Birthday paradox

[Priyanka]

Most of you know birthday paradox. And most of you might have dismissed its application in roulette.

Does anyone know what is the most popular birthday in europe and US? October 1st week. Why? Because hmm, christmas and new years eve. So something strange happening on new years eve which is making the odds tilt and make it better than 1/365 for days in october 1st week. So going back to the birthday paradox, does it tilt the odds of two people having the birthdays on October 1st week? I dont know the answer, but an interesting question to ask and think about. Where's Bayes. I could use his help.

Changing this to roulette, if there is a particular event out of a set of events that occurs in a higher frequency, is it highly likely that we are able to find a repeat of that event faster than expectation? Hmm.. good to think and thats what am thinking right now. Trying to find an event in a set of events that is highly likely to happen.

[falkor2k15]

I have discussed the bday paradox at length before with regards to Roulette, but could never figure out it's application.

With regards to your question we could look at the cycle length outcomes of Dozen Cycles: CL1, CL2 and CL3. I think these have been referred to as blue balls, green balls and red balls - each not equally likely. One type of balls representing CL2 always has a higher, faster chance of occurring than CL1 or CL3.

Although CL2 is "king", all CLs are comparable in terms of cost/payout. And I doubt that is the correct application of birthday paradox. It was hypothesized that one event has more chance of repeating when part of a larger group of events that repeat, but I doubt that as well. My guess is that the application is with parallel games that target the same event, hence more support.

[maestro]

tricky one...as we have in our room on constant 37 numbers

[Turner]

Its not really a paradox. Just maths.

If you go on Manchester City FC website and look at the first team, excluding loaned out players, 24 players can be seen and Aleix Garcia and Kevin De Bruyne both have the birthday 28th June 

[nottophammer]

Most of you know birthday paradox. And most of you might have dismissed its application in roulette.

Does anyone know what is the most popular birthday in europe and US? October 1st week. Why? Because hmm, christmas and new years eve. So something strange happening on new years eve which is making the odds tilt and make it better than 1/365 for days in october 1st week. So going back to the birthday paradox, does it tilt the odds of two people having the birthdays on October 1st week? I dont know the answer, but an interesting question to ask and think about. Where's Bayes. I could use his help.

Changing this to roulette, if there is a particular event out of a set of events that occurs in a higher frequency, is it highly likely that we are able to find a repeat of that event faster than expectation? Hmm.. good to think and thats what am thinking right now. Trying to find an event in a set of events that is highly likely to happen.

Your on a higher thought plain than me, but i'll tell you where to find your repeat.
In spins 1-10. if spins 1-10 don't produce, you've been unlucky. So spins 11-20 should give one of 1-10, it's a bit of a Vaddi problem. I'll leave it there 

[nottophammer]

Just now on MPR plop's on there, these 12 spins  end +45
33
35
15
19
24
27
8
26
14
9

20
24

[nottophammer]

Interblock ?

[Priyanka]

With regards to your question we could look at the cycle length outcomes of Dozen Cycles: CL1, CL2 and CL3.

Falkor the problem with them is that they are not equal events. One needs one bet where as others need more. Hence the risk also varies.  I am actually thinking there should be something which will carry the same payout but odds being different.

tricky one...as we have in our room on constant 37 numbers

I know.. I know..  indeed tricky one.  I have tried various things like cycles, stitching bets, VdW - nothing seem to be working here as none of them has been able to break the payout-odds balance. Tough nut!

Its not really a paradox. Just maths.

If you go on Manchester City FC website and look at the first team, excluding loaned out players, 24 players can be seen and Aleix Garcia and Kevin De Bruyne both have the birthday 28th June 

Hmm semantics. I believe it is called birthday problem rather than a paradox these days. Gunners has Lucas Perez and koscielny. Chelsea have victor moses and Nathaniel Chalobah. We could go on and on I guess.

What about turbos method of numbers exceeding expectations. Payout is same but do they have different odds as they hit more than the expectation?

[RouletteGhost]

a roulette event highly likely to happen

and above expectation

well we would all like to find that

[Turner]

Payout is same but do they have different odds as they hit more than the expectation?

many say  "yeah, but that doesnt change the odds so the HE will get you"

No one ever says "and here is how my method changes the odds"

[Priyanka]

many say  "yeah, but that doesnt change the odds so the HE will get you"
No one ever says "and here is how my method changes the odds"

May be because such a method doesn't exist

I am seriously looking at the following data point about repeaters. Most of the repeaters happen within the last 18 unique numbers. Defining "most" - only in approximately about more than 3500 repeaters do we have an exception and that is a large number. One side of me says that the odds of certain numbers are different from certain other numbers. Other side says that "huh! how can that be". But however, I truly believe because these repeaters happen within 18 unique numbers, "when" the repeat happened is something that might have the potential to be the method, where the payout is going to be 35:1 but the odds are going to be nearly half or a little more than 1/37. As always the issue is to find a practical working method and I think am struggling there.

[maestro]

<One side of me says that the odds of certain numbers are different from certain other numbers. Other side says that "huh! how can that be> if it was me i would listen to this side...but people are not the same..

[falkor2k15]

May be because such a method doesn't exist

I am seriously looking at the following data point about repeaters. Most of the repeaters happen within the last 18 unique numbers. Defining "most" - only in approximately about more than 3500 repeaters do we have an exception and that is a large number. One side of me says that the odds of certain numbers are different from certain other numbers. Other side says that "huh! how can that be". But however, I truly believe because these repeaters happen within 18 unique numbers, "when" the repeat happened is something that might have the potential to be the method, where the payout is going to be 35:1 but the odds are going to be nearly half or a little more than 1/37. As always the issue is to find a practical working method and I think am struggling there.

What do you mean by last 18 uniques? Most numbers repeat on spin 7-8 right? Max is about spin 25. So of course more will fall on the 7-8 side of 18 compared to the 25 side of 18.

In all my tests numbers have always been distributed evenly regardless of whether it was 4 (my Son's age) or my birthday (27 of next month), so I don't believe in superstition as the tests prove the maths to be correct.

[Priyanka]

What do you mean by last 18 uniques?

When a repeat happens, it happens on a number that is within the last 18 unique.

Most numbers repeat on spin 7-8 right?

It depends on where you start the counting from . I start counting back from the repeat to the place where the number appeared before.

[falkor2k15]

Yeah, you did show that counting from a repeat keeps the stats in order.

I still don't understand the last 18 unique part though... even if you retrack on a repeat there will seldom be 19-25 uniques compared to 1-18 before the next repeat occurs.