BEST TEAM ANALYZER PLEASE LOOK!

[Sean.Conway]

This has got to be the best one I've seen, it is very accurate based off of my scout team and is easy to use and has all information!
2006 Team Analyzer Current Data (http://www.chiefdelphi.com/media/papers/1830)

[Jeff Rodriguez]

http://www.chiefdelphi.com/forums/showthread.php?t=46813

[Lil' Lavery]

Definately one of the best. My only qualm is that the rating is based solely on qualification round data. If I were a scout this year I'd use it and this (http://www.chiefdelphi.com/media/papers/1820) to gather data. The problem with sw293's is that it is missing data from GTR, SBPLI, BAE, or UTC. It does it per event though, so you can track whether a team has an upwards or a downwards trend through the season.
Also, be wary of using pure statistical data, but rather use it as a single guage, along with pictures/video/first hand sight of a team's robot and how it performs. You can often find certain weaknesses and strengths to use to your advantage from actually watching machines, that you can't get from numbers.

[Sean.Conway]

Definately one of the best. My only qualm is that the rating is based solely on qualification round data. If I were a scout this year I'd use it and this (http://www.chiefdelphi.com/media/papers/1820) to gather data. The problem with sw293's is that it is missing data from GTR, SBPLI, BAE, or UTC. It does it per event though, so you can track whether a team has an upwards or a downwards trend through the season.
Also, be wary of using pure statistical data, but rather use it as a single guage, along with pictures/video/first hand sight of a team's robot and how it performs. You can often find certain weaknesses and strengths to use to your advantage from actually watching machines, that you can't get from numbers.


Its basically the same as Offensive2006 when you really look at it, and true, this is jsut some calculations, (the best, calculations I've seen) but it is no substitute for seeing it in person or ect., just data right?

[Lil' Lavery]

Its basically the same as Offensive2006 when you really look at it, and true, this is jsut some calculations, (the best, calculations I've seen) but it is no substitute for seeing it in person or ect., just data right?
The difference between it and Offensive2006 is that Offensive 2006 factors in the elimination rounds and breaks it up between compeititions. Each of those two differences have their pros and cons, thus why I'd use both of them together.

[OZ_341]

An awesome program for sure. A nice addition would be finding out who the teams partnered with in the playoffs.

In 2003 we did our homework and picked two Regional Winning Partners as our alliance partners. (236 & 175)

Having two partners that had already worked togehter and won, helped us to take Galileo in 2003.

Just a thought, but a great program just the same!!

Thanks for doing this!

[sw293]

Offensive Power Ratings have been updated for the SBPLI, BAE and GTR regionals. They do not include elimination rounds; elimination rounds foul up the "random" sample of the qualifying schedule.

Both of these rating systems are based on linear algebra. In both cases a matrix M (based on the schedule of the event) maps a vector v of average scores to the vector of ratings Mv. The difference is how the matrix M is defined.

I am partial to OPRs because you do not lose any information with them. Given the OPRs and the schedule from a regional, you can in fact figure out the average scores for each team in the regional :yikes:. Not so for scaled scores. With scaled scores, if you increase the scores in every match by one point, the scaled scores won't change. With OPRs, if you increase the scores in every match by one point, each team's offensive power rating will increase by 1/3 of a point. :D

[Bongle]

Something that would be cool in any of these analyzers is defensive power rankings. 1281, for instance, isn't very powerful at all on offense (we've maybe scored 30 points in total over 30 games), but we're very mobile and strong at defense. It would be neat if you could perhaps do another pass of your data to make defensive power rankings. I'm not entirely sure how you'd do it, but you might look at the average score of a given team and see if it is depressed by the presence of an opposing robot.