--- title: "Scoring providers" description: "Choose how rCTF calculates point values for decay challenges." order: 4 --- Scoring providers calculate point values for decay challenges. rCTF includes six algorithms that use solve counts, event timing, or both. ## Configuration ```yaml scoreProvider: name: scores/classic # Default ``` Each challenge sets a point range via `points.min` and `points.max` in its data. The scoring provider calculates the current point value from the solve count and other context. ## Score context All scoring providers receive a context object with: | Field | Description | | --- | --- | | `minPoints` | Minimum points for the challenge (floor) | | `maxPoints` | Maximum points for the challenge (ceiling, when 0 solves) | | `solves` | Current number of solves | | `maxSolves` | Maximum solves across all challenges (used by `scores/legacy`) | | `eventStartTime` | Competition start time (used by `scores/jammy`) | | `eventEndTime` | Competition end time (used by `scores/jammy`) | | `firstSolveTime` | Time of the first solve for the challenge (used by `scores/jammy`) | ## Providers Charts use a sample challenge with `minPoints: 100` and `maxPoints: 500`. Solve-based providers show the full provider family with the current tab highlighted, while `scores/jammy` uses its own time-based curve. ::::tabs :::tab[scores/classic]
0 100 200 300 400 500 0 10 20 30 40 50 Solves Points
scores/classicscores/sekaiscores/steepscores/genniscores/legacy
The **default** scoring algorithm. Uses a logistic decay function that drops points as more teams solve the challenge. ```yaml scoreProvider: name: scores/classic ``` Challenges start at `maxPoints` with zero solves and decay toward `minPoints` as the solve count climbs. The decay curve is moderate, which works for most CTFs. ::: :::tab[scores/sekai]
0 100 200 300 400 500 0 10 20 30 40 50 Solves Points
scores/classicscores/sekaiscores/steepscores/genniscores/legacy
Logarithmic decay with tuned constants (gradient=10, decay=60). Produces a gentler curve than `scores/classic`. ```yaml scoreProvider: name: scores/sekai ``` ::: :::tab[scores/steep]
0 100 200 300 400 500 0 10 20 30 40 50 Solves Points
scores/classicscores/sekaiscores/steepscores/genniscores/legacy
Similar to `scores/classic` but with a steeper decay curve. Points drop more aggressively with early solves, so challenges differentiate faster. ```yaml scoreProvider: name: scores/steep ``` ::: :::tab[scores/jammy]
100 200 300 400 500 0 20 40 60 80 100 Event elapsed before first solve (%) Points
scores/jammy
Time-based scoring that values challenges by when they were first solved. Unlike the other providers, this one looks at the first solve time relative to the event duration rather than how many teams solved the challenge. ```yaml scoreProvider: name: scores/jammy options: maximumScoreTime: 0.8 # Optional, default 0.8 ``` | Option | Default | Description | | --- | --- | --- | | `maximumScoreTime` | `0.8{:ts}` | Fraction of event duration used to ramp from minimum to maximum points. First solves after this threshold receive maximum points. | ::::note This provider needs `startTime` and `endTime` set in the configuration. It uses the first solve time to anchor the scoring curve. :::: ::: :::tab[scores/genni]
0 100 200 300 400 500 0 10 20 30 40 50 Solves Points
scores/classicscores/sekaiscores/steepscores/genniscores/legacy
A binary scoring algorithm. It returns `maxPoints` if the challenge has 0 to 2 solves, and `0{:ts}` points otherwise. Good for challenges that shouldn't contribute to the score once they're widely solved. ```yaml scoreProvider: name: scores/genni ``` ::: :::tab[scores/legacy]
0 100 200 300 400 500 0 10 20 30 40 50 Solves Points
scores/classicscores/sekaiscores/steepscores/genniscores/legacy
Hyperbolic tangent-based decay that normalizes against the maximum solve count across all challenges. This was the original rCTF scoring algorithm. ```yaml scoreProvider: name: scores/legacy ``` ::::note This provider needs `maxSolves` (the maximum solve count across all challenges) to normalize the curve. Expect surprising results if challenge solve counts vary widely. :::: ::: ::::