Law of 3rd by Drazen on an interblock machine.

[Azim]

For those that don't want too much of tracking. Here is a system for you to look at.

I have been playing this at the Casino. It has worked for me so far. I have only gone for 3 hits. 1 hit in each dozen.
As soon as the dozen hits, I stop betting it. I do use normal progression for the dozen if it goes beyond the 35 unit count.

I have 0 in the first dozen and 00 in the 3rd dozen.

Here is the system itself:

link:://:.rouletteforum.cc/index.php?topic=4960.msg49578#msg49578


In the 2 attachments are spins for 2 sessions I had today.

[Azim]

Session 3:

[nottophammer]

Marty the topic is main roulette board page 6.
System based on LOTT- need comments

[Interceptor]

I like flaw of the third, good one, easy to play. Hello

[nottophammer]

Interceptor or what aka. you use.
Collect 200 spins from Wiesbaden. You'll get anywhere between 180 spins never 200 as dealer change.
You see the remainder of the spins.
#2 hits consecutively

Would you not WIN.

[SWEET]

Go chatgpt, upload real data of 1million spins.
Ask like this.

Xxxxx
Here real data of EUROPEAN ROULETTE.

Make 74spins. "NON-OVERLAPPING SEQUENCES."
THEN track in every sequence, how many numbers UNHIT in each 74spins
non-OVERLAPPING sequence.
No need to show step by step.
Just show the SUMMARY of UNHIT .
XXXXXXXX

Why 74spins??? Because if you ask wrongly, you get wrong answer.
37 may get 12unhit.
74 may get 6 UNHIT.

The chatgpt will show you all you need to know, and show many interesting facts and suggestions too!!!maybe you could post here, all the facts here to share with us!

[SWEET]

Sampel from my chatgpt.


The zip file has been successfully extracted, and it contains the file spins1M.txt. Let me know what you'd like to do with this file — for example, analyze the roulette spins, search for patterns, or something else.
The data has been split into 27,027 non-overlapping sequences, each containing 37 spins. For example, the first 10 numbers of the first sequence are:

[21, 19, 29, 36, 7, 7, 16, 12, 20, 2]

Let me know how you'd like to analyze or process these sequences next.
Great question — let me clarify what I meant.

When I said "no single number", I was referring to individual roulette numbers (0–36). What we checked was:

> Out of 27,027 sequences (each 37 spins long), is there any number that appears at least once in almost every sequence?



Result:
There was no single number that appeared in 99% or more of the sequences — not even once per sequence.

This is interesting because you might expect all numbers to get roughly equal exposure, but randomness causes variability. Some numbers just don't show up at all in many sequences.

Would you like to:

Lower the threshold (e.g., check for numbers appearing in 95% or 90% of sequences)?

Look for something else that happens in almost every sequence, like:

At least 30 unique numbers appear?

Some numbers repeat within the 37 spins?

Certain dozens or columns always hit?



Let me know which direction you'd like to explore.

[SWEET]

You're asking the right question — and I totally understand your goal.

✅ What you want:

For every 74-spin non-overlapping sequence, count how many of the 37 roulette numbers (0–36) do not appear — i.e., how many go unhit. Then we can get a good average number of unhit numbers per 74 spins.


---

🛠� Why I haven't run it yet:

The Python analysis tool I use to process that full dataset (1 million spins you uploaded) is temporarily down — it's a system-side issue on OpenAI's end, and I don't have control over exactly when it comes back. It usually doesn't take long — sometimes minutes, sometimes a bit longer. When it's back, I'll immediately run the analysis for you.


---

🔄 In the meantime:

You've already pasted:

74 spins → 2 missing numbers

200 spins → 1 missing number


This gives a rough idea that the number of unhit numbers in 74 spins is usually between 0 and 6, and often closer to 2–3 in real-world data.


---

📌 When tools return:

Here's what I'll do right away:

Break your 1 million spins into 13,513 blocks of 74 spins (non-overlapping).

For each block, count how many of the 37 numbers didn't appear.

Give you the average, min, and max unhit count per block.

Optionally: histogram or breakdown.


Would you like me to notify you when the tools are back so I can run this instantly?

[Interceptor]

And the question would be how would you develop a system based on this data that ordinary people would understand? The answer is that the rule of thirds is not reliable and that 37 or 74 spins mean nothing because sometimes you will get a similar result with 111 spins and there is also a 2.7% house edge.