OPR

OPR, short for offensive power rating, is an estimate of how many points a team adds to its alliance’s score in a typical match. It is the number most of PitStop is built on.

FTC only reports scores per alliance, never per robot, so OPR works backwards. Across every match at an event, it finds the set of team numbers that best explains all the alliance scores at once. Each team’s number is its OPR. PitStop leaves penalty points out, so a team gets no credit for fouls the other alliance committed.

A handful of matches is not enough to tell teams apart, so each team starts from a guess based on its earlier events this season and how it finished last season. The matches at the current event take over quickly. Older information counts for less, and less again for a team that has been away for weeks, because robots change between events.

PitStop uses OPR because it tested best. Every traditional FTC event from 2021 to 2026 was replayed, predicting each match from only the matches before it. OPR with these starting guesses beat every alternative tried, including several versions of EPA, in every season. The accuracy page has the full results.

The same method runs on each part of the score, which gives auto OPR, hive OPR, tips OPR and the rest. Each part is solved on its own, so the parts will not always add up to total OPR exactly. Every variant is listed in the glossary.

Rankings: RS, TBP and NP

Official rankings come straight from FIRST. PitStop does not reorder them. RS, the ranking score, is a team’s average ranking points per match, and it decides the order. TBP, tiebreaker points, is the average match score without penalties, and it separates teams with the same RS.

NP means non-penalty: an alliance’s score with the other alliance’s foul points taken out. NP avg, NP max and NP min describe the scores of the alliances a team played on, and spread shows how much those scores move around from match to match. A high NP avg with a small spread is a team you can count on.

Ranking points

In BIOBUZZ qualification matches, each team on an alliance earns:

Win or tie
3 for a win, 1 for a tie
Swarm
1 when the alliance scores 16 or more leave and park points
Pollinator 1
1 when the alliance makes 4 or more hive tips
Pollinator 2
1 more when the alliance makes 7 or more hive tips

That is at most 6 ranking points per match. PitStop predicts each bonus before the match. The Swarm chance combines each robot’s record of leaving and parking with the alliance’s predicted leave and park points. The Pollinator chances treat tips OPR as an expected count and ask how likely the alliance is to reach each threshold. Both were checked against this season’s results, and the accuracy page shows how close they came.

Predictions and win chances

For a match that has not been played, each alliance’s predicted score adds up its two teams’ OPRs, shifts them up or down for how high-scoring the event has been, and adds the foul points the other alliance tends to give away.

The win chance comes from the gap between the two alliances’ ratings, turned into a percentage by an S-shaped curve fitted to five past seasons. The curve stays closer to 50/50 when the teams have played only a few matches at the event, or when the model is unsure about them for other reasons. An evenly matched game has about a 2.4% chance of a tie.

A 70% win chance means the favorite should lose about 3 times in 10. Upsets are expected, not mistakes. Replayed on past seasons, the predicted chances lined up with how often teams actually won to within a few percentage points.

Teams with no history this season are treated as weaker than average when picking a winner, because that is what the past data shows. For predicted points, a separate estimate without that penalty is used, so new teams’ scores do not come out too low.

Rank projections and playoff odds play out the rest of the schedule 10,000 times using these chances, then count how often each team lands in each spot.

At stake and what changed

What’s at stake

Shown on upcoming qualification matches at events with playoffs. For each team in the match, it gives the chance to finish high enough to be an alliance captain: as things stand now, if they win, and if they lose. PitStop gets these by playing out the rest of the qualification schedule many times, once with this match forced to each result. The wider the gap between the win and loss numbers, the more the match matters to that team.

What this match changed

Shown on played qualification matches. It lists each team’s rank and OPR just before and just after the match, so you can see who moved and by how much.

Queue times

Events rarely run to the printed schedule. PitStop estimates when each match will start from how the event is actually running: it measures the gaps between the last several matches played, ignores breaks, and moves the rest of the schedule along at that pace. Scheduled breaks, like lunch and overnight, are kept.

Your queue time is the estimated start of the match two before yours, but never more than 20 minutes before your own match is expected to start. At events that use Nexus for queueing, Nexus’s own queue times and calls, like now queuing and on deck, come first.

