*The Cauldron*

[insidebet]

The other night I was up 750 units (25$) each.  Then I woke up!

Seriously, I am interested in your method cos it sure is different.  Plus, god knows why, I sort of believe you when you say you actually win on the long run...

[nitrix]

I've done testing of 100'000 spins of TRNG from random. org.

I had it running on a monitor in the living room.
(10 seconds per spin so I can see the up and downs of the system and get a feel for it)

It took 11 days to process its work unit; I ended up with a bankroll of roughly 32'000$. . .
starting from 100$.

I'm actually REALLY impressed.  So far it's the only winning system I have seen on this forum!

I sure hope it's not a programming error.
The code is 300 lines long and everything is heavily documented, making the double-checking not always as easy. 

So I'll re-implement it in another language with safer classes. . .  you know, just to see if the results are consistant

[bombus]

Thanks for posting nitrix.

Nice result, and yes, let's hope the re-implemented code performs as well.

Did you run the test flat betting, or using the '1-2' progression?

I'd love to see your coding of the system, so PM me if you would like to share it.

In a way I hope there was a programming error because then we would have 2 winning systems! Lol.

Cheers.

[insidebet]

Nitrix,

You say you want to code it in another language.  Do you know Roulette Xtreme?  Could you possibly do it there?

Thanks.

Insider

[nitrix]

It was loosing money in the long run flat betting, so I added your 1-2 progression.  Then the bankroll went down.

No difference with all other systems there. . .  but I didn't want to give up already so I looked back into the code again, inversed and moved the triggers for the 1-2 progression (aka when it must switch to 1 and when it must switch to 2) and then the bankroll skyrocketed !

Beleive me, at 3 in the morning, that had nothing rational, I was just messing around like "I wonder if" and it worked.

I'll clean the sources (actually written in PHP).  Also rewrite it in C++ for god's sake. . .  and if it still works, implement a user interraction, like you give the program the numbers and it tells you what to bet for other ppl on the forum.

I won't write RX code because my main goal is to have a working bot, that can play Roulette online alone while I'm doing something else (: RX code is so simplified it makes it hard to understand for me ahah.

Anyway I'm not selfish
I'll let you guys know when I'm done and publish my work whatever the outcome is.

Back to coding (;

[bombus]

Interesting, nitrix.

When you're done we will see if you have got it right or not.

I could test it for you as I would spot any mistakes fairly quickly.

I suppose insidebet would like to see an rx script for this because with rx you can rapidly go through many thousands of spins, even a million spins literally within minutes.

It makes sense to me to test with rx if you can, before the drawn out live tests.

Good luck.

[bombus]

Don't forget,

Even though many people say it shouldn't make any difference, this is a wheel order system and as such was not designed for rng results.

Both appear equallly random but appearance isn't everything.

[nitrix]

aaaaand. . . .  it was a programming error.  Once I fixed the error, the program couldn't make any consistent winning.  So I'll stay away from it.  Maybe it's because I'm using RNGs and you warned me agaisn't them. .  anyway.

I'll look for something else (:

[bombus]

Ok.

Good luck, nitrix.

For the rest of you I would suggest you do your own testing and draw your own conclusions, because I can almost guarantee that nitrix got it wrong somehow. I'm not saying this is an infallible system, just that it is very likely nitrix's code was not accurate.

I say this because I have some experience with coding systems for testing and for bots, and from my experience it does not matter how clearly you initially explain the procedure there is always some correspondence between designer and coder before things are absolutely correct. In this instance I never heard squat from nitrix - not one question.

So everyone should decide for themselves.

Cheers.

[bombus]

Actually, this is easily validated.

Nitrix,

I will post a few hundred numbers here (say 300). Even better, you can post 300 numbers. You can run them through your program, and I will play them manually. We both use the "1-2" progression.

If our final bankrolls match up, then you have coded the system correctly.

It wouldn't take you long to run 300 numbers, so what do you say?  

[balint]

Ok.

Good luck, nitrix.

For the rest of you I would suggest you do your own testing and draw your own conclusions, because I can almost guarantee that nitrix got it wrong somehow. I'm not saying this is an infallible system, just that it is very likely nitrix's code was not accurate.

