There is a riddle on YouTube: Can You Solve The Missing Sock Puzzle?
The riddle goes like this:
A statistician keeps a simple wardrobe. He only purchases pairs of black socks and white socks, and he keeps all of the socks in a pile in the drawer. Recently one of the socks was lost in the laundry.
The socks have a mathematical property. If you select two socks at random from the drawer, the socks will match in color exactly $50\%$ of the time.
The statistician owns more than $200$ socks but less than $250$ socks, and there are more black socks than white socks.
How many socks of each color are there?
And, which color sock was lost in the laundry?
The solution in the video uses the "$50\%$ matching mathematical property" as a "fact" to conclude that the total number of socks must be a "perfect square", and since the only perfect square between $200$ and $250$ is $15^2$, the total number of socks is therefore $225$. And then further concludes that there must be $120$ black socks ($60$ pairs) and $105$ white socks ($52$ pairs plus $1$ unmatched white sock).
Now, this just seems wrong to me. I understand that probability is not (always) intuitive, but, I feel (intuitively) that there is something wrong (invalid) with the initial setup of the question.
Specifically, I have a problem justifying the "mathematical property" of the socks where "if you select two socks at random from the drawer, the socks will match in color exactly $50%$ of the time".
I know from the stated conditions, that there are more black socks than white socks, so surely, the probability can't be $50\%$. And merely stating it is $50\%$ doesn't make it so.
An explanation given in the comments for the video is that:
Probability of picking a black pair: $\frac{120}{225} \cdot \frac{119}{224} = 28.33\%$
Probability of picking a white pair: $\frac{105}{225} \cdot \frac{104}{224} = 21.67\%$
$28.33\% + 21.67\% = 50\%$
This argument seems compelling, but I think it's flawed.
I keep looking at it with these two examples:
If you start with $220$ black socks ($110$ pairs), and $5$ white socks ($2$ pairs plus $1$ unmatched sock), the chance of picking a pair is probably $95\%$ or better. And, if you start with $224$ black socks ($112$ pairs), and $1$ (unmatched) white sock, the chance of picking a pair is better than $99\%$.
Then:
- Start with $104$ black socks ($52$ pairs) and $104$ white socks ($52$ pairs).
- The probability of picking a pair of socks is $50\%$.
- You add an additional pair of black socks, and the probability of picking a pair of socks increases (to slightly more than $50\%$).
- As you add additional pairs of black socks, the probability of picking a pair increases more and more.
- Then, according to this "riddle", when you reach $120$ black socks ($60$ pairs), and you add an additional (single) white sock, the probability of picking a pair of socks suddenly drops back down to $50\%$.
What's actually happening here?