Understanding OPR, DPR, and CCWM

The idea behind OPR

OPR (Offensive Power Rating) estimates how many points a team contributes to its alliance on average, using only final match scores and the schedule. Its key assumption (from The Blue Alliance's Math Behind OPR): an alliance's score is a linear combination of its three members' individual contributions. So if teams A, B, and C scored 50 together, that becomes one equation A + B + C = 50.

Across a whole event you get many such equations, usually more equations (matches) than unknowns (teams). That is an overdetermined system with no exact solution, so OPR is computed with the least-squares method, which finds the per-team values that minimize the total squared error across all matches. More matches give the solver more equations to work with.

DPR and CCWM

Two relatives are computed the same way:

• DPR (Defensive Power Rating) estimates how much a team lowers the opposing alliance's score, used as a rough defense proxy. (It is calculated like OPR but using the opposing alliance's score.)
• CCWM (Calculated Contribution to Winning Margin) equals OPR minus DPR, an estimate of net contribution to the score difference (how much a team helps you outscore the other alliance).

Strengths

• Free and automatic. TBA computes it for every event with zero scouting effort.
• Year-agnostic. Because it only needs scores, OPR works for almost any season.
• A useful first cut. A team's OPR is a quick estimate of offensive strength and a sanity check on your scouting.

Serious limitations

The Blue Alliance's own OPR and You article is blunt about the weaknesses:

• Linearity breaks down. "The more the actual game deviates from the linear model, the less accurate OPR usually is." Many games have non-linear scoring (bonuses unlocked at thresholds, or actions that require all three robots), which violates OPR's core assumption, so its predictive accuracy varies a lot year to year.
• It can mis-credit. When robots interfere with each other or strong robots mostly play with strong partners, the math can spread credit in misleading ways, especially with few matches.
• It ignores defense. OPR is built purely from points scored, so a robot playing aggressive legal defense (which can suppress an opponent's score substantially) gets no credit in OPR. This blind spot is one of the most-cited reasons OPR cannot stand alone.

How to use it well

The community rule of thumb is exact: per OPR and You, "OPR can always supplement, but never replace, proper scouting." Practical guidance:

• Use OPR to rank-order offensive output quickly and to flag teams your scouting might have missed.
• Before trusting it, ask whether this year's game is roughly linear and whether contributions are independent. If not, weight it less and scout more.
• Always pair it with scouting for anything OPR cannot see: defense, reliability, consistency, and endgame success rate.

In short, OPR is a powerful, free starting point and a great cross-check, but it is an estimate built on an assumption that does not always hold.