OPR results from Week 2 is out. You can download the data from
http://www.chiefdelphi.com/media/papers/2174
If there is any mistake, please let me know.
I need to make some corrections on my recent post on 2/21. The reported score is not the seeding score as I thought before week 1. It is the actual number of goals minus penalties.
I am a little surprised how good the correlation is of OPR to results prediction. At most regionals/districts, the prediction is 70-80% accurate. It also correlates very well with the actual alliance selection. Look at the data and you can be the judge.
Also the correlations are better for Week 2 than Week 1. It probably has to do with the additional 5 seeding points for a win.