Let's say I've got a pool of 20 numbers, and each event chooses a number randomly. I'm trying to find the 50% point for one of these three:
- 50% chance that by this event, at least 1 duplicate number is pulled.
- 50% chance that by this event, at least 2 duplicate numbers are pulled.
- 50% chance that by this event, at least 1 triplicate number is pulled.
The closest I've come up to finding out a solution to this is using the Birthday problem for #1, wherein I get a range between 5 and 6. I feel like that's the direction I need to go to figure out the number for 2 and 3, but I can't seem to crack it.
Any advise or help is appreciated.
To be precise I assume that 3-cates (or 4-cates ...) do not fulfill the requirement of duplicates, and so on, which means for the 0.5 probab. margin we only count duplicates. Drawing a quadriplate, for example, does not count as 2 duplicates(!).
After 5 draws the prob. of 1 duplicate is 2907/8000 =0.363375 and the prob. of 2 duplicates is 513/16000 = 0.0320625, total below 0.5. (There are chances for 3-cates up to 5-cates.. ignored here). After 6 draws the prop. of 1 duplicate is 8721/20000 =0.43605, for 2 duplicates is 26163/320000 = 0.081759375, for 3 duplicates is 513/320000 =0.001603125, altogether beyond the 0.5 treshold. So the answer to 1 is between 5 to 6 draws as already stated.
In a short cut notation we can split the 20 into a sequence of i^f(i) which means an i-cate occurs f(i) times. E.g. 0^17 1^2 5^1 means 17 numbers have not been drawn, 2 numbers have been drawn once, and one number has been drawn 5 times. An incomplete table (which means entries containing 2^1 and 2^0 are dropped) after 7 draws with probabilites shown as fractions and floating point numbers:
with sum below 0.5. After 9 draws
which sums to 0.375.. below 0.5. After 10 draws
which sums to 0.484215, below 0.5 After 11 draws:
which sums to 0.5823... more than 0.5 So the anwer to 2. is that the transition is from 10 to 11 draws.
3. After 9 draws (incomplete table meaning entries containing 3^0 are dropped)
below 0.5. After 14 draws
which sums to 0.3807, not enough. After 15 draws
which sums to 0.50648, sufficient. So the answer to 3 is between 14 and 15.