Best in your State/Region

[A Mountain Man]

Hello again, all! I've missed you. Let's take a little detour.

I was wondering this same question to myself right after the championships: "What mountain teams were the best this year, and in years past?" However, I got carried away and I took my question further, instead asking "What team would be the best if the entire Rocky Mountain Region was in the district format during 2016?" So, I calculated by hand the number of district points each team would have received if the district system were present in the region this year. (I'm a mountain man. I don't know how to make software do things for me.) I have organized the list by the tiebreaker rules, and have come up with a very interesting result. You can find that document here.

A few important things about the document:

• All the teams with a green background would have qualified for Championships, as they are the top 20%. This, of course, does not factor in the recipients of the Rookie All Star, Engineering Inspiration, or Chairman's Award at a theoretical district championship, nor does it estimate what teams would have qualified for a district championship. This list also does not factor in the estimated number of points from a district championship, as one did not happen.
• I have finally decided on a geographic area to monitor, and it does/will not include the PNW region or British Columbia. That is why they are not present.
• I am not perfect. This list may be incorrect in some of the ordering.

I hope you all enjoy this interesting thought experiment. I'll be back soon with the answers to the main questions and more data! Stay tuned.

[Billfred]

Originally Posted by Kevin Leonard

Summer CD is boring, and these threads are less boring.

Who are the top 5 teams in your state/region:

• this year by OPR?
• based on this year's results at various levels?
• this year overall?
• during the past 5 years by OPR?
• during the past 5 years by results at various levels?
• during the past 5 years overall?
• in terms of culture(both culture awards and your experience with the teams)?

I'll comment in a little bit with some data on NY, unless someone else wants to do it.

I'm basing the South Carolina OPR data off of Palmetto; 1293 and 1319 weren't there, but neither got above .500 at any event during the season.

OPR:
1) 4451
2) 1876
3) 1102
4) 3490
5) 343

Results at various levels/overall:
1) 4451 (won Palmetto, Palmetto EI, Orlando Innovation in Control and QF, Carver Imagery Award and QF)
2) 3490 (won Rocket City with Robonauts and Bomb Squad--but hey, they did enable 100% capture along the way)
3) 343 (Palmetto finalists, Rocket City QFs)
4) 1876 (Palmetto QFs and 8 seed, Orlando semis and 7 seed)
5) 1758 (Palmetto #3 seed and QFs)

Truth be told, that list is pretty close for all the other ones too. 343 has a Championship subdivision finalist in 2015 that would push them up a bit more (fourth robot, never played, but scoreboard), but they were also off the pace for a few years going back. 4451's been on a hot streak where they've been head and shoulders above everyone else in the state--one of three teams to win Palmetto back to back, WFFA, EI, RAS literally everywhere they went--and with them helping to start a team the next county over I'm on the lookout for them. 3490 is always a threat in the state, especially at SCRIW, but just now broke through by winning the last-pick lottery (which I have absolutely zero room to hate on). 1876 never gets any press or buzz, but somehow they pull a rabbit out of their hats and gets in contention even at overstuffed events like Palmetto and Orlando.

[asid61]

Originally Posted by MichaelBick

973, 3476, and 330 are all definitely in the mix too for CA.

Very true. I have a very North-specific worldview, being from lower norcal myself.

[Kevin Leonard]

Originally Posted by Jay O'Donnell

Nitpicking but 229 should be 5th on the championship picks list since we were 8th overall pick on Newton.

Not saying we should've been, but we were.

You're correct. I recalled that 229 got picked, but I assumed it was much later in the draft. Updated as such.

[Brian Maher]

I'll post my analysis on MAR and New York sometime in the next day. In the meantime, I have one question about a response.

Originally Posted by GKrotkov

For teleop low goals:
1) 25, t = 7.6629 (they destroyed)
2) 1257, t = 4.6518
3) 708, t = 4.5852
4) 5113, t = 4.2006
5) 1923, t = 3.5787

For overall boulder volume:
1) 25, t = 6.3270
2) 708, t = 3.2628
3) 5895, t = 3.2018
4) 3314, t = 2.8022
5) 2590, t = 2.7175

How come 1257 is #2 on the low goal list, ahead of 708, but not even top 5 for boulder volume when 708 didn't score high? (or at least, not that I saw) Not that I take issue with it, data is data, but I'm genuinely curious how this result is possible.

Your analysis is pretty spot on. T/Z-scores are a nifty stat for this. I look forward to seeing how it compares to my own analysis.

[GKrotkov]

Originally Posted by BMSOTM

How come 1257 is #2 on the low goal list, ahead of 708, but not even top 5 for boulder volume when 708 didn't score high? (or at least, not that I saw) Not that I take issue with it, data is data, but I'm genuinely curious how this result is possible.

Thanks for asking! Yeah, you can also see a similar effect with 5895 coming in ahead of 3314 in overall boulder volume but behind in teleop high goals.

