Roulette Ideas

The wheel was invented as a Random Generator.
Later someone thought you could bet on the results.
So the main idea of the game is to guess which number will hit in the next spin.
And then the the brothers Blanc added the houseedge.

Now our basic Knowledge:
If we bet on any single number our chance to hit is 1/37.
If we bet this number 37 times:

and don´t hit: we will lose 37 units
If hit in the first spin we will win 35 units plus get back our bet unit
If we hit in the 37th spin we did bet 37 units and win 35 units and get back our bet unit that will leave us with a loss of 1 unit (not to mention tronc)

Nowadays we know (or should know)that there is no way to beat the houseedge.
And we know that we can´t do accurate predictions.

So what can we do?
We can reduce the houseedge by betting less spins.
If we bet less spins, we will lose less
and we will be able to leave the table with a Plus in a win.

This gives us the following rules:
Bet 34 spins. So if you win only once (no matter what spin as long as we bet every 34 spins) we will win 1 unit.
But if we stopp at any win (we will be in Plus for sure)we will increase our wins.
But if we have to play all 34 spins and don´t win we will lose 34 units.
The losses will ad up to more units than the wins we collect.
Why? because we don´t win the units if our numbers appear a second or even a third time.
We need these wins to reduce our total losses.

With basic statistical math:
If we bet each number we will lose 14 attacks 14*34 = 476 units
and we will win 23 attacks with an average of 18 units = 414 units

Am I missing something here? While there is a theory of optimal stopping for some scenarios (e.g. the Secretary problem), inserting stop points in a random sequence of outcomes doesn't affect the returns.

It may work, but only once! So if you get a hit before spin 34 and quit, you can say you've won, but keep going back to do it again and the returns will be no different than if you had used a different stop point or none at all. The boundary between sessions doesn't change the randomness of the outcomes.

But maybe I'm misunderstanding something...

Quote from: BlueBuzzard on Jan 10, 2023, 10:48 AMinserting stop points in a random sequence of outcomes doesn't affect the returns.

It's for this reason that "hit & run", which is so often espoused by system players, doesn't work.

Quote from: BlueBuzzard on Jan 10, 2023, 10:48 AMAm I missing something here? While there is a theory of optimal stopping for some scenarios (e.g. the Secretary problem), inserting stop points in a random sequence of outcomes doesn't affect the returns.

As I tried to show in the other thread we have to get away from theory.
A Stopp Loss will cut off wins
A Stopp Win will cut off further wins and collect all total losses.
A Stopp Win or Loss could equal this out

e.g.
34 spins lost -34
1. spin won +34
17 spins stopp loss -17
16. spin win +18

I´m just creating a statistic showing the spin of the first win.
Due to Law of Third there should be a "break even point" (my theoretical thinking. Bound to be proofed)

There is something else, that bothers me:

If a straight hits
Also the line hits
also the double street hits
Also the double street hit
Also the finals hit
Also the wheel-sector hits

How often do we see: Just the chance next to the one I bet hits.
I bet 1 the 2 hits
I bet 8 the the 30 hits (wheel neighbour)
I bet final 9 and final 0 or 8 are the favorits

I posted a strategy on the roulette life forum that exploits this.

I might repost it in my area as some others said it performed quite well.

Quote from: TwoUp on Jan 11, 2023, 12:07 AMI posted a strategy on the roulette life forum that exploits this.

I might repost it in my area as some others said it performed quite well.

I posted the strategy in the TwoUp area here: Russian doll parachutes

Quote from: winkel on Jan 10, 2023, 01:04 PMThere is something else, that bothers me:
If a straight hits
Also the line hits
also the double street hits
Also the double street hit
Also the finals hit
Also the wheel-sector hits
How often do we see: Just the chance next to the one I bet hits.
I bet 1 the 2 hits
I bet 8 the the 30 hits (wheel neighbour)
I bet final 9 and final 0 or 8 are the favorits

you need to make up a streams tracker winkel ..then break the results down...

Quote from: winkel on Jan 10, 2023, 12:31 PMI´m just creating a statistic showing the spin of the first win.

