Roulette Ask Me Anything About Betting the EC's (Even Chances)

s oxfan were you in hiding ?

 

Greed is great & controlled ..

 

That is explained by variance n×p×q where n is the number of spins, and p = 18/37 is probabilty to win and q = 19/37 is the probability to lose.

So in 37 spins the variance is 37 × 18/37 × 19/37 = 342/37 = 9.24

The standard deviation is just the square root, so 3.04, 2 standard deviations is ≈ 6 for 95% confidence, 3 standard deviations is ≈ 9 for 99.7% confidence.

Given 37 spins we expect 18 ± 9 wins on an EC which is up to 1.5x what you expect based on the mean (37 × 18/37 = 18).

 

It's like calculating variance.
I need something else.
I'll ask the question differently.

Let's say we're not talking about roulette. Let's say it's disco throwing. A discus thrower does training and throws the disc 100 times.
After that, he has some kind of variance in the distances the disc has flown. He knows the maximum, the minimum, the average, and the median.

It is likely that the variation forms a Gaussian curve with a peak not far from the average.

Suppose the minimum is 52 meters, the maximum is 70 meters, the average is 61, the median is 62 and the peak is 63 meters.
Supposing knowing all this, the thrower marks an interval of 66-60 meters, i.e. +/-3 from the peak of the curve, and counts that 50 throws of 100, i.e. half of all was in that interval.

For some reason, he needs to know with what probability the disc will hit the marked interval at least once out of three, once out of four, once out of five, etc, and how that probability will change, if he will hit the interval 60 times, 40 times ...
( here also can be interesting to have probability to hit the interval 2 times from 5 or 2 times from 6 or any other combination )

The talk is about the next training with the same 100 throws.

This example is not about roulette, but me very similar situation and the same problem have in the roulette game.
I did not find a way how to calculate, then did simulations and choose results experimentally.

However, there is no certainty that what I have received and what I use are correct...

 

Maybe you want the binomal mass function to compute the probability of a certain number of occurances within a defined number of trials.

 

Frozen tundra would be my guess. Lol.

 

Something similar, but I wonder how this is affected by the total number of hits.
For example - I predict 18 numbers and know that I hit them an average of 57 times out of 100.
I want to calculate the probability of not hitting them 4 times in a row.
Or what is the probability of hitting in 7 spins less than 3 times?
All that is as examples...

And how would those probabilities be affected if I averaged 52 hits, 47 times... and so on.

 

Easy, the probability of not hitting 4 times in a row is:

1-(57/100)^4 = 0.8944 = 89.44%​

Binomial mass function for strictly less than 3 is:

0.1282 = 12.82%.​

You can use these online calculators or excel:

• https://www.gigacalculator.com/calculators/binomial-probability-calculator.php
• https://stattrek.com/online-calculator/binomial

 

TwoUp, you have demonstrated over and over again why the Martingale progression will not work unless you have an edge. Do you recall me ever coming back to you arguing otherwise? Have I not been up-front with my postings never claiming to have an unbeatable strategy? Have I not revealed exactly how my strategy is played so it could by tried by anyone and critisized after trial? Have I not shown pictures of my win/loss documentation just to show where I was at with it at the moment? Do you recall me bragging about that? So, what I am wondering is why there are those who jump down my throat about any of this. Am I missing something that you are aware of in my approach of sharing only what I have done up to the moment?

 

I always forget that are such calculators . Thank you for reminding me.

I will try them and will look if that is similar to what I got accordingly to the simulation program.
I many take data from my real games, of course, the amount is not big data, but it is real.

I have developed some betting strategy it is slightly similar to that Massionello's betting only trigger for changing betting value is not lost or win...
Or say lost or win is not the main factor...
But overall it let me win with quite a high %, even in my understanding that % is too high and that is the reason why I look where and what is not good, maybe I am simply in some good flow... hard to believe that all that "must" be...

 

Who do you have in mind?

 

These forums are a crazy place with very hostile personalities. When I first joined I almost deleted my account, because it was like defending 1+1=2.

There are some pretty strong views against the martingale, which is fair as many have been burned by it.

I've tried to be balanced explaining the trade offs, even working out the plausibility of your results with confidence intervals. I use a mini marty in my tiers based positive progression.

I have also explained the fact that a martingale is the only mathematically certain way to beat the house, but if and only if you had unlimited funds and no bet limits.

Fundamentally the math says the probability of winning approaches 1 or 100% with a martingale. But it's a dead end because of the limits and fund requirements.

Ultimately you will too leave it behind at some point when you get the slight run of variance that goes against you. That positive variance that has lifted you up takes the stairs and is much slower than the rapid slippery slide down.

Every simulation I've ever run shows this behaviour, a nice slope up and then kaboom, through the floor on a few streaks.

 

Benas I have a solution to get the median using a very easy calculation where you get the 50% likelihood of something happening during a specific occurrence.

For example, let's say I want to know how many singles I can expect before I get a series of three and get a 50/50 situation and the median value.
Then you take a sample of random bits from random org with the numbers 1 & 2.

Now start counting each sequence with singles and write down the number of singles before the series of three or higher hits.
Let's say you get the following results.

1 5 15 2 6 8 2 4 9 2 3 ... and so one, maybe you get a sample of 100 sequences.

Now you put them in order where 1 1 1 1 1 1 2 2 2 3 3 4 4 4 4 5 5 6 and so on
Then you calculate from left to right and cut the amount by half.

That number you get in the middle is your median number where you have a 50% likelihood to reach below and 50% to reach above.

Also works fine with numbers where one single number might have the median value of 16 where the expectation is 50% to get that number hitting within 16 attempts.
You can use this formula on numbers or trends or anything you like.

I attach the original file that makes a much better explanation.

 

Thanks, I read that but did not understand what the author want to say...Find the median is not hard that is simply counting, but how to use that for benefit ...

In the play, we do not know which will be median ...to use some old?

 

Well, the occurrence of the recent variance during testing will affect the results.
Let's say nr 22 has a median value of 16 over 1000 trails.
Then you look at the random distribution, for example, each time nr 22 hits.
Then you have a 50% likelihood that nr 22 will hit again for the next 16 trails.
If the current stretch of variance would change that number would not be correct.
But is apparently gamblers who use such parameters when gambling.
To describe the phenomenon as periodic cycles.
The reason is that there are as many singles as series no matter numbers or even money.
Now someone might want to wait for fictive losses and apply progression while others want to flat betting.

 

If there is no advantage over the casino, martingale will not help, but by studying statistics you can have it or be in the near-zero zone and then a variation of betting may be beneficial.
But again, martingale is a progression that inflates dispersion, and using it you will lose simply because of dispersion.
If you compare the martingale and say simple D'Alambert you will see that most of the time D'Alamber is ahead of the martingale.

 

Yeah that was what I thought too, the median would have to be from past results, or a running median that gets continually updated which would be a pain for B&M.

 

Yeah I have posted a comparison of progressions previously...

https://www.gamblingforums.com/posts/145941/​

 

The main place in the game is where to bet. We won't win if we bet where the ball won't hit...

Finding such a place helps statistics if you can accumulate a lot of data. If not, then only correct calculation can help about what will be, assuming there must be a natural distribution giving a Gaussian curve, and then naturally bet where we expect the peak of that curve.
I haven't seen a player play differently and win consistently yet...

 

Well as I've shown for 99.7% of all outcomes with 37 spins we can expect any EC to hit 18 ± 9 times, so as little as 9 and up to 27 in a 37 spin cycle.