Candidate grades

What an A+ candidate actually means.

A grade tells you how much of a candidate's performance comes from their own strength versus the environment they ran in. It is not just a win-loss record.

Browse the leaderboard Margin convention

Offices scored

Senate · Governor

Only office_id 3 and 5

Baseline

State presidential

Same cycle, same state

Grade tiers

A+ through D, percentile-based

The concept

Winning isn't enough

A candidate who wins a safe seat by 12 points when the state leans 20 points their way is actually underperforming. A candidate who loses by 2 points in a 15-point environment is overperforming dramatically.

Overperformance, not raw margin

We compare each candidate's margin to the state's presidential margin in the same cycle. The difference — positive or negative — is their overperformance for that race.

Career score, not single race

A single great race does not make an A+. Career scores are the sum of race scores across all graded contests. Candidates need at least three scored races to earn a career grade.

Step 1

The baseline

State presidential margin, same cycle

We use the state's presidential margin in the same election cycle as the race being scored. For a 2020 Senate race in Michigan, the baseline is Michigan's 2020 presidential margin.

Why state, not national

National baselines distort scores for partisan strongholds. A Mississippi Democrat winning by 5 points is extraordinary, not average. State baselines correctly anchor the comparison to the ground truth of that state's politics.

Updated every four years

Presidential baselines shift as states shift. A state that was D+2 in 2012 and R+8 in 2016 will show that swing in how candidates are scored across those cycles.

Non-presidential cycles

For midterm races (no presidential on the ballot), we still use the most recent state presidential margin. It is the best single indicator of the state's partisan baseline for that era.

Step 2

Incumbency adjustment

Incumbents get a penalty — not a bonus — on their overperformance score.

Why a penalty

Incumbents carry built-in advantages: name recognition, fundraising networks, franking privileges, case work. A challenger who beats the baseline without those advantages deserves more credit than an incumbent who beats it with them.

The direction, not the number

We apply a three-point downward adjustment to incumbent Senate and Governor candidates' race scores. Lower-profile statewide offices get smaller adjustments. The exact values are calibrated against the observed incumbency advantage in our historical data.

Plain-language summary

Race score = overperformance + incumbency adjustment

Overperformance is the margin advantage over the state presidential baseline, signed so positive means the candidate beat expectations. Incumbency adjustment subtracts a few points from incumbents. That's it.

Step 3

Career grade

Grade scale

Grades are percentile-based against all scored candidates. Percentile ranges here are approximate — we recalibrate as the pool of scored candidates grows.

A+Top 5%

Elite

Consistent, dramatic overperformance across a full career.

ATop 13%

Excellent

Regularly beats the state environment by meaningful margins.

A-Top 23%

Very good

Strong overperformance most cycles, occasional softness.

B+Top 35%

Good

Noticeably better than the baseline in most races.

BTop 48%

Above average

Slightly outperforms the environment on balance.

B-Top 60%

Average

Roughly matches the partisan environment of their state.

C+Top 72%

Below average

Underperforms the environment more often than not.

CTop 82%

Weak

Consistently loses ground relative to the partisan baseline.

C-Top 90%

Very weak

Underperforms by wide margins across multiple races.

DBottom 10%

Worst

Substantial, repeated underperformance against the state baseline.

Minimum three races

Candidates need at least three scored races to earn a career grade. Short careers are listed without a grade rather than graded on too little evidence.

Percentile, not absolute

Grades are percentile-based against the full pool of graded candidates. As the pool grows or political dynamics shift, we recalibrate. A 2020 A-minus might read differently when 2030 data is in hand.

Out of scope

What we don't score — and why

Scoring everything would produce misleading numbers. Some offices are legitimately not comparable against a state presidential baseline.

Presidential races

The candidate IS the baseline

Trying to score a presidential candidate against the presidential vote is circular. We exclude presidential candidacies from scoring.

U.S. House races

State baselines are the wrong yardstick

A state's presidential margin averages districts that vary enormously in partisan lean. Scoring a district-level race against a state-level baseline produces nonsense.

Statewide executive offices

Ticket-splitting destroys comparability

Lieutenant Governor, Attorney General, Secretary of State, and similar offices show extreme ticket-splitting in some states. Scored against presidential baselines, West Virginia Democrats would appear to be +48 performers, and Vermont Republicans would appear to be -28. The numbers are real; the comparison is meaningless.

What we do score

Senate (office_id 3) and Governor (office_id 5)

Both offices are statewide, partisan, and face the electorate under roughly the same environment as the presidential race. Scoring is defensible. Everything else we leave unscored.

Worked example

How a score gets computed

Conceptual walkthrough using sample values — the actual pages use live data.

Imagine a Democratic Senate incumbent in a state that voted R+8 in the most recent presidential cycle. They win reelection by 3 points.

Baseline: R+8, so the state’s partisan environment is 8 points against Democrats.

Raw margin: D+3. The Democrat won by 3 points.

Overperformance: +11 points. The Democrat performed 11 points better than the state’s baseline would predict.

Incumbency adjustment: -3 points. Incumbents get a downward adjustment.

Race score: +8 points. A strong single-race score that contributes to a good career grade if sustained over multiple cycles.

Repeat across every race in the candidate’s career. Sum the scores. Rank against the pool. Assign a letter grade.