[alThor]

[ QUOTE ]
[ QUOTE ]
I ran a sim with n=10k

If you simmed it, I can live with our difference.

[/ QUOTE ]

Unless his sim was on a much smaller sample, the difference is very unlikely due to chance. 10k should be more than sufficient here to get your answers closer than that.

The SE for your sime is about:

(.79 * (1 - .79)) / 100 = 0.001659, 1.6%

So you should be within 3% of the true answer, probably closer.

gm

[/ QUOTE ]

??
You're missing a square root, so I stand by my (rough) margin of error Also, I was within 2%, which is borderline close if Micky were simming also, with similar margin of error. However, he was doing exact, so I guess there was an error in my sim. Unfortunately I don't have my sim with me, so I won't get around to debugging it.

I will withdraw my estimates, and go with Mickey's numbers.

alThor