Spartronics Scouting Process, 2016
@ Mount Vernon Regional
Jack Stratton, Team 4915
Intro
Scouting for alliance selection has been a pretty hair-tearing-out process in the past. Nobody likes staying up until midnight sorting through papers trying to extract some meaningful data, yet that was what had to happen. The aim of the scouting papers and data analysis this year is to assist scouters in collecting and looking through the information collected about other teams, efficiently.
Day 0: Before a Competition
Six lead scouters are assigned at the meeting before a competition. Any suggested changes to the scouting sheets are made, and (12 / 2) * (# teams at event) + 10 sheets are printed. A script is also run which generates a spreadsheet customized for this event, based on data pulled from The Blue Alliance.
Day 1: Qualification Match Scouting
The sheets themselves have boxes to put in the team number, match number, number of points scored, information on what defenses the team can cross, etc. There's also boxes for qualitative data like "how much work did this team do" and "other comments".
They're given a clipboard labeled R1, R2, R3, B1, B2, or B3. These correspond to alliance stations on the red or blue side, numbered 1--3. Clipboards, and the people using them, don't have assigned teams. They scout whatever team is in their slot at the time. Clipboards have a stack of these sheets (printed 2-up). Once both halves are filled out, they're given to someone to record the information in the pregenerated spreadsheet.
Back at the hotel, the lead scouts, team captains, some mentors, and anyone else involved gathers to discuss team rankings. Using a combination of the scouting spreadsheet data, The Blue Alliance team rankings, and background knowledge, lists are created. Disregarding the top six alliance captains, unless we're one of them, we sort the remaining teams into three categories -- "definitely pick", "maybe pick", and "never pick". These lists are then sorted by preference, so that when a team on our list is picked, we can cross it off and move down the list to find the next best match.
Day 2: Decision Making
Scouting continues even after making the three lists. Usually there are more qualification matches than fit in one day. Based on our observations -- mostly qualitative at this point -- we rearrange teams on the list until the very last match. The final list is written down and given to the driveteam coach to bring to the alliance selection stage.
We bring the information regardless if we're selecting in the first round, because the data we gather can still be used by the alliance that picks us to help find a third alliance member.
Does it Work?
We think the data gathered is really useful in identifying who would work well with us. As far as the scouting process, the response of the scouts have been overwhelmingly positive. Overall this was a great improvement over the previous scouting processes.
Previous Method
The last version of this process was pretty different for the scouts in the stands. We had a hanging file folder box and one labeled folder for every team at the event. The lead scout would find a blank sheet, label it with the team number, and give it to the scout with the clipboard associated with the team's current position in the alliance station. After one match, the lead scout takes all the filled out sheets and files them back in to the right folder. This was overall a bad decision.
We bought the folders intending to improve organization of the papers. It succeeded, but wasn't at all worth the effort. Since the information on the sheets was being (almost) immediately entered into a spreadsheet, there's actually very little necessity to keep track of the files so meticulously. Storing and retreiving six sheets from six folders every single match takes an incredible amount of time.
Everyone involved prefers the current method to this file folder system.
Hindsight
The scouting sheet itself places a bit too much emphasis on qualitative things like how much work a team did. Reorganizing the sheet to have dedicated places for numerical data like the number of times they crossed a defense and moving the qualitative parts into the combined "other comments" box would make analysis easier. The sheet also neglects a tally of how many times a particular defense was crossed, only tracking how many times they crossed any defense (and which defenses were easy to cross). Confusion arose when people were unsure whether to track only neutral zone to courtyard crossings or include the reverse in their tallies.
The hanging file folder has been obsoleted by Ruby and spreadsheets. It's now a bit of a waste to try and use it. Recycling the file folders, minus three -- blank, inbox, and processed -- is probably what we'll do.