Average Team OPRs going in to Einstein

[bimbunky]

Some food for thought:

1: 2056 - 71.3922396241
2: 148 - 70.0640372296
3: 1678 - 61.6746235366
4: 1690 - 60.0464764582
5: 195 - 58.734126484
6: 330 - 57.3404216112
7: 987 - 56.802826136
8: 4678 - 56.7850558028
9: 2481 - 54.7325881845
10: 1986 - 51.9589946392
11: 2052 - 50.8941352961
12: 1501 - 49.789168129
13: 2122 - 49.1143966134
14: 217 - 48.4589076373
15: 3476 - 47.205222354
16: 379 - 46.7165911692
17: 3339 - 42.7481699705
18: 5050 - 41.7598876809
19: 120 - 40.5139406869
20: 41 - 40.177310552
21: 1065 - 39.646558455
22: 3015 - 38.9890248366
23: 1086 - 38.3333221328
24: 364 - 36.0320531408
25: 3538 - 35.8519716521
26: 1511 - 35.1514886306
27: 2990 - 34.4945152421
28: 694 - 33.6461602452
29: 188 - 33.1854360695
30: 1197 - 30.6396225515
31: 1405 - 28.8537837093
32: 4828 - 25.4473861927

[TheMilkman01]

Hey, thanks for doing this. It's really cool to contrast and compare each team and alliance on this list. Stats are always cool

[bimbunky]

I actually work with our team's scouting system so this is something that I really like to do

[Ether]

@bimbunky: Please define what you mean by "average" team OPR going in to Einstein.

i.e. what data are you including and how are you doing the averaging?

[bimbunky]

The average of their OPRs of all of their respective regionals up until their subdivisions

[jijiglobe]

Just wanted to take this opportunity to point out how OPR is a pretty weak indicator of team strength. Our robot was projected to rank 53rd in curie before champs, but instead, we ranked 8th and captained our alliance to Einstein. We even took a game off of the world champions.

[bimbunky]

Indeed. But a team should have a generally strong OPR throughout their regionals also because their presence should allow them to score generally higher in their matches compared to other teams'. Even though OPR is not always accurate, across regionals there must be a chance of a good team being put on an alliance with bad robots, so their OPR is not high just because they have good robots on their team in each quals match.

[christheman200]

On the contrary, a team may have had coms problems, etc. in one or more of their regionals, leading to a lower OPR. I wouldn't judge based off of averaged OPR for any other than the very top or bottom of the stats. That being said, I don't mean to bash the poster. I actually found this quite interesting.

[bimbunky]

I'm glad you found it interesting! You're correct, OPR uses a linear regression (bunch of fancy stuff) that depends on the sum of all match scores for a team so in a way comes issues and such could potentially bring it down. Over several games, however, these are generally minor inconveniences although I agree OPR in the first place is an estimate and not always an agreeable number

[Arhowk]

Originally Posted by jijiglobe

Just wanted to take this opportunity to point out how OPR is a pretty weak indicator of team strength. Our robot was projected to rank 53rd in curie before champs, but instead, we ranked 8th and captained our alliance to Einstein. We even took a game off of the world champions.

Yep. We were... 18th? in OPR coming into the competition and ended 40th and didn't even get picked. (Not to say that our crappy intake or our two matches where this whole concept of "team" was flown out the window wasnt to blame, but nonetheless it happenes here and it happens everywhere)

[angelah]

The fact that OPR doesn't correlate well with rankings doesn't necessarily make it useless. There is more than one logical conclusion to draw. I think the fact that it seems to correlate less this year than in previous years (from my random observations) is pretty interesting actually. I honestly wish I had the time and smarts to crunch the numbers more: OPR versus both rankings and alliance selection, for this year and the last two.