Does cold become hot?

[superman]

Over the last few weeks there's been plenty of discussion about hot and cold numbers so I decided to code a tracker, it runs against BV NZ rng in real money mode (all my tests are conducted in real money mode), what it does, see screenshot, it tracks each number and displays the gaps between hits, what I did notice was that after a number has slept for over 100 spins and then hits, very often it hits again within 10 to 20 spins, there will be the odd one that goes back to sleep as always but I have yet to see back to back > 100.

I have circled the gaps in red, most recent gap is on the left, oldest to the right, the tracker shows the last 10 gaps for each number, LH = spins since Last Hit and AVG = average spins between hits, fairly useful info and fun to watch.

So armed with that info I have botted a method to wait for a number to sleep > 100 spins, put it in a "watch these" array then check each consecutive spin for a match, so if a number is being "watched", as it slept for over 100 spins, and the last spun number is in that array, that number now gets moved to the betting array and gets bet on, the betting array is 9 big so the max numbers to bet on will be 9, as a new number appears to wake up the oldest drops off, simple. Bot printout also attached.

Still toying with the idea of removing a number when it rehits BUT so far the ones that slept seem to hit multiple times after waking up, not sure if 9 numbers is best either, further testing will probably help with that.

[Nickmsi]

Fascinating information . . . would never have known you could have a betting array of 9 numbers all of whom have slept for over 100 spins.

Keep us posted . . .

[superman]

I don't think "hot ever really becomes cold",

The question was does cold become hot, we know hot numbers, repeaters, dont stay hot for long enough, so what about cold numbers? when does a number become cold? and when does it stop being cold?

would never have known you could have a betting array of 9 numbers all of whom have slept for over 100 spins

It gradually builds itself up, current test I am at spin number 210 and there is 6 numbers that have slept over 100 spins so far.

[MrJ]

Everything switches....in time. It might be a short time or a long time. It makes no difference if its individual numbers or sections of the wheel. Cold becomes warm, then hot. Hot becomes warm, then cold etc. Thats why I use rotating numbers, add a number, drop one from
the list etc. Lets say I'm betting on a hot number from the last 22 spins, the #34.

I get three wins off it and then guess what happens? Its no longer hot and DROPS.......a different number (warm or hot) then takes its place. I'm NOT saying its an automatic win of three hits but the point is, the #34 dropped from the last 22.

Ken

[Gizmotron]

Nice thread. This is how you get smart about randomness.

Many years ago I wrote a program for myself that tracks all the gaps between every number's hits. After 300 spins the simulation would sort the gap values in an ascending  order. With more hits you see more gap values for each number. In this way I could start to understand the characteristics of typical hot number distributions.

You are the first person I've seen that has investigated with a simulation to discover these things. Keep up the good work. There is a reward for all this. In the long run you will find that the simplest method is the best method.

[superman]

Everything switches....in time. It might be a short time or a long time. It makes no difference if its individual numbers or sections of the wheel. Cold becomes warm, then hot. Hot becomes warm, then cold etc. that's why I use rotating numbers, add a number, drop one from the list etc. Lets say I'm betting on a hot number from the last 22 spins, the #34

Yes agreed, as I said I am testing different amounts of numbers, up to now I am working with 9 as a new one appears I drop one off as you do, still need to decide wether to remove A number after a hit or two, undecided at the moment. Don't forget Ken, this is RNG I am testing against.

You are the first person I've seen that has investigated with a simulation to discover these things. Keep up the good work

Thanks for the kind words Gizmotron

[Skakus]

Looks to me like you'll still need a progression to float this.

What if any hit that gives an overall profit dumps the whole bet list and not just the hit number?

Then you could possibly eliminate the max number of 9 bets and see how far it can go?

e.g., #6 makes the list / 42 spins later #14 makes the list / next spin #6 hits / total drawdown now -44 units / pay +36 units / down -8 units / 9 spins later #6 hits again / total drawdown -26 units / pay +36 units / profit +10 units / dump the list & start fresh tracking, etc.

A real money account test could be fun!     

[superman]

Yes Skakus it's looking like it will/may be better to remove the number from the list when it hits, as sometimes they go back to sleep. Will keep plugging away at it.

[superman]

Thanks ego but we are talking about RNG fella, there is no bias in RNG but some numbers stay cold and some numbers stay hot, in my opinion it is due to the seed value generated at that specific time/date/whatever, so the hot ones must eventually cool off and the cold ones must eventually warm up, I am trying to find the time scale for this happening.

As you are talking real wheel I will remove your posts to avoid confusion for other readers.

[Gizmotron]

Superman - " in my opinion it is due to the seed value generated at that specific time/date/whatever,"

That computer algorithm has an accurate simulation of accuracy out to 500,000 spins per each setting of the random seed. The hot / cold number characteristic is common in an RNG that does not manipulate the results. Your computer at home is an example of an RNG that does not manipulate the results.

[Gizmotron]

There are many that claim that a computer's RNG is completely different than a real wheel. I say that the average values for characteristics are the same. This is simple to prove. Someone needs to compare their computer's RNG to real spins to find the average values for ECs, Doz/Col, Streets, streaks, singles, doubles, triples, etc...

[superman]

@ Gizmotron, yes, and as far as we know BV does not manipulate the results, today I stopped letting the missed numbers build up and keep rebetting them, I have now let them build up and use them for 1 win only, so far the largest bet was 16 units on 4 numbers, far from scary.

Will continue testing.

EDIT:

Someone needs to compare their computer's RNG to real spins

In the past I have used php random, c++ random, RXtreme random, local random (excel)  and found no real noticable differences

[Gizmotron]

It's  an interesting topic, at least to me it is. Here is what I've noticed. All hot numbers in 300 spins have a longest gap between hits. This number can be above 60 spins and yet this is still the temporary hottest number for the time being. After this larger gap this same number fires up with small gaps.

[Gizmotron]

" and found no real noticable differences..."

Most modern languages use the same algorithm in their base library. The casinos have moved up to this same algorithm  for security reasons. ...But a casino can still cheat the results with other algorithms designed to interrupt  the results.

[superman]

It's  an interesting topic, at least to me it is

It is to me too, and I am sure a few others, well those who have been doing this for a good few years anyway, it has nothing to do with table or wheel layout, RNG is just a selection between 1 & 36, BV NZ or 0 & 36 for all others, it has nothing to do with roulette.

After this larger gap this same number fires up with small gaps

Exactly what I am looking at, in the last session I have just finished, number 36 was sleeping then was beton for 100 spins as it slept again after appearing once only, at that point it was back in the watch array so when it did hit it was still being bet on, hope that makes sense, anyhow, after about 400 spins number 8 did exactly the same! so the gaps would have looked something like

127,4,165