Showing posts with label Simulations. Show all posts
Showing posts with label Simulations. Show all posts

Saturday, February 12, 2011

Is Semyon Varlamov Elite?

Well, let's take a look. I ran 100,000 trials of Varlamov's career to date, assuming a .915 talent total and .920 at even strength (approximately NHL average).

Varlamov thus far has been .930 at ES and .918 total (he had a brutal PK sv% last season, about 80%, only better than Vesa Toskala).

Results:

          0.904    0.909    0.914    0.919     0.924     0.929     0.934    0.939    0.944    1.000
ES      2472    6380     14007    22662    24754    17897    8691    2629     451       57
Total   1969    5982     16916   24357     25097    18152    5889    1485     143       10

The numbers denote the number of times I got a save percentage less than or equal to the one heading the column, but greater than the one heading the previous column (so first column is save percentage under .904, second is .904 to .909, etc).


So there's still roughly a 11% chance Varlamov is simply average and overperforming at even strength. That's not all that big. It's tougher to say where Varlamov lies overall.

Looks to me like George McPhee has a 22-year old stud on his hands.

Code is after the jump. I think it's fairly straightforward should you ever want to edit it. If you aren't familiar with Java and want standard deviation and mean coded in, shoot me an email and I'll send you a revised program.

Thursday, February 10, 2011

Caps Shutout Sim

Reposted from a comment I made over at Japers' Rink.


Two notes:

One, the link points here.

Just a thought that occurred to me: I might have miscalculated the confidence interval, although at first look it looks alright (every team finished within that interval last season).

I copy-pasted the code I used after the jump. Results may vary slightly from what I got.