Scouting Gears with "Abandoning Ship"

An interesting part of the game this year is that for many gear oriented robots, the most intelligent strategy in a match may not be to score as many gears as possible. This usually occurs when the alliance knows they won’t score any more rotors so instead of wasting time scoring more gears, the robot either defends or scores fuel after scoring gears or slows down their gear scoring cycle time by mixing in defense or fuel scoring.

This raises a problem for scouts: by simply writing down the number of gears the team has scored, you reward them for poor strategy. So the question is, how do you accurately gauge a team’s gear scoring ability without rewarding poor strategy?

2830 will be developing its scouting system in the near future for our week 4 competition so we’re wondering what other teams are doing about this.

Cycle time.
Timed values like placing a gear, a retrieval zone cycle, and other gear-related numbers allow you to accurately analyze how they might fit into your alliance, or the oppositions alliance. This allows you to make calls on how many rotors to go for, or if defensive play is worth it on an opposition that is likely to just barely engage X rotor without defense.

add something like “number of gears delivered that scored no points” as a datapoint for poor strategy.

While I’m generally of the mind that the more data you can gather the better, at a certain point you hit diminishing returns in regards to the value of that data, and the accuracy of said data.

Point being, I wouldn’t necessarily expect our final scouting system to account for cycle time, especially not for even more specific things like “player station cycle” or “neutral zone crossing”. Things like that are asking for students to look away from a given robot, even for as small amount of time as it takes to find/start a timer, something that should be minimized in my estimation of a good scouting system.

To address the OP’s point regarding numbers missing a team making a strategic decision to stop scoring gears, that is exactly why you want to have some sort of qualitative info in addition to quantitative data. I’d also wager that Team WXYZ (a pure gear-bot) would generally have a similar number of cycles match to match, and if there were any outliers then maybe those are worth diving deeper into to figure out why (match video, qualitative info, etc.).

Counterpoint: Knowing a robot can run 6 gears on average makes that robot far more valuable than one who runs 3 on average, even though they’re worth the same amount of points. And it’s much harder to determine if a robot ran unnecessary gears just because, or if they were expecting a partner to score gears that didn’t happen, resulting in not enough for the next rotor.

While I do agree with that, wouldn’t that be evident with whatever else they do?

If an alliance of teams A, B, and C does 3 rotors then decides that they cannot get the last in time, the amount of high goals/low goal scored will show that the gears weren’t their primary focus. Let’s set up an scenario:

A - Only gears, nothing else: Put on 2 gears

B - Gears, and climbing robot: Put on 2 gears & climbed

C - Gears, climbing and high goal robot: Put on 2 gears, shot ~10 kPa into high goal then climbed

While those teams put on the same amount of gears, though not completing the 4 rotor requirement, it is what they do with the extra time that can show worth in which team can fit into your alliance.