This is where I am just going to unleash some of the ideas that come into my head from time to time regarding roulette.
Feel free to add your own. :thumbsup:
Nice topic.
One idea i had is to start with the engine room first, before you plan how to play.
Make up you mind what kind of loss-limit you should have and win-goals.
That way you can not play as you might want to play.
You narrow down the game to just be able to use some progressions and not all.
Also that you narrow down the game to just be able to use some bet selection with tight variance.
Its like starting from the right end of things.
First you set you money management plan into action.
Then you pick bet selection and progression into testing.
One MM that i find being among the best i read is Brett Morton's solution.
And one other is John Patrick's solution.
Here is something that has always intrigued me.
[attachimg=1]
So that's 372 spins. Probability suggests each number should appear 10 times.
However three of the numbers (9, 27 and 34) have appeared 16, 17 and 17 times respectively.
I have looked at a lot of these bar graphs and it's not uncommon to see a few numbers showing 50% + of the expected hit rate.
It seems to get more out of whack once you start going past the first few cycles.
The day before.....
[attachimg=1]
291 spins suggesting an average of 8 hits.
Yet numbers 11, 12 and 14 are working overtime.
None of the other numbers really get close.
There is a decent method in here somewhere.
Bally do you read and speak german.
Cheers
Hello Sputnik
No! However my Brother-in-law is German and I am always asking him to translate things from the 'paroli' forum that I have trouble in understanding even after basic google translation.
cheers
The best way to go forward with this in my opinion is to go through some cycles and see how many numbers can keep close to the average at each stage. I want to find out if there is a point where a higher proportion of numbers start lagging behind. That might then give me something to work with.
Here is the first cycle from a set of 370 spins I have.
There are 15 numbers that have not appeared.
I am going to go through 10 cycles just to see what happens!
[attachimg=1]
After the second cycle and there are 17 numbers that have not shown up the average of 2 hits over the 2 cycles.
[attachimg=1]
[attachimg=1]
[attachimg=1]
So just over half of all the numbers have not appeared on average 4 times over the 4 cycles.
I am breaking it down cycle by cycle because I want to take a good look at things. The danger as always with roulette is in making quick assumptions. I need to try and find something which happens consistently over all these cycles or maybe it's a case of trying to find an extreme situation which only happens at certain points.
[attachimg=1]
One thing I noticed looking at the pic is how six of the numbers have appeared in every cycle. I put a red dot next to those. Could be something or nothing.
[attachimg=1]
If you look back in the forums archives for posts by me tracking numbers appearences over X spins, if that's what it is you are doing, I have done this over 100s of thousands of spins RNG real and airball, YES there patterns BUT at no stage before you have your results can you determine where to bet, reliably anyway, you get the same percentages as you would expect with any inside number method. Good thinking though, funny what you find/see happening.
If I aint following properly, give us a brief idea of what you are looking for, I may already have it coded somewhere.
One thing this shows me is that playing lots of numbers is a disaster waiting to happen. You are going to get hit by some bad variance when your group of selected numbers matches the majority of these underperforming numbers and there are a lot of them as you can see.
However there does always seem to be a small group of numbers that are performing at a rate of 50% better than average. The problem as you have likely figured is these numbers can change quickly. A hot number can go to sleep and a cold number can wake up.
So maybe one way of catching that consistency is to look for numbers that have appeared regularly in the previous cycles (see the red dots in the pics) You could also keep an eye on this for waking up numbers.
Superman
I am just gathering some data at the moment and then I am going to have a look at it. There are no real concrete ideas just yet.
cheers
Bally how about to do things the other way around.
Lets say there exist a threshold where numbers divide naturally from being hot or could.
I find the spread between 15 and 21.
Assume you need a past sequential window with no repeats to catch does hot repeating numbers being slight biased or due towards random fluctuation.
So when you play a real wheel, then you know that if the wheel has a slight bias you will be on that train and if not then you will catch does hot numbers.
We can discuss this in private if you want and i can show you some very interesting information about why past spins can influence future ones.
Is not made up from me, it is based upon and from names like Laurance Scott and Snowman.
Methods like this on good days can reduce the house edge and if we use the Kelly staking plan we might have the best there is catching slight bias numbers or hot numbers based upon one physical phenomena.
Nice thread Bally.
Making observatory pit stops along a growing number stream is a good way to approach roulette, I think. So you should get some strong methods out of this type of analysis.
You are basically looking at averages to see if there is any edge obtainable. You are stopping every so often to reassess the averages. By using averages you can group together and quantity of numbers you like.
Why not stretch the averages from two ends then bet for a correction?
One end would be to find the average of 1 number hitting in 36 spins = 1 then shorten the spin count to 27 spins while still focusing on numbers that have attained the average for 36 spins = 1.
So a pit stop every 27 spins instead of every 36 spins, but using the average hits for 36 spins.
The other end to stretch would be to find the average of 1 number hitting in 36 spins = 1 Then only focus on numbers that have hit at least 1 more time than the average.
So a pit stop every 27 spins instead of every 36 spins, and focus on numbers that hit at least 1 above the average for the spin count.
Pit stop @ 27 =1 Focus on numbers with 2 or more hits.
Pit stop @ 54 =2 Focus on numbers with 3 or more hits.
Pit stop @ 81 =3 Focus on numbers with 4 or more hits.
Pit stop @ 108 =4 Focus on numbers with 5 or more hits.
Etc.
This should show really hot numbers.
But how to bet them? Well I wouldn't.
I would group together a few numbers then bet for an average correction of sorts.
Let's say you have tracked 81 spins and numbers 6 & 14 both have 5 hits, and numbers 18 & 27 both have 6 hits. Everything else is hitting at or below the average.
Make some 3 number clusters. They would be 4.5.6 – 13.14.15 – 17.18.19 – 26.27.28
The middle number from each set has hit above the average for 81 spins. They might continue to hit or they might go to sleep. Don't bet them. Bet on the numbers either side of the hot numbers to maintain the high average.
The bet would be 4.6.13.15.17.19.26.28.
An eight number bet in this example. You can only bet 8 numbers 4 times before a progression step is required. Your choices are to stop after any win, or stop after 4 losses, or use a martingale type progression on the betted numbers until a hit.
Hope this helps with crazy ideas. :thumbsup:
Quote from: Bally6354 on February 16, 2013, 02:03:49 PM
This is where I am just going to unleash some of the ideas that come into my head from time to time regarding roulette.
Feel free to add your own. :thumbsup:
Thanks for the thread Bally.
This way we don't have to worry about reading the entire thread before posting a comment or tweak/idea : )
My first idea.
Research the new methods posted each week. Select 1 to 3 methods to play for only one week.
I think that this is better then just randomly guessing what is going to land next, regardless if it is a EC, dozen or line etc.
Someone has looked at the method and in the vast majority of cases it worked well, short term.....
Take the previous number with the time in which it was spun then calculate the next number.
32 at 16:32 X ??? - ??? =next number >:D
Regards
Max
:nod:
Quote from: Marshall Bing Bell on February 25, 2013, 10:57:20 AM
Pity this thread fizzed out.
100 or so crazy ideas might had really led somewhere.
I am still looking at it however I have not had that much time.
It did lead to the idea for the 'second cycle' thread. (although I am sure that idea is not a new one)