51% chance that he won't survive past his 13th point-cycle toss

~There is an 88.43% chance that an SRR-7 shooter will NOT survive past his fourteenth point-cycle toss (meaning that, on average, there is an 8.8-out-of-10 chance that his hand WON'T survive his fourteenth point-cycle toss).

The simple fact is that more than 88% of ALL your point-cycle hand-durations will never double-past (be 2-times longer than) your current in-casino point-cycle SRR-rate.

MP

[bobr]

Ok, cool, that all makes sense. So then it seems like most every bet on the table is a bet on what is going to happen on the next roll, not on the next 10 rolls. The firebet/all-tall, etc are exceptions, and the pass-line bet is a contract, but for every other bet on the table you are betting on the next roll. My $30 six will either lose, win, or push.

[Dave73]

While avoiding the 7 may make your hands longer it wont necessarily make you a winner. To me, having a non-seven number to bet on is MORE important than just not hitting a seven. A 6.0 SRR can still be a winner and can still be influenced dice. All you need is to roll LESS of some numbers and MORE of other numbers. I would venture to say it does not matter WHAT numbers you have more or less of as long as there is some consistency and you KNOW what those numbers are.

[Mad Professor]

Hi Tabletop,

I have no problem whatsoever with the whole anti-math/"SRR-rate has no place in polite dice-influencing conversation" direction that many guys here want to take this board in.

I'm just making the information available; what you, or Heavy, or any other reader does with it, is entirely up to the individual. You can ignore the math if you want, but you have to know, that that is exactly what all of the casinos hope most skilled-shooters will continue to do.


MP

[Mad Professor]

Hi Tabletop,

Rest assured that I didn't take your post as a personal attack at all.


MP

[shunkaha]

Or maybe...

If the shooter is throwing the dice randomly, there is a 1 in 6 chance that you will lose all your bets on the next roll. On the roll after that, the odds of losing all your bets is still 1 in 6. If the shooter rolls 20 numbers, the odds on the next roll is still 1 in 6. If this were not true, then in order for the overall probability to be one in six, the odds of a PSO would have to be a lot less than 1 in 6, and then the odds of a seven coming up would increase, and anyone with money on the table after six rolls would be a chump. Pressing bets kind of assumes the opposite, that a good roll will keep going because it's a Good Roll. Odds are still just 1 in 6 each time, though.

So maybe a good shooter can change that to 1 in 7. (I'm probably closer to 1 in 6.5...) In my experience, that's because I'll have a couple of good rolls, and a lot of random ones, but I don't get to know in advance which rolls are going to be good. I'm trying every time, so I assume that the distribution of my random rolls is...random. So if you're betting on that 1/7 shooter who can influence the dice I would guess that, at least for the first dozen rolls, your odds on each roll of losing all your money is 1 in 7. Chasing someone who is influencing the dice is a little different because you have things like fatigue, increasing chips on the table, etc. weighing against positives like the shooter getting into a nice groove and maybe improving their chances even more.

Still, though, I think that using statistics as a justification for increasing or decreasing your bet as the roll progresses looks more like gambling than science.

- Bob

No, making your move based on statistics is more science [math to be precise] than gambling. Would you walk up to a card counter and say, hey just because your count says you should increase your bet so you can make money... I say you shouldn't because it seems like gambling to me. In fact one countermeasure against counters is to make sure they can't raise or lower a bet.

Lets put this in a proper perspective, lets say I invent a game that pays even money on bets and the mechanics of it are that statistically 89% of all first decisions favor the player, 91% of all 2nd decisions favor the house. Given a choice would it not make sense to you statistically to bet the 1st decision with a higher dollar amount and the 2nd with a lower amount? If you still feel that betting based on statistical probability is gambling, would you flat bet that new game through a 1,000 1st and 2nd decisions? If the rules paid 9x on the first decision as long as you parleyed your bet but the second paid even money... would you play that rule or would you say that statistical probability said you shouldn't play it that way?

[bobr]

No, making your move based on statistics is more science [math to be precise] than gambling.

My point was just that I don't think it's a valid use of statistics to use the results of a few previous rolls to indicate what will happen on the next roll. Overall, with a lot of data, you could certainly get an idea for what to expect from someone influencing the dice and what numbers might show a profit, but just because a seven hasn't shown up for three rolls (in the past) doesn't mean that a seven is statistically more or less likely to happen on the next roll (in the future).

(I realize this has already been discussed in this thread and my goal isn't to re-open a topic when I don't have anything new to say, I'm just clarifying what I meant to shunkaha)

[heavy]

I believe if you have done sufficient tracking through the years you will get to the point that you recognize certain patterns and trends within your own sphere of competence and be able to capitalize on them. As DF said many years ago - the dice don't have a memory but they do have a history. So when you're looking at relatively short series of number you toss at a live game - you're able to fall into history repeats itself mode and capitalize on the current trend. For example, when using the 3-1 2-3 set I know that when the sixes start showing I can just take down my bet on the eight and press the six. The eight just isn't going to pay for itself. I know when Irish uses his mutant V-2 set, which is designed to produce even numbers, if the eleven shows I'd better slip into see a horn bet a horn mode. I suspect others have experienced similar situations.

[shunkaha]

A 6.0 SRR can still be a winner and can still be influenced dice.

They may be a winner due to variance, but with a large enough sample of data, someone with an SRR of 6.0 will not alter outcomes in a meaningful way. Theoretically it's possible, but theoretically it's possible to throw 1500 tosses without a seven or theoretically it's possible that if you have a million monkeys pounding on computer keyboards, one will write the great American novel, or at least Twilight.

Irish, you're wrong... I'm pretty sure it only takes one monkey pounding on a keyboard to manage to write Twilight, but really, glittery vampires?