t-Luck Algorithm

How to measure luck

Accurately measuring luck, or rather trying to predict the gaps of a roulette chance in the short term is pure utopia, however as the number of spins increases, thanks to statistics the forecasts begin to become less and less approximate. in essence, the gaps that determine our luck or misfortune in betting a chance at roulette, are actually measurable.

A possible way to measure gaps is the one already described in โ–บ this post, when I tell you about the famous Marigny coefficient.

However, the Marigny coefficient has limits, as it is based only on opposing and equiprobable chances, ie without taking into account the presence of zero, which unfortunately constitutes a serious error of assessment.

In fact, if we consider for example 40.000 spins on roulette, according to Marigny we will have that our maximum luck (equal to 5 times the square root of the spins played) will be 1.000 units won, but it is a pity that in 40.000 spins we will also have encountered 1.081 times zero, so as you can see with roulette bets on Red or Black at even mass (flat bet), reaching 38.000 / 40.000 spins, due to zero it is mathematically impossible to win even a single unit!

This limit, however, is much greater if we consider the bets on the single number, in this case in fact by always aiming for even mass (flat bet) we can survive even over 200.000 spins!

The simulation of the previous image was obtained with the software bot โ–บ Roulette Bias Sniper, as you can see after 215.000 spins played flat bet, there are still 2 numbers that would have made the player win the equivalent of about 30 single winning numbers, so over 1.000 units! But this is a topic that we will discuss in more depth in another post.

Another method of measuring gaps, but much more precise than the previous one, is the โ–บ Student's t-distribution, which I will illustrate to you immediately.

The first pillar of this method is the unit of measurement for gaps, called standard deviation (sqm).

The standard deviation is equal to the square root of the product of the total number of events (n) times the favorable probabilities (p) and the opposite probabilities (q).

sqm = RADQ (n * p * q)

for example if we consider 1.369 spins of roulette we have

sqm = RADQ (1.369 * 1/37 * 36/37) = 6.

The second pillar of the t student รจ average of an event (m), which is equal to the product of the number of events (n) and the favorable probability.

m = n * p

again in relation to the 1.369 spins above, if we consider a single number, we have:

m = 1.369 * 1/37 = 37

These two values, mean (m) and mean square deviation (sqm), are of absolute statistical value, because they allow any gap to be reduced to the same unit of measurement, regardless of the event in which it occurs.

This important reduction is achieved precisely by t student, which is the ratio between the deviation (understood as the difference between the favorable events U and the mean) and the mean square deviation.

We therefore have that:

t = (U - m) / sqm

Again in relation to the hypothetical 1.369 throws of a roulette ball, if for example the number 13 comes up nineteen times, we have that

t = (19 - 37) / 6 = - 3

The + or - sign indicates hyperfrequency or hypofrequency.

The coefficient t student it is therefore very useful because there are statistical tables that can also be found on the net, which indicate exactly the percentage of probability that certain values โ€‹โ€‹of are exceeded t.

It is commonly assumed that the maximum limit of t student be equal to 4, that is the statistical limit for which it is agreed that the probability of exceeding it is practically nil.

Before continuing remember that on ThatsLuck you can also find free content, if you want to stay updated on publications subscribe to the channel on โ–บYouTube.


The 2 mistakes of Marigny

Clarified what the t student and how it is calculated, I tell you right away that this method of measurement is decidedly more appropriate than the Marigny coefficient, because in the results it produces it also takes into account the tax (zero).

A big mistake of Marigny was to think that once a chance reached difference 3 or higher, it had to necessarily re-enter, so he suggested to aim for the immediate re-entry of the gap.

The first mistake of Marigny was not to consider zero, because if it is absolutely true that the gap must be returned, it is equally true that no one can establish a priori in how many strokes this gap must occur.

If a chance reaches for example gap 4 (very high Marigny coefficient since the maximum is 5), who can assure us that a phase of alternation between red and black that lasts even hundreds of spins cannot begin?

Not bad, someone will think, in the alternation phases you do not win but neither do you lose ... but no, because in any case the zero will come out according to his expectation, eroding in advance all the advantage that we could achieve when the gap really returns towards the natural equilibrium.

Second and most serious mistake of Marigny: considering the spins collected over several days and from different roulette as a single permanence (also known as "personal permanence").

I empirically tested this fascinating concept and after a few million simulated spins I came to this conclusion: for the purposes of concrete statistical reliability, the gaps of roulette must be measured exclusively in a series of spins referable to the same generator that produced them. in an uninterrupted series of launches.

In other words, if we want an analysis of 1.000 spins to be reliable, we must record 1.000 spins continuously at the same roulette and not for example 10 tranches of 100 spins taken on different days and from different roulette.

Always remember this concept in the future, because it is very important and obviously does not apply when we are looking for a roulette bias, because in this case the sum of all the data will still be indicative, indeed it will confirm the presence of the defect or not, but this too is a topic already covered in a โ–บ other post.


t-Luck Algorithm (the theory)

Now let's see on which statistical assumptions I based the new software t-Luck Algorithm.

Let's analyze the table above again:

Based on the data reported, if for example the red reaches a value t student equal to 3,00 means that the probability that this value reaches 3,50 is just 0,02%!