I say this because I have some experience with coding systems for testing and for bots, and from my experience it does not matter how clearly you initially explain the procedure there is always some correspondence between designer and coder before things are absolutely correct. In this instance I never heard squat from nitrix - not one question.

So everyone should decide for themselves.

Cheers.

I agree with you Bombus everyone should decide for themselves.

**does not matter how clearly you explain the procedure there is always some correspondence between designer and coder before things are absolutely correct**

True and not any method can be coded at 100% all rules (same thing happened to me).
One more thought....human decision is not same with machine decision(always same decision),
we the humans do make changes according with the prezent situation....

cheers

[Bayes]

Actually, this is easily validated.

Nitrix,

I will post a few hundred numbers here (say 300). Even better, you can post 300 numbers. You can run them through your program, and I will play them manually. We both use the "1-2" progression.

If our final bankrolls match up, then you have coded the system correctly.

It wouldn't take you long to run 300 numbers, so what do you say?  

I say - good idea!

I'm going to code this system over the next few days, so it will be a useful check. 

@ balint,

I disagree with you in regard to coding. Any system can be coded as long as the rules are clear and unambiguous. If there is any subjectivity then the system creator is just guessing, in which case you can code in the "fuzzy" logic or make a random decision (within defined parameters). The point is, a program is just a series of instructions, and if the system creator can't explain them clearly enough, that's not a weakness of the computer, which is just a dumb box of chips.

[bombus]

That's great, Bayes.

Same offer for you. When you're done, post some numbers and I will play through them for you as a check reference.

It would be equally great to prove or disprove the system.

Cheers, and nice to see you posting here. 

[nitrix]

Okay, I like the sound of it (: It's true that a computer can't take guesses.  Like "Raise or hold" and "lower or hold" your bet. . .  that's not a decision a computer can do.  It needs triggers.  Like if bankroll > than our last bankroll, then we can take a guess. . . . . . . .  or just increase/lower it anytime you face the choice. .

It's ridiculous how a simple thing can add-up in 100'000 spins !

Anyway, adding a few `verbose` so you can a a visual of what the program's betting and when.

/*
    STAGE 1 = Mark off around the wheel the last 10 individual numbers hit.
    This could take as little as 10 spins or many more – I've seen it take up to 21 numbers.

    STAGE 2 - Now starting from the zero (or if zero is part of the first group of 10,
    the next available un-hit number in a clockwise direction (you can go anti clockwise if that is more comfortable)
    mark each un-hit number in succession around the wheel with the numbers
    1-2-3-1-2-3-1-2-3-1-2-3-1-2-3-1-2-3-1-2-3-1-2-3-1-2-3.

    STAGE 3 - We use the next spin number as our first trigger.
    Note what group that number belongs to and bet that group once with 1 unit.

    STAGE 4 - Betting the same group as the trigger.
    STAGE 5 - Betting the same group as the first trigger & the second trigger.
    STAGE 6 - Betting the remaining group (gambler's fallacy: being due).

    STAGE 7 - Do nothing.  Take the losses, wipe off and start over.
*/

Every numbers spun are marked with an arrow like this
> 13(2)
This is number 13 spun, and the program even show us what group it belongs to.  (Of course at Stage 1 it cannot predict in what group they will be marked so -1 appears).

I added a few commentary when entering a new stage like so:
[ *] Stage 4 - Betting the same group as the trigger (1)

Almost everywhere in the program (where I though it was relevant), I have placed outputs like so:
Bankroll: 1076

It also tells when the program detects a win or a loss.

ANNND

what numbers are being bet.

Placing 2 unit on numbers 32 25 6 11 10 33 31 29 12
(Temporary) Bankroll: 1058

Temporary bankroll is just the interval between the chips placed on the board and the actual wheel spinning.

Good luck bombus.  If I missed anything, let me know.

P. S: Okay it made 65 units in 300 spins. . .  sometimes it'd go up 1200 or lower, or higher, back and forth.  But it was definitively losing in the long term for me.  In fact, with a 1000 unit bankroll, it cannot go past 15'000 spins.

[nitrix]

Oh and I changed the progression to match yours as close as possible with the info given:

Multiplier is set to 2 on a win
Multiplier is set to 1 on a lost

(This is the amount of chips placed on every numbers when betting).

[nitrix]

-