When I was working on it initially, I realized that consistency was overvalued in the standard error statistic*, and a team that was consistently slightly better than the average team came out far ahead of everyone else, even those that had small inconsistencies but were generally better (significantly higher average.) This is because using the t-distribution isn't precisely telling us how good a team is, but rather how unlikely their performance is given that we assume that they are the average team.

The way I solved this is by restricting analyses of individual fields to a select set of teams rather than all teams at the competition, which raised the average comparatively and reduced the overvaluing of absurdly consistent teams. I kept the ones with all teams analyzed, but honestly reality-checking the latter made me realize that restricting the number of teams for more specific fields could be helpful. For example, I did not include 1712's data in the high goal t-score calculations. This caused averages to change and thus some of the strange cardinal results you see in the final order sort. So, in the example of 708 and 1257: 708 has a higher average and higher standard error than 1257, so, with the low goal specific analysis the higher general average resulting from eliminating teams that aren't competitive low goalers makes standard error more important and thus 1257 does better relative to 708 in the low-goal specific one rather than the boulder volume one.

* at least, overvalued for my purposes in picklisting.

[Karibou]

Originally Posted by GKrotkov

The way I solved this is by restricting analyses of individual fields to a select set of teams rather than all teams at the competition, which raised the average comparatively and reduced the overvaluing of absurdly consistent teams. I kept the ones with all teams analyzed, but honestly reality-checking the latter made me realize that restricting the number of teams for more specific fields could be helpful. For example, I did not include 1712's data in the high goal t-score calculations. This caused averages to change and thus some of the strange cardinal results you see in the final order sort. So, in the example of 708 and 1257: 708 has a higher average and higher standard error than 1257, so, with the low goal specific analysis the higher general average resulting from eliminating teams that aren't competitive low goalers makes standard error more important and thus 1257 does better relative to 708 in the low-goal specific one rather than the boulder volume one.

* at least, overvalued for my purposes in picklisting.

How did you determine the cutoffs for which teams to include for each analysis? I imagine it was pretty clear-cut for high goal scoring since that was an "either you can do it or you can't" ability for the most part, but where did you draw the line for low goal scorers? I know 25 was good, but is their t-score so dominant compared to the rest of the teams because they were so good, or because there was a wider spread in low goal scoring ability, lowering the average compared to how well 25 was performing? (does that question make sense? Statistics really isn't my strong suit)

Also, is this data from just quals, just eliminations, or both?

[ezygmont708]

Originally Posted by GKrotkov

Thanks for asking! Yeah, you can also see a similar effect with 5895 coming in ahead of 3314 in overall boulder volume but behind in teleop high goals.

When I was working on it initially, I realized that consistency was overvalued in the standard error statistic*, and a team that was consistently slightly better than the average team came out far ahead of everyone else, even those that had small inconsistencies but were generally better (significantly higher average.) This is because using the t-distribution isn't precisely telling us how good a team is, but rather how unlikely their performance is given that we assume that they are the average team.

The way I solved this is by restricting analyses of individual fields to a select set of teams rather than all teams at the competition, which raised the average comparatively and reduced the overvaluing of absurdly consistent teams. I kept the ones with all teams analyzed, but honestly reality-checking the latter made me realize that restricting the number of teams for more specific fields could be helpful. For example, I did not include 1712's data in the high goal t-score calculations. This caused averages to change and thus some of the strange cardinal results you see in the final order sort. So, in the example of 708 and 1257: 708 has a higher average and higher standard error than 1257, so, with the low goal specific analysis the higher general average resulting from eliminating teams that aren't competitive low goalers makes standard error more important and thus 1257 does better relative to 708 in the low-goal specific one rather than the boulder volume one.

* at least, overvalued for my purposes in picklisting.

Gabe,

Do you have stats for climbs?

Thanks,

Z

[Kevin Leonard]

One more list I played with- Sorted by Average OPR Rank during the past 5 years while removing each team's worst year:

1. 5254* [1]
2. 1507 [4.75]
3. 3015 [7.75]
4. 340 [8]
5. 20 [9]

*Now 5254 only has two years to calculate from, but whatever.

[Chris is me]

Originally Posted by Kevin Leonard

One more list I played with- Sorted by Average OPR Rank during the past 5 years while removing each team's worst year:

1. 5254* [1]
2. 1507 [4.75]
3. 3015 [7.75]
4. 340 [8]
5. 20 [9]

*Now 5254 only has two years to calculate from, but whatever.

Okay dude now you're just purposefully creating ranking algorithms that make 5254 look good

[GKrotkov]

Originally Posted by Karibou

How did you determine the cutoffs for which teams to include for each analysis? I imagine it was pretty clear-cut for high goal scoring since that was an "either you can do it or you can't" ability for the most part, but where did you draw the line for low goal scorers? I know 25 was good, but is their t-score so dominant compared to the rest of the teams because they were so good, or because there was a wider spread in low goal scoring ability, lowering the average compared to how well 25 was performing? (does that question make sense? Statistics really isn't my strong suit)

Also, is this data from just quals, just eliminations, or both?