In reality, however, this is not the case, because perhaps the question we should really ask ourselves is: once a chance reaches t = 3,00 how many times does it arrive at t = 3,50? I have not yet done this verification, but it will not take long and I imagine that the table above should be read more correctly as follows: on an indefinite number of tranches of 1.000 spins those that will have a value of t = 3,00 will be 0,13 % while there will be no tranche with t greater than 4.

However, wanting to consider as reliable the suggestive hypothesis that a tranche with t = 2,50 can exceed t = 3,00 only in 0,13% of cases, I wanted to set the t-Luck Algorithm on a particular logic, in the sense that both the Marigny coefficient and the t student, when they reach extreme values, they are in fact representing a very strong trend of a given chance, which as we have seen before, could return after who knows how many hundreds of spins, while we continue to pay the tax at the counter due to zero.

To confirm what has been reported so far, I propose these two graphs, referring to 1.000 spins analyzed both in relation to the value t student (first graph) and the trend of the gap of the Red chance.

As you can see, the first graph confirms that once a value t = is reached -2,5% after about 200 spins (we are therefore in the presence of a hypofrequency of red, i.e. black has come out many more times) the value of t student begins to rise, indicating that the Red chance gradually begins to rebalance its frequency with respect to the opposite Black chance.

The rise, however, is not sudden, but we see that the balance (value t student close to zero) practically reaches 1.000 spins, so we play about 800 spins in which we pay the beauty of 800/37 = 22 zeros and in fact as you can see in the second graph due to zero the hypothetical cash of the player who started betting after 200 spin (cash / gap value -45 in the second graph), closes the 1.000 launches with a handful of pieces won, because most of the advantage deriving from the gap closing was eaten by zero.

What would have been the optimal strategy for the player in this case? It would have been to start playing at t = -2,5 (at spin 204) and stop as soon as a few pieces of profit have been obtained (at spin 246) with value t student climbed back to -2,00 thus winning 3 pieces of profit. Seems little? The player in question would have won 3 pieces in 42 spins, or 7% of Roi!

From all this derives ours first rule: start betting only when the t student reaches a value of +/- 2,5 and stop as soon as a profit is made.


Middle Trends

The second pillar of the t-Luck Algorithm is to look for this value of the t student 2,5 not in the chances that go into a strong gap as in the graph above referring to the Red, but in the chances that instead have a more stable trend, softer than the others and that I have renamed with the term Middle Trends.

But if these chances don't have a big gap, how do they reach the value t student 2,5?  

Here's an example of what I mean by right away Middle Trends.

The two graphs above always refer to the Red chance, this time simulated on 100 spins.

If you look at the first graph you will notice that the value t student enough left stable, namely between +1 and -1,5 in practice, in the first graph this value obviously started from 0, then rose to +1, then fell to -1,5 and finally returned to +1.

So far nothing strange, but if we count the value t student according to minimum and maximum values reached we will have that from +1 (max) it dropped to -1,5 (min), so there was one deviation between minimum and maximum value of + 1 / -1,5 or 2,5 points!

Here we have found our reference value 2,5 and therefore when around the spin 20 of the graph the gap of 2,5 has been created and we begin to focus on Red (because at -1,5 we are in a hypofrequency situation) here is that the fate (and statistics) rewards us, in fact playing up to t student = +1 we would have won 15 units in less than 80 spins!

Obviously based on rule 1 above we would have stopped after the first profit, however with this example I hope to have clarified the concept of Middle Trend and how to count the t student basing it on the gap between the minimum and maximum values โ€‹โ€‹encountered.


t-Luck Algorithm (the Software)

All clear so far? Ok, don't worry, the software will do all these calculations t-Luck Algorithm, the player will only have to enter the numbers as they come out and possibly bet exclusively on even mass (flat bet) when signaled by the Software.

After activating  t-Luck Algorithm with the code you already know how to find, just open a game table and start entering the numbers that have already been released, to do so just click on one of the buttons in the central column numbered from 0 to 36.

When you click on a number, it also appears in the box at the bottom left (Last) as our reference reminder.

Be careful when you register the numbers, because if you enter a number wrong there is no way to fix it and you have to click on the logo ThatsLuck at the bottom right, which basically resets the session and then you will have to start all over again.

In practice there is nothing else to do, when one of the chances to monitor which as you will see are:

โ–บRed / Black

โ–บEven / Odd

โ–บLow / High

โ–บDozens

โ–บColumns

โ–บSestine

produces a student t-value gap of 2,5 immediately in t-Luck Algorithm a warning is activated indicating which chance to aim for!

As you can see in the image above, in this case it is signaled to try to bet on the first sixth (SES 1), which as you can see in the two columns on the right (which represent the Frequency of sortie of the various chances), it is neither the most frequent sestina (which is SES 2), nor the least frequent one (SES 3 and SES 6 never released).

In the event that a number between 1 and 6 should come out, the value of the student t will drop below 2,5 and then the warning will disappear, clearly until there is a warning you do not bet and simply record the winning numbers according to their chronological order of release.

Obviously it will also happen to bet more chances at the same time and, in this case, you could try to bet even a few units of lower value on the numbers in common between the chances to bet, just like I did in the image below, where I crossed the COL 1 with SES 2 and therefore I also bet on the two common numbers 7 and 10.

I hope I have provided a thorough analysis of the project t-Luck Algorithm, my recommendations are quite simple: never increase your bet and establish from the beginning how many units to win before stopping (Stopwin), a value that I recommend setting at 10, then of course do as you please, as important as always is have fun at the bank's expense!