Re: Basic probability question concerning E(X) and number of rolls of die
- From: "\"Luis A. Afonso\"" <licas_@xxxxxxxxxxx>
- Date: Tue, 29 Nov 2005 13:01:04 EST
The 68% interval [0.00109, 0.00213] does not contain 0.00217 but it does 0.00202. (The last two runs of 4m2 I performed).
However if we consider all the 3 values: 0.00212, 0.00217, 0.00202 which mean is 0.00210, and 3*400´000 = 1´200´000 trials leading theoretically to 1200000*0.00206 = 2472 successes and 1/sqrt(2472) = 49.7 we find the relative std. error equal to 2.0%. Being the bias +0.0002, +0.0007, -0.0004, the greater one in percent is 100*0.0007/0.00212 = 3.3%. These values (2.0% and 3.3%) are sufficiently close to safely infer that the problem is correctly solved.
_______________licas (Luis A. Afonso)
.
- References:
- Prev by Date: Re: Permutations and ONE source sample replications
- Next by Date: playing with stocks
- Previous by thread: Re: Basic probability question concerning E(X) and number of rolls of die
- Next by thread: Questions on the F-Test *VS* the Chi-Square Test
- Index(es):
