FRC Scouting Guide: How to Scout and Build a Picklist

Winning in FRC is rarely about having the single best robot in the room. It is about understanding every robot in the room well enough to predict matches, pick the right partners, and play to your alliance's strengths. That is what scouting does, and good scouting is one of the highest-leverage things a rookie or mid-tier team can invest in.

What scouting actually is

Scouting is the process of systematically recording what robots do during qualification matches so your team can make better strategic decisions later. There are two kinds, and you want both:

• Quantitative scouting records numbers, match by match: how many game pieces a robot scored, where it scored them, whether it climbed or completed an endgame task, and how many seconds its scoring cycle took.
• Qualitative scouting records judgment that numbers miss: Is the robot reliable or does it brown out? Can it play defense? Does the drive team make smart decisions? Did it sit dead on the field?

Most teams assign one scout per robot per match (six scouts for the six robots on the field) and log data on paper sheets or a shared app, then aggregate it into a spreadsheet. The goal is a clean, per-team summary you can sort and compare.

If you are setting up a scouting system from scratch, start with our Scouting and Strategy guides before you build a spreadsheet, because the structure of your data sheet decides what questions you can answer later.

OPR and EPA: what the metrics mean

You do not have to compute everything by hand. Two community metrics give you a fast first read on any team.

OPR (Offensive Power Rating) estimates a team's average point contribution to its alliance using linear algebra: it takes every alliance's final score across the event and solves a least-squares system to assign each team a number. OPR is free on The Blue Alliance and requires zero manual scouting, which is its biggest strength. Its weakness is that it assumes scoring is linear and shared cleanly across three robots, so it gets fuzzy in games with non-linear scoring, heavy defense, or penalties. As The Blue Alliance puts it, OPR "can always supplement, but never replace, proper scouting."

EPA (Expected Points Added), from Statbotics, is a newer model. It builds on the Elo rating concept but reports results directly in points, so you can read it much like an OPR. EPA is a running average of a team's contribution that updates after every match, and Statbotics reports it as more predictive and better calibrated than both Elo and OPR.

Treat both as starting points, not verdicts. They tell you who is generally strong; they do not tell you whether a robot can play the specific role your alliance needs.

Turning data into a picklist

A picklist is a ranked list of the teams you want to play with in the elimination rounds. Because a standard FRC event fields eight alliances of three robots each, your list should comfortably cover the top 24 or so teams. Most teams keep three lists:

• A first-pick list: complete robots that would be strong alliance captains' partners.
• A second-pick list: specialists who fill a specific gap, such as a fast defender or a reliable endgame robot.
• A do-not-pick list: teams to avoid because of unreliability or safety, regardless of raw output.

Build the lists by combining your scouting data with strategy. Ask what your robot does well, what it cannot do, and which partner abilities would complete the alliance. A team with a mediocre OPR but a flawless climb or strong defense can be worth more to you than a higher-rated robot that does the same thing your robot already does.

How alliance selection works

Knowing the selection mechanics keeps your list realistic. At official FRC events, the eight highest-ranked alliance captains pick in order from Alliance 1 to Alliance 8 for their first choice. For the second choice the order reverses, running from Alliance 8 back up to Alliance 1, a serpentine pattern that gives lower seeds an earlier second pick. Each invited team accepts or declines on the spot, and a team that declines cannot be invited again.

Because picks vanish in real time, never bring a single ranked column. Keep your list dynamic and cross teams off as they get chosen, so you are always looking at the best robot still available.

Scouting well is a team-wide habit, not a one-person job. Get your data structure, your metrics, and your picklist process right, and you will out-strategize teams with flashier robots. For step-by-step templates and strategy breakdowns, head to our Scouting and Strategy department.