Choosing the Right Metrics

Less is more

The most common scouting mistake is collecting too much. If a scout has to track twenty fields per match, they will miss action and the data will be noisy. Choose metrics that are (1) easy to record accurately in real time and (2) directly useful for a decision. If a field would not change a pick or a match plan, drop it.

Categories worth tracking almost every year

• Scoring counts by location/level. The core of any year's metrics: how many game pieces, and where. Most modern games reward higher or harder locations more, so separate them. In the 2026 game REBUILT presented by Haas, for example, robots score FUEL into a HUB, so you would tally fuel scored.
• Autonomous performance. Whether the robot moved/left its starting zone, what it scored in auto, and which routine it ran. Auto is high-value and consistent, so it is very predictive. (Note that the auto period length is set by the game manual each year — in 2026 REBUILT it is 20 seconds.)
• Cycle count or cycle time. A cycle is one full trip from acquiring a game piece to scoring it. Cycle count per match is one of the best proxies for offensive output.
• Endgame. What the robot achieved at the end of the match (climb level, park, etc.) and whether it succeeded. Endgame points are often decisive and binary, which makes them easy and valuable to track.
• Defense. How much time the robot spent playing defense and how effective it was. This is critical because point-based public metrics like OPR completely miss defense.
• Reliability flags. Disabled, tipped, fell off a mechanism, no-show, or experienced a major malfunction. These flags catch robots whose averages look fine but who fail when it matters.

Ranking points and the scoreboard

FRC qualification standings are driven by Ranking Points (RP). The exact RP rules are redefined each season in the Game Manual, but the structure is always: RP for the match result plus bonus RP earned independently by hitting that year's performance thresholds. In the 2026 REBUILT season, each qualification match awards 3 RP for a win and 1 RP for a tie (0 for a loss), plus up to 3 bonus RP — ENERGIZED (score at least 100 FUEL), SUPERCHARGED (score at least 360 FUEL), and TRAVERSAL (score at least 50 TOWER points) — for a maximum of 6 RP per match. Tracking whether opponents and partners tend to earn those bonus RP tells you how matches will be won and helps you predict the final rankings, which in turn predicts who will be an alliance captain.

Design metrics to the game, every year

Each season FIRST releases a new game, so your exact fields change every January. A good workflow:

1. Read the Game Manual scoring section the week of kickoff.
2. List every way to earn points, in auto, teleop, and endgame.
3. For each, ask: Can a scout count this reliably in real time? Will it change a decision?
4. Keep the survivors. Group them as auto / teleop / endgame to mirror how analytics tools split scoring.

A concrete starter set

For a typical game, a lean per-match form might be: Left starting zone in auto (Y/N), Auto pieces scored (count), Teleop pieces scored by location (2-3 counts), Endgame result (dropdown), Played defense (None/Some/Lots), Reliability flags (checkboxes), Notes (free text). That is roughly 7-9 fields, recordable by one focused scout. You can always add a field next event; you can rarely recover data a tired scout never captured.