I threw out any team that had an average low goal score of <1. I don't have any insightful reason for that, but I think that it's fair to assume that a competitive low goaler at MAR Champs will score one boulder per match, on average. There weren't any of these, but if I found a team with a standard error of 0, then I'd have to throw them out, too. Not for any great reason, but just because the formula for t-scores divides by the standard error.

The question about 25 makes perfect sense - and it's a really good one, too. It has a multifaceted answer. For one, the t-distribution flattens out near the extremes. That means that you have to increase relatively more t-score to gain a similar amount of area under the curve. That is, t-scores don't scale linearly. A team with a t-score of 4 isn't twice as good (or even twice as unlikely) as a team with a t-score of 2. As for the spread of teams, eliminating teams with <1 low goal average really tightened the spread, rather than widening it. I haven't tried to prove it, but I imagine that this could help 25 by reducing the margins between everyone else's average and the population average. I do think that even with those mitigating factors, 25's margin over everyone else is still remarkable.

This is from qualifications only. Dawgma reduced our scouting to a watchlist after we got 9 matches for each team, but I've filled out some of the scouting via recordings since then.

Originally Posted by ezygmont708

Do you have stats for climbs?

I have the data that Dawgma & 708 collected from MAR Champs, but it'd be kind of pointless to use the t-distribution on scales, since you won't do more than one per match. Probably just as good to look at the ratio of successful scales to attempted scales. That gives us:
1/2/3) 708 [7/7], 341 [5/5], and 869 [6/6] tie with a perfect record.
4) 25 with 8/9
5) 365 with 6/7

Major caveat there, though. I didn't do the nonboulder scouting to fill out the scales, so we only have a limited # of matches to get that data from. Also, since 25 was on our watchlist we watched them more, more chance for us to catch them in a bad match. If someone else has different data, I'd go with that.

[Brian Maher]

Originally Posted by ezygmont708

Gabe,

Do you have stats for climbs?

Thanks,

Z

Originally Posted by GKrotkov

If someone else has different data, I'd go with that.

Here's what 1257's data says on scale rate (9-12 matches observed for all teams):

1. 5401 (100.0%)
2. 25 (78.5%)
3. 4573 (70.0%)
4. 708 (66.7%)
5. 341 (64.3%)

Here's the top five for average endgame points (scales + challenges):

1. 5401 (15.0)
2. 25 (12.9)
3. 869 (11.4)
4. 341 (11.1)
5. 4573 (11.0)

If anyone else has stats requests from MAR CMP, feel free to ask and I'll see what I can do.

[Gregor]

Originally Posted by Kevin Leonard

One more list I played with- Sorted by Average OPR Rank during the past 5 years while removing each team's worst year:

1. 5254* [1]
2. 1507 [4.75]
3. 3015 [7.75]
4. 340 [8]
5. 20 [9]

*Now 5254 only has two years to calculate from, but whatever.

Are you a meme or is this a real post.

[Whatever]

Originally Posted by jajabinx124

Here they are!

By Championship Selection Order for MN:
1) 5172, 3rd overall pick
2) 2823, 4th seed captain
3) 2052/3130 (tie), 4th overall pick
4) 2987, 8th seed captain
5) 4607, 9th overall

By Championship Rank:
1) 2823, ranked 4th
2) 2052, ranked 8th
3) 3130, ranked 9th
4) 5172, ranked 10th
5) 2987, ranked 11th

2987 is flying under your radar.
Quote is editted to add 2987

[Rangel(kf7fdb)]

Originally Posted by A Mountain Man

Hello again, all! I've missed you. Let's take a little detour.

I was wondering this same question to myself right after the championships: "What mountain teams were the best this year, and in years past?" However, I got carried away and I took my question further, instead asking "What team would be the best if the entire Rocky Mountain Region was in the district format during 2016?" So, I calculated by hand the number of district points each team would have received if the district system were present in the region this year. (I'm a mountain man. I don't know how to make software do things for me.) I have organized the list by the tiebreaker rules, and have come up with a very interesting result. You can find that document here.

A few important things about the document:

• All the teams with a green background would have qualified for Championships, as they are the top 20%. This, of course, does not factor in the recipients of the Rookie All Star, Engineering Inspiration, or Chairman's Award at a theoretical district championship, nor does it estimate what teams would have qualified for a district championship. This list also does not factor in the estimated number of points from a district championship, as one did not happen.
• I have finally decided on a geographic area to monitor, and it does/will not include the PNW region or British Columbia. That is why they are not present.
• I am not perfect. This list may be incorrect in some of the ordering.

I hope you all enjoy this interesting thought experiment. I'll be back soon with the answers to the main questions and more data! Stay tuned.

I think the list would be more accurate if you took the average of the events each team did or double the total points if a team only attended one event. As it stands now, a team that only attended one event is below where they should be.