Ok, it doesn’t have Croatian
But you still can do test - translate from another language say to English with a few translators and compare.
For me really looks like Deepl does that better than Google…
Thanks
I started this theme which is like a series of articles, I wanted to finish it by sharing with you some of my thoughts that I use during the game.
However, a couple of weeks ago I had a conversation with one … I can’t call him a player (you may remember me mentioning that there are players and there are gamblers) and after that conversation, I changed my mind. And now I have to think about how to end that series of articles.
I didn’t notice and started writing in a part of the forum that everyone could see. Maybe that’s a good thing, maybe it’ll work as an advertisement for this forum… I am, after all, one of its oldest visitors, and I want it to evolve somehow.
So in the near future, I will write something from my experience that will be interesting for readers, I am almost sure in that, but I can change the location where I will write that because some minds will be not for all.
So if you want to be sure that can to read, maybe think about the possibility to register for the forum.
These days the models that APs build, and the actual predictions are usually quite accurate. Far more accurate than they really need to be in order for APs to win big. There are really two main reasons for losses.
-
The AP finds that they’re betting in the “forbidden zone”…the location where your bets can actually lose at a rate that vastly exceeds the house edge. These days, that’s almost always because the AP hasn’t accounted for the RRS function. Practice predicting the RRS direction BEFORE you do anything else. or find a wheel where you can bet late enough in the spin that it doesn’t matter.
-
Variance in the scatter. Playing on a low edge wheel and not properly accounting for the scatter. We all see the little short term graphs of scatter (30 or 50 spins) but that’s not nearly enough on a weak wheel. Weak wheels require 300 to 1000 spins or more. Otherwise, you’re not going to know what you’re real edge is going to be or what to expect with variance. The scatter looks good for short periods and then “whiff” gone again. Players need larger samples in order to know how large the variance will be. If you’re playing a low edge wheel (5 to 7%) then you’re going to have losing days or even weeks if you’re not playing for very long. The lower the edge, the longer you need to play. Monte Carlo simulations will help the AP better grasp the fluctuations and will provide another level of confidence when playing.
Tracking prediction to strike number will also enable the player to overcome both problems above.
What is Monte Carlo simulations? Thanks
You meant monte carlo method?
What I’m referring to with “Monte Carlo simulations” means, testing on a biased RNG stream that replicates your edge, and the width of your betting arc.
Here’s how you can do it. Create a biased RNG function. Next, drop the RNG stream into a betting simulation or you can play it manually. Graph the results in bundles of 50, 100, 200 etc and see how often your biased section wins or loses. Calculate the results.
If you need a stream, then let me know and I can create one for you.
1 Like
Yes, please, that would be great if you can send me one.
I can generate a stream for you. How big would you like the edge to be, and how wide is the arc?
Hello Mr. Snow
Welcome back, hopefully snow storms will be reigning across Europe soon again.
Just a quick one, do you recommend for weak wheels to collect 1000spins for both directions or would it be for both directions.
Regarding the ball scatter…I know some will disagree with me, but if you know the wheel is weak, then get a real sized sample and keep at it until you see the section bell up at or above five standard deviations. The more spins you get, the better.
Another word for brute force method
Or trying out 1 billion spins on a computer and seeing Odds from the trial results