Data Scouting

Hello, everyone! First, I’d like to congratulate everyone on a successful season and I hope to see you all in St. Louis. I’m from team 2959, the CW Tech Robotarians, and I’d like to know what data everyone else takes during matches. Before Worlds, I’d like to refine our data collection process as much as possible in order to ensure quality data. (We use an app to take data, and we’d be more than happy to share the source code with you for that.)

When I scout I just take down truss, and Auto high or low goal shooting and the amount of experience they have like the year they were a rookie etc.

Congrats on getting to Worlds!

Our team keeps track of robot performance during each match by filling out a paper form with three main sections.

The first is a small section for autonomous where scouters fill out whether the robot started with a ball, whether they got mobility points, the goal they tried to score in, and if they made it. (In the rare case of a double ball auto, scouters specify the outcome in the notes section).

The second is a big notes section for jotting down qualitative assessments and miscellaneous observations (had awful partners this match, got pushed easily, fast driving, played killer defense, incurred a pinning foul, etc.)

The third section is a fill in the blank about what the robot did with the ball during each cycle, and how long they possessed the ball. For example:

  1. role = middleman*, goals = truss, shots = 2, score = 10, time = 10s
  2. role = inbounder/scorer**, goals = high/low, shots = 3/1***, score = 1****, time = 45s
  3. role = defense
  4. role = inbounder, time = 8s

*passed between two alliance partners
** got inbound and went straight to score
*** 3 high goal shots, one low low goal shot
**** made the low goal after missing the high goal shots

This process requires six scouters (one for each robot) and takes up a lot of paper (one half-sheet form per robot per match), but the completeness of the data really helped us evaluate the competing robots’ capabilities and helped a ton in strategizing for our matches and picking our winning alliance.

What variables did you guys watch for? Was it just quantitative or also qualitative? I’m always interested in what other teams are looking at.

Our team, 4362 tried a similar system to record information – it gave us a lot of data but we had issues keeping up with data entry. We tried to put all of the data from the sheets to update a spreadsheet and this proved to be far to timely at best, and led to not even finishing entering half the matches of data by the end of the competition.

I was wondering what you guys did to go about processing your data as our current system of the spreadsheet is clearly far too inefficient.

We kept all our data in a 2 inch binder with sheet protectors (one labeled for each team). We had a seventh scouter who’s job was to collect the six half-sheets at the end of each match, look over them and give any necessary feedback, and then put each sheet into their respective teams’ sleeve. Whenever we had a match coming up, I’d take the binder down to the pits, browse the pages on our allies and opponents, then discuss with the drive team and our allies what strategies would work best for the match.

As far as alliance selection goes, we looked at all the data Friday evening, made a preliminary pick list, and kept an eye out on Saturday morning for any developments that might warrant changes to the list.

So more or less you just use the sheets themselves for compilation and run them by the drive team before each match?
That’s almost a perfect compromise between the two systems we’ve tried so far. Have you tried a sheet for data that the coach can hang on to or do you just use it to get a feel for the teams in qualifications to develop a strategy?

Thank you for helping by the way, this is our first year where we’ve had enough of a robot and a team to start working on a scouting system. It also happens to be our first year going to worlds so in the next couple of weeks our goal is to get something more concrete set up.

With our current system, it’d be hard to retain multiple copies of the data. I do use it to to get a feel for the robots and develop strategies to discuss with the drive teams, but I always have the binder with me during these discussions so anyone can see what data these judgements are coming from and provide their own insights.

Okay that makes sense. Thank you again for your insight and good luck at worlds!

No problem, best of luck to you too!:slight_smile:

I’ve posted our scouting system here. In the next day or so we will be updating it to account for ground/handoff passes. We’ve successfully used for three events this year.

In terms of pre-match strategy, the strategy team in the stands write down the stats of the 6 teams on a sheet of paper and outline our suggested strategy. This sheet then gets relayed the to drive team in the pit.

With our four data collectors in the stands, we only take quantitative data to gather the hard numbers. For that, we look at how many scoring cycles each robot was in and how much their scoring cycle was worth. We also look at how many high goals, low goals, and truss shots were made by each robot. We found that by finding out how many scoring cycles each robot was in and how much their cycles were worth, it gave us a very good idea of how well that robot was performing. For example, if a robot was in 6 scoring cycles during a match but only made 1 point during those six cycles, we would not consider it superior to a robot who was only in 5 scoring cycles but made 20 points. Did that make sense at all? I hope it did. So, for all the other teams reading this, do you guys use paper to take data? Or do you use an app? Or even another program to do it?