Questions about Roulette Mining (RM) program - atoms, algorithms, offers

foxography wrote:
- Will it be possible in the future to edit the entry level of the existing deposit, which Shpilevoy wrote about a little earlier? For example, today I have $50, this is the so-called first level. Based on the size of the deposit, the bot could regulate the size of the bets proportionally to the risks. As an option. I just asked. Well, no, so no

We build the game with even entries of 1000 chips. Depending on their financial capabilities (or better yet, on the profit generated by the bot), the player decides on the SCALE of the upcoming game.

The scenario is as follows:
If you have $50, you can play 0.10 and count on 50% working leverage of 500 chips.
The bot lost several $50 entries, then withdrew the balance to $200.
The player returned $100 to his game fund and then played a full session with the remaining $100.
Bot made x5.
The player withdrew $100 into the game fund and played with $400 at a scale of $0.20.
If you're lucky, $400 turns into $1000.
Gradually, the player increases his game fund, creates a safety cushion for future games and receives a hefty profit from casino mining.

Casino mining develops further depending on current results.

Today we analyze the statistics of the RM program games and for the sake of extended session histories we have lowered the aggressive swing of dispersion, so exits to x5 x10 are quite rare. But we control the potential of powerful passes and as soon as their share becomes stable and frequent enough, the program will start showing really amazing results:

RM task for 09/25/20.
It is necessary to enable high dispersion of bets and remove numerous missed spins. The RM program should enter the game more actively and periodically show x10 passes, and our task is to increase the frequency of such passes.

Ok. We're working.
Ideally, you need to achieve a positive longrun. Then, so that each player with RM gets an automatic opportunity to trust the virtual Shpilev with their money. To do this, the program must very closely copy the master's habits.

Thanks for the updates, continuing work on developing and improving the bot.
Question with a prelude:
It seems that in the history of this thread it was noticed that sometimes only 1-2 atoms carry the session. Quite a common phenomenon and there is nothing unusual about it. Someone suggested to alternately turn on atoms and see which one carries, which one does not and, accordingly, make a decision either to adjust the game tactics, or to continue it "as is", or other options.
On the one hand - you'll get tired of turning atoms on or off every spin, of course. But on the other hand, you can see which of the turned-on atoms is "working" by turning them all on in the "expert" mode (I don't remember what it's called exactly, and I'm too lazy to look for it, sorry). Right?
Now the question is: how can these procedures be made more convenient? Is it possible, in principle, to see active atoms WITHOUT including them all, working only on some of your own set of atoms in a given period in a given session, but also not to miss the moment when another atom begins to actively bring profit? And whether to include it or not, the decision will be up to the user. The main thing is that a person sees maximum statistical information and can make timely decisions on it with the support of a bot. As an option.
Or something like bot recommendations. Like, dude, you're working on the included atoms "1, 3, 5, 6", there's an alternative to connect #2, the dynamics are positive there.
Attention! I understand perfectly well that this very positive wave has begun and can end at any moment, and with the connection of new atoms, the rates will inflate even faster, the bankroll may simply not withstand and burst, but what I wanted to convey is that maybe greater maneuverability and greater information in front of the user's eyes will still allow him to make better decisions? In tandem with a bot, you will get a unique and powerful weapon in front of the casino.

I tried a super perversion. I launched a bot on 10 emulators at the same time. But each had only one atom enabled. I fed the same data to each bot. I saw how this barely perceptible moment of confirmation of the work of a specific atom was born. One or two. Or three. Or drained everyone
I worked on active atoms for a while. There were ups and downs. And I saw it all, could partly control it, read the situation. BUT! It's such a perversion that to hell with it. My laptop is already weak, and when he saw what I was up to, he started calling 112 and asking to take me away for treatment. In general, everything was very slow and very resource-intensive. At the same time, I got the result, took a sedative, made conclusions. Here. And I wrote a question. Sorry if it's stupid.

foxography wrote:
Now the question is: how can these procedures be made more convenient? Is it possible, in principle, to see active atoms WITHOUT including them all, working only on some of your own set of atoms in a given period in a given session, but also not to miss the moment when another atom begins to actively bring profit? And whether to include it or not, the decision will be up to the user. The main thing is that a person sees maximum statistical information and can make timely decisions on it with the support of a bot. As an option.
Or something like bot recommendations. Like, dude, you work on the included atoms "1, 3, 5, 6", there is an alternative to connect #2, there is a positive dynamic there.

Our ultimate goal is full automation. Eliminate the human factor and guessing (choice) options.

If you write down the algorithm of your actions for choosing the best combination of atoms, all this can be automated and the program will do it itself.

Alternatively, each RM spin goes through all the combinations of atoms, superimposes them on the previous game, and looks for the maximum credit point. The combination found is activated for the next prediction, and all other atoms are disabled.

To do this, you need to keep a history of which atoms worked on which spin and have 2 calculation passes - a rough one (enumeration) and a clean one (actual history). Only then will the balance not break and reflect actual events in the real game.

We will come to this version of the internal expert later. Now it monitors entry points and the general BRM for a fixed set of atoms that the user included at the very beginning of the session.

Our ultimate goal is full automation. To eliminate the human factor and guessing (choice) options.

Yes, I remember. This has been written about more than once. I understand and support. In my messages, I myself draw attention to the fact that they are dominated by an attempt to control the user over the bot. It is explained simply - the algorithms used do not yet bring the expected result and by interfering with its work, an experienced user is able to correct the situation when the casino reads the bot like an open book and stupidly drains. Or the bot is not yet able to recognize all the variations, its full potential is not revealed or or or.

If you write down the algorithm of your actions for choosing the best combination of atoms, all this can be automated and the program will do it itself.

I'll think about it, I'll try, but there are problems with this. The same as teaching AI human knowledge, life experience. In our case, gaming experience.

Yes, and to explain the choice of the best combination of atoms, each spin requires seeing the behavior of each atom.
Um... and here, so to speak, there are some minor difficulties, which I tried to complain about in the previous message. As my laptop would say, "what a minor difficulty"!! But I will, of course, try. Maybe I can formulate something intelligible

Alternatively, each RM spin goes through all the combinations of atoms, superimposes them on the previous game, and looks for the maximum credit point. The combination found is activated for the next prediction, and all other atoms are disabled.

To do this, you need to keep a history of which atoms worked on which spin and have 2 calculation passes - a rough one (enumeration) and a clean one (actual history). Only then will the balance not break and reflect actual events in the real game.

Well, that's it! That's about what I would like. What the bot currently lacks for greater positive dynamics and more confident win passes.

We will come to this version of the internal expert later. Now it monitors entry points and the general BRM for a fixed set of atoms that the user included at the very beginning of the session.

And this cannot but please. Thank you!