Roulette Outside the BOX

Yes both stdevs' have grown, but the ratio of the standard deviations (the bet selection to random selection) has stayed the about the same.

Currently I'm stepping through the code which generated the data with the debugger to make sure there isn't an error.

 

I look at testing as a simple process. First, there's no need for simultaneously testing a randomized number or control selection at the same time you're testing whatever bet selection you've chosen. You can simply calculate the actual hit rate probability of the random or base choice, lets say it's 1:37 for ease, and compare it to your bet selection's hit rate over thousands of actual spins.

1. Is your hit rate higher than expected/probability? (you win)

2. Is your hit rate the same as what is expected/probability? (house wins, payout is wrong)

3. Is your hit rate lower than what is expected/probability? (house really wins, payout is wrong and your BS is actually worse than the expected hit rate)

In all but 2 or 3 of the 200+ methods/systems I've tested over the years, the answer is in numbers 2 and 3 above. Given enough spins, even the best methods come back down to EV if not based on VB/AP biased wheels or dealers. That includes Palestis' Single Dozen method which was talked about briefly yesterday. It was no problem breaking that under testing and showing there was no actual edge, it simply provided the ability to dance between the raindrops for a very very long time.

Only when it comes to repeaters, and specifically repeaters under certain conditions, has #1 occurred.

Currently through 5197 spins, the base method in this thread has a hit rate above probability by +1.9%. That's high, I've only seen it breach that level in shorter test runs with other systems, but it's still well within "luck" or a good run. As Caleb correctly says, you need a BS to hit higher than expected, for many many many spins before you can begin to think it's giving you an advantage.

How many spins is a good question, system players bristle at the "1M spins" requirement, but similarly 10,000 - 20,000 spins simply isn't enough either.

If you have a BS that can withstand 50,000+ spins, I would say that's strong, easily within 0.01% of the "best" ways to play non VB/AP provided that it isn't pushing spread barriers (table limits) and doesn't rely on a negative occurrence not showing up for an extended period of time (progression where your BR is wiped say 1 in 1000 attempts, but you "win" the other 999, and just haven't hit the 1000 wipe event yet).

So applying that to this thread, something about numbers that have been dormant for 37+ spins, waking back up with a first hit, and then hitting far beyond EV for the next 150 spins maximum is definitely a thing.

I don't know what the "thing" or phenomenon is, but it goes beyond just catching back up to 0 STD. It happens over and over, so strangely, to the point where in 85% of my tests one of the "awoken" numbers becomes the leader for those 150 spins, hitting more often than even numbers that hit 3-4x within the very first 37 spins...and had a huge lead in the so-called horse race before the first unhit appeared in that cycle.

Now hitting at +1.9% is not going to win flat betting. The payout is still wrong, and obviously wrong enough that you're still losing even at that advantage. This isn't blackjack, +1.9% means nothing to actually being able to net + units, it's simply a good start.

Where the repeats on unhits seems to excel is with a progression, specifically a positive progression. I'm shocked at how well it's done over the 5917 spins I've completed so far, the net % edge in the 34 session wins is laughable, I won't even mention it as it seems so ludicrous. Meanwhile the losses of the 5 that didn't come in are at the same average net loss per session as the won units per session. Thus far it's as if you had an EC bet selection that hits at 87.2%....

We'll see if it continues, I test because my BR is precious, I don't give a fuck about the method itself or the system or the creator. This is money.

I look to attack, debunk, and eliminate, any system, because that's the purpose of testing. I don't test to prop up fallacy, or add to the shrine of ego of whomever developed the approach.

But I will say this, nothing has produced numbers this high for this long in anything I've ever submitted to the crucible of how I test. I'm not optimistic as my aim is to make it fail as I have all of the others, but I can absolutely see how Turbo and others have come to the point they're at in terms of focusing on repeats, how they occur, when they occur, why they occur, and how to take advantage. It's the only phenomenon that seems real, particularly when combined with binomial distribution and how the wheel's 37 pockets never change.

Will keep going with what we're doing here, let's see how it plays out at 10,000 spins, 15,000 spins, 20,000 spins, and beyond.

 

Jerome, thank you for your excellent work by the way, very much appreciated. I admire those who possess skills like yours, and am thankful anytime people like yourself take valuable time to participate in an experiment that they may not believe is feasible from the start.

Great job!

 

Indeed, Jerome!

I`m not sure people realize the importance of the results.
I would ask all anti system players to have a good look at the graphs.
Looking at them, it is obvious that the bet selection can be improved. Is this the way to go? CERTAINLY.
It started out with a simple OBSERVATION.
The framework soon followed.
We gave it an interval. We set up the progression.
Math says all numbers return to their expected EV, Give it enough time, IT WILL HAPPEN.
It is just math.
What we do not have is TIME.
What we need to improve is the attack interval. I never go more than 185 spins. But give it enough variations, and the permutations will soon break anything.

We are David against Goliath. It is a titanic effort. Or so it seems.
I`m waiting for the results of the 740 spins. Obviously in BM casinos a system like that is not usable.
We`re just testing it, (crash testing it) until it breaks and gives us what we need. Limits.

The big STOP button.

To be continued.

 

It may not seem a pretty graph.
It`s all Flat bet.