Here's a breakdown of when the first win occurs, the number of winning sessions (out of 500,000), and the average number of spins before a win:


No. Wins @ spin  1 : 13982
No. Wins @ spin  2 : 13608
No. Wins @ spin  3 : 13010
No. Wins @ spin  4 : 12579
No. Wins @ spin  5 : 12333
No. Wins @ spin  6 : 12089
No. Wins @ spin  7 : 11921
No. Wins @ spin  8 : 11453
No. Wins @ spin  9 : 10819
No. Wins @ spin 10 : 10824
No. Wins @ spin 11 : 10522
No. Wins @ spin 12 : 10266
No. Wins @ spin 13 : 10103
No. Wins @ spin 14 :  9723
No. Wins @ spin 15 :  9344
No. Wins @ spin 16 :  9058
No. Wins @ spin 17 :  8802
No. Wins @ spin 18 :  8703
No. Wins @ spin 19 :  8239
No. Wins @ spin 20 :  8150
No. Wins @ spin 21 :  7819
No. Wins @ spin 22 :  7638
No. Wins @ spin 23 :  7367
No. Wins @ spin 24 :  7263
No. Wins @ spin 25 :  7043
No. Wins @ spin 26 :  6987
No. Wins @ spin 27 :  6675
No. Wins @ spin 28 :  6567
No. Wins @ spin 29 :  6316
No. Wins @ spin 30 :  6041
No. Wins @ spin 31 :  5904
No. Wins @ spin 32 :  5828
No. Wins @ spin 33 :  5597
No. Wins @ spin 34 :  5578

No. of Winning Sessions : 308151
Average Wait for a Win : 9.13766 spins

So the win rate is 308,151 * 100 / 500,000 ~ 62%

Quote from: BlueBuzzard on Jan 12, 2023, 08:08 PMHere's a breakdown of when the first win occurs, the number of winning sessions (out of 500,000), and the average number of spins before a win:

Thank you very much for this statistic. that helps me a lot.

Quote from: winkel on Jan 10, 2023, 12:31 PMAs I tried to show in the other thread we have to get away from theory.

Most of the time theory matches reality pretty closely. If it doesn't, that's good news because we may have discovered an anomaly which could put the edge in our favour.

In my previous post I showed the results from counting the number of sessions in which the 1st, 2nd, 3rd etc wins occurred, which I obtained from running a simulation. They can also be generated by probability, like so:

The chance of a win on the first spin for a given number is 1/37.
Multiplying by the number of sessions (500,000) gives you 1/37 * 500,000 = 13,514 (number of sessions in which you get a hit on first spin)

If the number doesn't hit on the first spin, but wins on the second, the probability is 36/37 * 1/37, (L,W), and multiplying this by 500,000 gives 13,148, which is the number of sessions in which you get a hit on the second spin.

If the number loses on the first two spins but wins on the third (L,L,W), the probability is 36/37 * 36/37 * 1/37, and multiplying by 500,000 gives 12,792 (number of sessions in which you get a win on spin 3.

Continue like this - multiplying the previous number by 36/37 (the chance of a loss) - until you have the number of sessions for all hits up to spin 34.

BTW, this is called a Geometric distribution and produces a characteristic curve. Plotting both sets of results - the previous list I posted together with those generated by theory, I got this:



As you can see, there isn't much difference between the curves!

Quote from: BlueBuzzard on Jan 13, 2023, 01:37 PMIf the number doesn't hit on the first spin, but wins on the second, the probability is 36/37 * 1/37, (L,W), and multiplying this by 500,000 gives 13,148, which is the number of sessions in which you get a hit on the second spin.

Here is the problem that confuses me.

Shouldn´t it be 500 000 minus 13 154 (hit s in first spin?

My idea is: I start 31 sessions in a month. If I hit 3 times I continue betting with the left 28 sessions and not with the starting 31 sessions.
That means I only make a second bet if I lost.

like this:

Quote from: winkel on Jan 13, 2023, 02:17 PMHere is the problem that confuses me.

Shouldn´t it be 500 000 minus 13 154 (hit s in first spin?

That wouldn't make sense. The idea is to find what proportions of the whole sample are wins after 0 losses, 1 loss, 2 losses etc. If you removed a proportion of the sample at any time it would change everything. Suppose you wanted to divide a cake in the ratio 1:2:3. There are 1 + 2 + 3 = 6 equal parts. One part would be 1 sixth of the cake, another part would be 2 sixths, and the final part would be 3 sixths, altogether adding up to the whole cake. But if you removed 1 part of the cake the original ratios would no longer apply. It might mean something, but I don't know what, lol.

Your spreadsheet is a mystery; it's just a mass of numbers with no explanation of what they mean.

Sorry for not explaining

horizontal day 1 to 31 of a month
day 1 we start betting #1 till it hits
day 2 we start betting #2 till it hits

... max 34 bets a day.
in this picture I tried a progression: If a day loses next day is to bet with +1 unit.

most right column: count of wins in spin x through all 31 days

Any further questons? Try me