Recent content by nccr

hmm...94Mb/77Mb http://www.speedtest.net/result/400197703.png

Umm... am I the only one that caught the massive amounts of sarcasm in this response video? IMO the video is clearly real and this is Matt ****ing with us. I love it.

Same here.

I don't think anyone has mentioned this yet, but it is a lot more feasible if you don't populate all your ram slots (like 4x2gb). The reason I commonly hear is that it stresses your NB a lot more. My overclock was wrecked (3.8+ to 3.6 on an E6850) when I did that.

Yeah, I'm a coterm. I generally try to go before late afternoon when it starts getting really crowded.

Good old Arrillaga. What year are you?

Try Cos instead of cos.

Nah, I used R. I didn't save the code, but it only took a few minutes to write. I just sampled with replacement multiple times from {-1,1}, took the cumulative sum, and counted a failure whenever it dropped below 1.

Ah, I misunderstood the problem as first. I'll think about this. In the mean time, I ran some simulations (10000 trials of n flips) n=3, p=.2512 vs .25 actual n=4, p=.1867 vs 0.1875 actual n=5, p=.1881 n=6, p=.1604 n=7, p=.1561 ... n=10, p=.1249 ... n=20, p=.0915 ... n=100, p=.0403

The only problem here is that you assumed that p = 1/2. Anyways, if n is even, the probability that the number of heads = the number of tails is given by nC(n/2) * 1/2^n.

This is just the sum from i = m to n (where m = n/2 + 1 if n is even and (n+1)/2 if n is odd) of nCm * p^m * (1-p)^(n-m)

I left it on for the first few months I used Vista. It never triggered as a safety mechanism during that time, only as an annoyance. I got very fed up with forgetting to run R as an administrator all the time and having to save my workspace, run R as admin to be able to install a package...

R http://cran.r-project.org/

Strange. I'm running Vista 64 and a 4850. I used 8.6 with CCC (didn't use hotfix) and played the entire way through CoD4.

I was able to run at 800MHz 4-4-4-12 1T with 4x1GB, but only 800MHz 5-5-5-15 2T with 2x1GB and 2x2GB.