Which other teams is your team most similar to?

[SoftwareBug2.0]

There are lots of papers ranking teams based on their on-field performance, but teams could also be compared based on the kind of awards that they win. So for every team I've made a ranking of the top 10 teams that do similarly at winning awards.

The format is as follows:

team_number[most_similar_team,2nd_most_similar,...]

Technical details:
-Data is taken from The Blue Alliance; fictitious teams exist if TBA says they exist.
-Similarity is calculated as follows: Sum of the squared differences between between each team's tendency to win an award type.
Where

award_tendency = ((# of times won) - (expected # of times won)) / (events attended)

where

"expected # of times won" = sum over all events attened( # of times award given out / # of teams attending )

Note that time doesn't come into this formula at all - winning something in 1992 is counted the same as winning it this year. Also note that this counts winning only "Engineering Excellence" and winning only "Innovation in Controls" as more different than winning only one of them and winning nothing.

Also, a list of least similar teams is not very interesting because it's almost always going to be a team that has only competed for one year and won a bunch of awards.

[Max Boord]

A google sheets version.

[nardavin]

Originally Posted by Max Boord

A google sheets version.

It appears that the parser is broken (not sure how, haven't looked too much at the code). The raw data is correct, but for some reason it mangles the team numbers when parsing it.

[Max Boord]

Originally Posted by nardavin

It appears that the parser is broken

Fixed.

[MARS_James]

So does this count all awards? Or just judged teams award?

Basically does this count regional winner/finalist? if so is that different than district event winner/finalist?

Does this count Woodie Flowers Finalist and Dean's List Finalist?

Also are Championship awards different than regional/district?

[SoftwareBug2.0]

Originally Posted by MARS_James

So does this count all awards? Or just judged teams award?

Basically does this count regional winner/finalist? if so is that different than district event winner/finalist?

Does this count Woodie Flowers Finalist and Dean's List Finalist?

Also are Championship awards different than regional/district?

Yes, it counts all awards that are awarded to teams including regional winner/finalist. It does not differentiate between regional winner and district event winner or between regional finalist and district event finalist. However, the formula does take into account event size, so while a 35-team district event is counted the same as a 35-team regional it is counted differently than a 60-team regional.

Woodie Flowers Finalist and Dean's List Finalist count if there are team numbers noted.

Championship awards are not counted specially. Each of the divisions and Einstein are counted as a their own events.

I am using exactly what The Blue Alliance calls "award_type", the complete list of which is documented here: https://github.com/the-blue-alliance...rd_type.py#L15

I've actually been thinking about seeing what I get if I reduce the number of categories. For example, maybe just four categories like so:
1) Culture: Chairman's=1; Engineering Inspiration=.5; Rookie All Star=.5
2) Winning: "Winner"=1; "Finalist"=.5; "High Rookie Seed"=.5;
3) Technical: Industrial Design=1; Innovation in Control=1; etc.
4) Other, 1 each

[Bryce2471]

Originally Posted by SoftwareBug2.0

I've actually been thinking about seeing what I get if I reduce the number of categories. For example, maybe just four categories like so:
1) Culture: Chairman's=1; Engineering Inspiration=.5; Rookie All Star=.5
2) Winning: "Winner"=1; "Finalist"=.5; "High Rookie Seed"=.5;
3) Technical: Industrial Design=1; Innovation in Control=1; etc.
4) Other, 1 each

I'd be interested in seeing this version of the list.

[SoftwareBug2.0]

Originally Posted by Bryce2471

I'd be interested in seeing this version of the list.

Your wish is my command.