First 4297 spins live westspiel- attach
ed at the end.

 

THIS IS IMPORTANT

@ Mako, and possibly Frodo and others, I got a pm from Dr Sir a while ago, he said :

Now obviously, using those spins is likely to have an impact on your test results. I guess Dr Sir thought this was pretty funny, but I don't. Anyone who has been testing using those spins has basically wasted their time and has every right to be extremely pissed off.

I tried to send Mako a pm but couldn't (do you need to have made a certain number of posts before you can send pms?).

So if you've been using spins provided by Dr Sir, stop using them for testing!

 

So Mako or anyone else who has those spins, could you upload them here? I'd like to run a few tests. Cheers.

 

Hi Jerome,

Thanks for the heads up. I am not too worried about that to be honest, as i tested them beforehand.
The simple test is to add them in RX and press the RUN button( no bets)
Look at the numbers statistics,
WestSpiel20180101_20180717_0230 is a joke sample!

 

Thanks Jerome, I hadn't gotten to his 00 wheel spins yet, was saving them for use only if a method actually survives past the 30x spin sets I use that no method has ever gotten even half way through before failure.

I was going to ask him if they were from a biased wheel though before employing them as I'd prefer neutral wheel results for sure.

I'll shoot you a PM with my email, feel free to use it anytime, will sent you the spins Caleb sent me.

 

Here you go Jerome: They include a few cow jumping over the moon pics as well, which is a bonus.

 

@ Frodo & Mako, ok it's all good then. Actually it will be interesting to see how much difference that weak bias makes to results. It didn't seem right to let testers continue to use biased spins unknowingly, because they might end up playing the system for real (convinced that it works), and then lose money. Either Dr Sir didn't think of that, or he didn't care.

Thanks for the spins, Mako.

 

Here's how the numbers played out for that spin set:

 

That's the right attitude. As the great physicist Richard Feynman said :

"The first principle is that you must not fool yourself and you are the easiest person to fool."

BTW, these results you're getting, are they for the most recent bet selection? the one where you remove numbers that are on or above expectation on each new cycle?

 

I don't really get why you're surprised by two different variance from 2 different system...

If for example i play Red+3rd column or Red +2nd column the variance is not the same either, so?

 

Sharp, they should be the same because I'm playing the same amount of numbers.

They should be the same as regards variance. Why would they be different? I'm talking about "long term". Obviously if the 2nd column hits more frequently in a given session than the 3rd column there will be a difference, but over many sessions? no. And don't forget we're talking about large (statistically significant) differences here, not something that can be explained by "chance".

 

I've done 5197 spins on what I've termed the "Brute Force" version, and 2219 spins on the newer version Eddy developed a few days ago.

The Brute Force version is simply tracking 37 spins, identifying the unhits for that cycle, then continuing to spin while watching for any unhit to hit. Once one does hit, I begin betting on it with 1u and continue on it with a 1u-2u-4u-8u-16u-8u progression until the session ends.

I don't remove any of the unhits that wake up ever, as this began as just an experiment for me to learn how unhits arrive, vanish, or repeat, within a typical playing session (120-200 total spins). I wanted to get a feel for them before I began limiting them by cycle or individual profit and removing or adding them.

What I've been surprised to find is this strange higher than expected incident rate where one of the waking unhits then goes on to hit far above EV for the playing session, delivering an unexpected + net unit gain the vast majority of the sessions.

Since I've gone through 39 of the sessions, it's been easy to see when it's going well (8u and 16u hits come earlier than average), or when it's going to fail (8u and 16u hits are later than average), allowing identifiable escape points for both + or - unit situations when compared to the total units being bet per spin at any given moment versus the amount of larger unit bets that are active.

I haven't been able to get as good of a mechanical grip on the newer version, where we're removing numbers that wake up, catch up to their expected rate (and in doing so allow a profit as we never bet on them in that initial 37 spin cycle) and are removed. The draw downs are larger in that version for me so far, and it's not as obvious as to when you should quit either way, if up OR down. One wrong move with that one is either disaster, or you missing out on a massive gain. I'm going to keep experimenting with it though as I think it has merit.

Once I get to 100 sessions of the Brute Force version, approximately 10,000 spins worth, I'll publish more data so anyone can see how each individual session played out and what the highs/lows were, etc.

 

Guys,
What also maybe of importance is with how many unhit numbers you start out with!
What i also found out in my testing is that it isn't always the better option to remove numbers at the end of a cycle where we delete the ones that have made it back to their AV.

 

The reason i bring this up is that i saw that after a cycle has ended, the removed numbers keep hitting!!! why this happens? i don't have a clue.

 

Yes it's pretty wild, unexpected for me personally, probably very expected for guys with an expertise in this area like you and Turbo.

You see an unhit arrive at say spin 55, then it hits again at 78, 86, 90, 110, 117, and with the 1-2-4-8-16 progression it's so profitable that you wind up hugely ahead and just reset. And it happens again and again and again, session after session.

Sometimes you don't even get to the 16u level in the progression because you had so many 8u numbers hit that they propelled a +500u or more profit which forced you to quit/reset before you even got the home run ball.

Fun as hell.

 

So maybe it's better to leave all unhit on the table once they have showed up but adjust the progression?