The accuracy page shows how close the start time estimates came on this season’s events, next to the printed schedule.

Data sources

FIRST FTC Events
The official source for events, teams, schedules, scores, rankings, alliances and awards.
FTCScout
Season-wide match history behind season OPR and world ranks, the season’s event list, and a backup when the FIRST API is slow or down. Event pages say when they are showing FTCScout data.
Nexus
Live queueing status, at events that use it.

During a live event, PitStop checks FIRST for new scores and rankings every 15 to 20 seconds, and open event pages refresh themselves every 30 seconds. Pit screens check for new results every 15 seconds and queue status every 10 seconds. Season ratings are rebuilt about once an hour. Scores appear once the event’s scorekeeper sends them to FIRST, so a short delay after a match ends is normal.

Impossible scores

Now and then a score is entered wrong. In one case a real match was recorded as 7,308 to 8,355. Because OPR solves every match at an event together, a single bad score can throw off every team at that event, and even the season-wide numbers.

So PitStop leaves a match out of its ratings and predictions when an alliance total is above 600, or when auto and teleop points do not add up to the non-penalty score. The match still appears in results with its official score. It just does not feed the model.

Limits

  • OPR assumes points simply add up. Defense, partners who work especially well or badly together, and one robot carrying an alliance all get spread across the teams involved.
  • Early in an event, ratings lean on earlier events and last season. A team that rebuilt its robot will look off for its first few matches.
  • Win chances ignore fouls. They were tested and turned out too random to predict per team.
  • BIOBUZZ is new this season, so ranking point chances, the tie rate and the score spread were fitted on this season’s matches alone. They will settle as more matches are played.
  • Teams with no history are assumed to be weaker than average. That is usually right, but it undersells a strong first-year team until it has played a few matches.
  • The numbers are only as good as what is entered at the event. Corrected scores show up on the next refresh.
  • None of this is official. For official rankings, schedules and results, use FTC Events.

Glossary

Every short label used in PitStop’s tables. Hover over one in a table to see it, or tap it to come here.

#
Rank
RS
Ranking score, average ranking points per match
TBP
Tiebreaker points, average match score without penalties
Record
Wins, losses and ties
OPR
Offensive power rating, estimated points per match
Season OPR
Offensive power rating over the whole season, estimated points per match
OPR here
Offensive power rating at this event, estimated points per match
Starting OPR
Season OPR coming into this event
Best event OPR
Highest offensive power rating at any one event this season
Rank by OPR
Where the team would rank at this event if sorted by OPR
World rank
Rank among all teams by season OPR
World rank by OPR
Rank among all teams by season OPR
Partners' OPR
Average OPR of qualification partners
Opponents' OPR
Average OPR of qualification opponents
Auto OPR
Offensive power rating for auto, estimated auto points per match
Teleop OPR
Offensive power rating for teleop, estimated teleop points per match
Hive OPR
Offensive power rating for the hive, estimated hive points per match
Flower OPR
Offensive power rating for flowers, estimated flower points per match
Garden OPR
Offensive power rating for the garden, estimated garden points per match
Park OPR
Offensive power rating for parking, estimated park points per match
Tips OPR
Offensive power rating for tips, estimated hive tips per match
NP avg
Average score without opponent fouls
NP max
Best score without opponent fouls
NP min
Worst score without opponent fouls
Spread
Standard deviation of scores without opponent fouls
Avg tips
Average hive tips per match
Avg auto
Average auto points per match
Avg park
Average park points per match
Swarm
Swarm ranking point, leave and park 16+ points
Poll. 1
Pollinator 1 ranking point, 4+ tips
Poll. 2
Pollinator 2 ranking point, 7+ tips
Bonus RS
Average bonus ranking points per match from Swarm and Pollinator
Pred.
Predicted score
Auto
Points scored in the autonomous period
Teleop
Points scored in the driver-controlled period
Hive
Teleop points from pollen and nectar in the hive
Flower
Teleop points from flowers the alliance owns
Garden
Teleop points from pollen and nectar in the garden
Park
Teleop park points
Tips
Hive tips in auto and teleop
Fouls
Points given by the other alliance’s fouls

Type a team number or name, a city or state, or an event name or code.