There are total of 13 players in a game, and 4 of them are assigned "evil" randomly. Some players have noted, that some players have been evil multiple times in a row (specifically, a player was evil 5 times in a row over 5 games). I am wondering about this specific probability, that is, what is the chance that we would observe this behavior, for any player (not just this particular player).
In addition to that, I want to know the general probability for this problem: M players, K assigned "evil", and over N games any 1 or more player being assigned "evil" N times in a row.
Bonus question, a curiosity: What about if there weren't 5 games, but L games in total, and a streak of N or more?
The best thing I can solve for now is simply the probability that a specific player will be "evil" 5 times in a row:
$(\frac{4}{13})^5 \approx 0.2757%$
I tried some things with $\frac{\binom{12}{3}^5}{\binom{13}{4}^5}$, but that simply yields the result for a single specific person, and doesn't take into consideration the fact that multiple people could overlap too
I attempted to code up a simulation which would simply check this (link to jsfiddle, freezes browser for a bit). My result seems to be around 3.55342%, which seems oddly close to $(\frac{4}{13})^5 \times 13= 3.58530%$, which seems surprising.
I also tried enumerating each possibility, however, it seems like this method wouldn't be fruitful, as the number of total possibilities is around $\binom{13}{4}^5 = 186865965446875$ (Or $\binom{M}{K}^L$), which is something my computer couldn't be able to handle. Possibly there are ways of optimizing this though...