So the other day I must have been possessed by a boomer, sorry about that. Continuing where I left off:
Quote:
The flops with 3 possible are {2x6 to 9xK} & Axx & [the 6 subtracted from above], so...
so 4(8*3 + 2*6 + 6) = 168
The answer will be [24*P(success | 1 sf flop) + 64*P(success | 2 sf flop) + 168*P(success | 3 sf flop)] / C(52,3)
If we're 9-handed:
Let p = [7*C(39,2) + C(7,2)*39 + C(7,3)]/C(47,4)
Let q = [6*C(39,2) + C(6,2)*39 + C(6,3)]/C(46,4)
Let r = [5*C(39,2) + C(5,2)*39 + C(5,3)]/C(45,4)
P(sf>nf | 1 possible) = 9*8 * C(4,2)/C(49,2) * p ≈ .01244
Explanation: there are 9*8 two-player permutations. The chance that the first one flopped a SF is C(4,2)/C(49,2) because there are 4 spots for the 2 SF cards out of a possible C(49,2) places those two cards can be. I multiplied that by P(2nd player NF | 1st player SF) which is p.
P(sf>nf | 2 possible) = 9*8*[2*C(4,2)*p/C(49,2) - 4*q/C(49,3)] ≈ .0244
Explanation: the factor of 2 before the C(4,2) is because there are 2 possible SF's. However, 2*C(4,2)/C(49,2) double-counts the six-card SF's. The chance of a six-card SF is 4/C(49,3). That's multiplied by a separate conditional probability "q" because in that case, 3 cards are known.
P(sf>nf | 3 possible) = 9*8*[3*C(4,2)*p/C(49,2) - 2*4*q/C(49,3)] - C(9,3)*C(4,2)Čr/C(49,4) ≈ .036
Explanation: 3*C(4,2)/C(49,2) double-counts the 6-card SF's, of which there are 2, so we subtract them. After that, we don't need to add the 7-card SF because it's still counted (now only once). Straight formations aren't the same as combinations in that not all the sets intersect, which is why inclusion-exclusion behaves differently with them. Anyway, when 3 SF's are possible, it becomes possible for two players to flop a SF, and this possibility was double-counted and therefore gets subtracted. A third conditional probability "r" is necessary for that scenario. There are 3*C(9,3) ways to pick a player to have a NF and two to have SF's; the chance of the 4 SF cards being sorted correctly is 1/3, so I omit the 3 and 1/3.
answer ≈ 1 in 2797
Most likely you'd get a very similar result using only "p" throughout, but my OCD wouldn't let me submit an answer whose nth decimal place is wrong
Next time I'll tweak a few #'s to get the 6-max answer.
I'll also investigate whether another method (such as permutations) is more efficient. What I did above already hides some of the inner workings by calculating the players separately, which I think might negate the advantage of perms, but we'll see.