Scouting score

Last updated 27 August 2026

A weighted average of six things a scout watched, on a 0-100 scale. It is computed on your device from your own team's scouting records — no match results, no VEX data, nothing from anyone else's scouting.

The rule

score = 100 \times \frac{\sum_{i \in A} w_{i} v_{i}}{\sum_{i \in A} w_{i}}

the traits actually answered — "Unsure" is not in it

w_i

that trait's weight, from the table below

v_i

the answer's value on a 0-1 scale

The denominator is the weight answered, not 100 — so an unanswered trait costs confidence, not points.

Weights

They sum to 100, and each one is an argument about how Override scores:

Trait	Weight	Why
Scoring	22	Cycles are the bulk of the points.
Autonomous	20	The 12-point bonus, plus AWP eligibility.
Under defense	18	What separates a qualification partner from an elimination partner.
Driver	16	The same robot in different hands is a different robot.
Reliability	14	A robot that breaks down scores nothing at all, whatever else is true of it.
Stability	10	The Midfield is 8 points per robot at the buzzer, and a tipped robot gets none of it.

What each answer is worth

Trait	0	Middle	1
Scoring	Slow 0.15	Steady 0.55	Fast 1.00
Autonomous	None 0.00	Works 0.55	Strong 1.00
Under defense	Folds 0.10	Holds 0.60	Immune 1.00
Driver	Weak 0.15	Solid 0.55	Strong 1.00
Stability	Tippy 0.15	Stable 0.65	Planted 1.00
Reliability	1 0.00	3 0.50	5 1.00

Only reliability is linear — a 1-5 rating mapped as (n − 1) / 4. On the three-way scales the middle sits above the midpoint: the gap between folds and holds decides whether a team is pickable at all, while holds to immune is a bonus on top. Floors are non-zero everywhere except autonomous — a slow, tippy robot still occupies an alliance slot, but a missing autonomous earns no bonus and no AWP.

Confidence

Scoring "Unsure" as a zero would mean a scout who caught one match rated a good team badly, so it leaves the average entirely and shows up here instead.

confidence = \frac{\sum_{i \in A} w_{i}}{100}

Below 0.5 Vantage shows the count of answered traits instead of a number. Half rather than 40%, because scoring and autonomous alone are 42 of the weight — a lower floor would publish a confident "100" off two taps.

Worked example

Fast · auton Works · Holds under defense · Strong driver · Planted · reliability unrated:

Trait	Answer	w	v	w · v
Scoring	Fast	22	1.00	22.0
Autonomous	Works	20	0.55	11.0
Under defense	Holds	18	0.60	10.8
Driver	Strong	16	1.00	16.0
Reliability	Unsure	—	—	—
Stability	Planted	10	1.00	10.0
Total		86		69.8

score = 100 \times \frac{69.8}{86} = 81 confidence = \frac{86}{100} = 0.86

What is not in it

Threat level and the play verdict (Pick / Rush / Defend / Ignore / Avoid) are conclusions, not observations. Scoring them would read a scout's own verdict back to them.

This replaced a single 1-5 "overall" rating, which put eight teams on every value at a 40-team event and sorted a pick list into five blocks with no order inside them.

What it cannot see

These weights are not fitted to anything. Vantage Points was tested against every match of a season; this is a rubric. OPR, CCWM, TrueSkill and EPA all come from match scores and none can see whether a robot folds under defense, so there is no standard here to implement — these numbers are an argument you are free to disagree with.
Scores are comparable only inside one team's scouting. An 81 from your scouts and an 81 from someone else's came from different eyes and different matches.
It does not know what happened. No result, ranking or margin reaches this number — it is the half of scouting that results cannot give you, meant to be read next to a team's VP rather than instead of it.
Confident is not the same as right. Six answers from one match count the same as six from ten.

Vantage is not affiliated with, endorsed by, or sponsored by VEX Robotics, Innovation First, or the REC Foundation. Questions or corrections: [email protected] · Vantage · Docs · Privacy