A team's APEX rating estimates how many points they are expected to contribute above an average team, adjusted for opponent strength, alliance partners, event difficulty, and recent performance.
Experimental, AI-assisted analytics
FRCinsights uses AI-assisted experimental analytics. Predictions, rankings, ratings, and insights are not guaranteed to be accurate and should not be treated as official results or professional advice. Data may be incomplete, delayed, miscategorized, or incorrect. FRCinsights is a new project under active development, and accuracy will improve over time.
The core idea, in one sentence
Take everything a team has done this season, figure out how hard it was given who they played with and against, weight recent and high-stakes matches more heavily, and express the result as a single number: expected points contributed above average.
Release notes
Every APEX model update, newest first · currently running APEX 0.5.1
Now liveAPEX 0.5.1August 20, 2026
Fairer credit sharing, so strong fields stop inflating everyone
APEX now works out each robot's share of the score by looking at the whole event at once instead of just that team's own matches. Teams at very strong events are no longer lifted by their partners, and teams at weaker events are no longer held back.
78.2%
Winner accuracy
held-out, in time order
75.6%
Championship fields
toughest matchups
18,788
Matches replayed
2026 season
0.004
Calibration error
lower is better
01
Your rating measures you, not your field
Before, a team's number was worked out from only the matches they played, so a robot at a stacked championship division soaked up the scoring level around them. Every robot at the event is now solved together, so the number reflects what you actually add.
02
Qualification matches lead, playoff repeats count once
Qualification play shuffles your partners, which is what makes it possible to tell robots apart. Playing the same three-robot lineup several times in a row is really one lineup seen repeatedly, so those replays now share credit instead of each counting in full.
03
Penalties and broken matches no longer count as scoring
Points handed to your alliance from opponent penalties are removed before ratings are worked out, and clearly broken matches (field faults, no-shows, wild outliers) are dropped. Expected penalties are added back only when predicting a score.
04
Weakly supported numbers are labeled, not hidden
If an event gave too few matches or kept pairing you with the same partners, we mark the rating with a small question mark and pull it gently toward the field average. Genuine improvement still shows through.
05
Checked against matches the model had never seen
Across 18,788 completed 2026 matches replayed in time order, winner accuracy stayed at 78% with better calibration, and last event's rating now predicts how a team actually does at their next event noticeably better than before.
Better predictions on strong fields, and honest defense and clutch numbers
APEX 0.5.0 · released August 12, 2026
A major prediction upgrade. Alliance scores are no longer just three ratings added together, defense is measured by how much an alliance holds opponents below what they normally score, and clutch only counts matches that were genuinely close.
Three great robots do not simply add up
Elite alliances share the same field, the same game pieces and the same cycles, so their combined score is less than the sum of their parts. Predictions now account for that crowding, which mostly fixes over-predicted championship matches.
Opponents can hold you down
A strong opposing alliance lowers what you are expected to score. That suppression is now part of the prediction instead of being ignored.
Ratings are frozen before each match
When we grade ourselves, a match is always predicted with the ratings as they stood before it was played, so nothing is scored with hindsight.
Defense is measured, not guessed
We first work out how far an alliance pushed its opponents below their usual output, then share the credit among the robots that were repeatedly on the field for it. A quiet third robot next to two big scorers is only called a defender when the evidence repeats.
Clutch means pressure, not just wins
Clutch now compares how you did against what was expected, and only in matches that were genuinely close going in. Both defense and clutch are centered at zero, so average play reads as zero rather than negative.
Teams arriving from earlier events
What a team did at earlier events is carried in as a starting point, which makes the first matches of championship divisions far more accurate.
Ratings that match a team's most recent event, plus live updates
APEX 0.4.0 · released July 7, 2026
A team's main APEX score now reflects their most recent official event instead of a season-long average, small and noisy playoff appearances no longer swing ratings, and the site updates in near real-time while events are happening.
Your APEX now reflects your latest event
A team's main APEX ranking now reflects their most recent official event. You may have noticed teams like 1690 showing a super high APEX because they didn't play an official event until championships, which made their season score look extreme compared to everyone else. Now the headline number follows current form, so a strong championship run shows up right away.
Fewer surprises from short playoff runs
A handful of playoff matches or a quick appearance in a very tough field (like Einstein) used to move a rating too much. Those small samples now count less, while a team's full event still counts in full. The result: ratings are steadier and less thrown off by a few high-pressure matches.
Einstein counts, but won't tank a strong division
The Einstein finals still count toward a team's overall season picture, but a team's championship division result stays their main recent event. That way a few finals matches can't drag down an otherwise strong showing.
Offseason events are off by default
Casual offseason events are left out of the main rankings by default, since relaxed play doesn't reflect how teams really stack up. You can turn them back on with a toggle whenever you want.
Near real-time during live events
While an event is happening, results and insights update about every 2 minutes instead of every 20. Matches that are already finished are saved and not re-downloaded, so only new results are pulled and updates stay fast.
Everything agrees now
The league strength chart, the rankings, and the top-teams lists all use the same latest-event APEX, so a team's strength looks the same no matter where you see it on the site.
