Week 5 OPRs are out with the exception of data from Hawaii from Week 4 and Mount Olive which will finish today. The World ranking has been recalculated. I will post an update once Mount Olive is done. The spreadsheet is at http://www.chiefdelphi.com/media/papers/2174
Brian seems to have factored Mount Olive into (preliminary?) MAR rankings. [strike]Although the FIRST website doesn’t have Saturday match results or awards information*, I guess from Brian’s rankings that the event actually finished on schedule yesterday.[/strike]
EDIT – Doh! Now I see what The Lucas was doing (numerical prognostication), and finally caught on to the unusual Sat-Sun schedule for Mount Olive. Ed is right. We need to wait and see how the REAL matches play out today. :o
*as of this writing
Mount Olive is not done yet. They finished 66 out of 84 scheduled matches yesterday. They are playing right now. Brian included the points that teams got on Saturday to see who has clinched a spot for the MAR Championship. I don’t understand where he got the number 56 from. It seems very high. That happens to be the number for Michigan before this weekend’s district competition to clinch a spot for MI State Championship. I didn’t do the calculations for MAR but I thought the clinching number should be lower based on the number of spots and number of teams there.
Ed,
I have noticed some major exceptions this year in OPR predictions for the playoffs. Orlando, Oregon, Traverse, Montreal, and several other regional OPR predictions are not correlating with the actual winners as in the past. Statistics for OPR are off. Do you see any other predictors that improve the results?
Ether, thank you for your effort to reconstruct the Hawaii regional standings data. You are doing the FIRST community a big favor. Please make sure you add up the cumulative HP, BP, TP and CP and put it in the same format as how FIRST report it. Then I can run the results through my spreadsheet. Let me know when it is available. Thanks.
P.S. It was nice meeting you in person last week. Sorry we didn’t get a chance to chat longer as we were both busy at the event.
I did not get a chance to analyse the data extensively this year. Overall I think OPR still represent the strength of each team fairly well. Keep in mind that the combined OPR to predict outcome has only been around 70-75% in the past few years. In most cases this year, it was still correct most of the time. The examples you gave were exceptions. Perhaps it was because of triple balance at the end since those extra 20 points were not available during qualifying rounds and therefore not represented in the OPR numbers.
When I am in St. Louis and get bored, I will look into it in more details. Will I see you there this year?
Ed,
I agree with the 70-75% prediction of the combined OPRs during the qualification rounds. However usually the playoffs are above 85% in prediction correlation. I looked at 59 regionals and the playoff correlation is 69.7%.
Thanks for your efforts this year to make the scouting data manageable,
Team 86 is finalizing plans to be in St Louis this year. I hope to meet you there.
Okay, you will have to download it again. All regionals and districts up to Week 5 are now included. Everybody please thank Ether for painstakingly recreating the Hawaii standings and put it in the same format to make my life easier. I also included the Mount Olive district results and recalculated the World Rankings.