What goes in
The model blends many signals. None on its own tells the whole story, the rating comes from how they fit together.
Points a team adds through scoring, above an average robot.
Defensive rating
Estimated from how far below expectation each team's opponents scored, split across the team's alliance and adjusted for sample size. Labelled as an experimental estimate.
Auto rating
Expected contribution during the autonomous period.
Endgame rating
Expected contribution from endgame and climb tasks.
Every metric, explained
What it means · how it's calculated · what feeds it · when to trust it.
The adjustments that matter
Partner & opponent adjustment
Your numbers are corrected for the strength of who you played with and against, so easy schedules don't inflate ratings.
Event strength context
Event pages show how strong the field is, so users can compare raw event results with the level of competition.
Current-form ranking
The main rankings use a team's latest official event APEX first, which better reflects the version of the team playing right now.
Playoff and Einstein inclusion
Playoff matches and Einstein matches count in the model, while tiny samples are pulled toward the team's season baseline so one unusual run does not dominate.
Handling missing & messy data
Missing breakdowns: overall APEX can still use final scores, while phase ratings show less detail until official score breakdowns are available.
Tiny samples: ratings are pulled toward the mean until enough matches exist, preventing wild week-1 swings.
Small playoff samples: a short, highly competitive run is blended with the team's season baseline instead of taking over the ranking by itself.
Delayed data: stale values show an older "last updated" timestamp and reduced confidence.
How confidence scores work
Confidence (0–100) combines several factors: sample size (more matches → higher), data completeness and source agreement, recency (fresher data → higher), and schedule diversity. Low-confidence outputs are labeled, not hidden.
How predictions are tested
Every completed match feeds an audit comparing predicted vs actual winner and score. We track winner accuracy, average score error, and confidence calibration: whether a stated 70% really wins ~70% of the time. See the full Accuracy Report →
Avoiding overreaction
The model uses recency-weighted averaging with regression to the mean and outlier damping. One blowout win or a broken-down match nudges a rating slightly rather than rewriting it, so APEX tracks genuine improvement without chasing noise.
Future REBUILT scouting metrics
Definitions, not active public team ratings yet.
These are the kinds of game-specific scouting fields FRCinsights plans to support when the data is reliable enough to show on team and match pages. They are not currently used as standalone public rankings.
Game Piece Throughput
Average game pieces scored per match.
Estimate
Auto Reliability
How often the autonomous routine executes as intended.
AI-assisted
Cycle Time
Average time for one acquire-and-score cycle.
Estimate
Scoring Accuracy
Successful scores divided by attempts.
AI-assisted
Preferred Scoring Location
Where the robot scores most often.
AI-assisted
Driver Efficiency
Path efficiency and decision quality under pressure.
AI-assisted
Defense Resistance
Ability to keep scoring while defended.
Estimate
Endgame Reliability
How often the endgame objective is completed.
Estimate
Ranking Point Contribution
Estimated RP this team helps its alliance earn.
AI-assisted
Penalty Risk
Likelihood of incurring fouls (lower is better).
Estimate
Alliance Role Fit
Best complementary role on an alliance.
AI-assisted
The data pipeline
How official FRC results become ratings.
1
Fetch
Pull raw events, teams, matches and score breakdowns from public APIs.
2
Validate
Reject malformed payloads, flag missing fields and duplicate keys.
3
Normalize
Map every source into a single internal schema for teams, events and matches.
Derive APEX contributions, consistency, momentum, schedule difficulty, etc.
6
Assign Confidence
Score each output by sample size, data completeness and recency.
7
Store
Cache computed outputs so pages render instantly without recomputation.
8
Display
Render with provenance labels, freshness timestamps and uncertainty.
FRC knowledge the model assumes
Qualification matches
Randomized round-robin matches that set the ranking order.
Ranking points
Bonus points for wins, ties, and meeting game objectives, separate from match score.
Alliance selection
Top-seeded teams draft partners to form playoff alliances.
Playoff brackets
Double-elimination bracket that decides the event winner.
District vs regional
Districts use a cumulative points system across events; regionals advance winners directly.
District points
Points earned at district events that determine district championship qualification.
Championship divisions
Large championships split teams into divisions that feed a final playoff.
Strength of schedule
Some teams simply draw harder partners and opponents, APEX corrects for it.
Coopertition
Game-specific incentives that reward both alliances for cooperating.
Auto / teleop / endgame
The three scoring phases, each modeled separately.
Fouls & penalties
Rule violations award points to the opponent and add noise to raw scores.
Surrogate matches
Extra matches that don't count toward a team's ranking.
Backup robots
Substitute robots called in during playoffs, flagged so stats aren't misattributed.
Event advancement
How teams progress from districts to district champs to worlds.
Season improvement
Teams iterate their robots across weeks; recency weighting captures it.
Raw score vs true contribution
A high score can come from strong partners, APEX isolates a team's own contribution.
A note on honesty
APEX is an original model built for FRCinsights. It is not affiliated with, nor a copy of, any other analytics provider. It runs on real official FRC results, and ratings are estimates, they describe expected performance, not certainties.