How Does the Kattis Ranklist and Scoring Work?

Updated

Kattis ranks users by total score: the more points you accumulate from solved problems, the higher you appear on the ranklist. The site publishes three separate leaderboards — individual users, affiliations (usually universities), and countries — so the same solving activity feeds into several competitive views at once. The exact point value of any single problem is set by Kattis and shown on the problem itself; the homepage only displays a difficulty label and a point figure for each suggested problem.

The three ranklists

The Kattis Problem Archive homepage shows a "Ranklist" section alongside "Affiliations" and "Countries." Each is a top-8 snapshot rather than the full standings.

Ranklist What it ranks Example from the homepage
Ranklist Individual users by score 1. Nick Wu — 18448.0
Affiliations Institutions by score 1. National University of Singapore — 7,926.9
Countries Countries by score 1. United States — 9,673.7

The user list is ordered purely by score, descending. The affiliation and country lists use the same ordering logic but aggregate across everyone associated with that institution or country.

How scoring works

Each problem carries a point value. On the homepage's "Suggested problems" table, values run from 1.3 pt (Cetvrta, labeled Trivial) up to 7.8 pt (Pipe Stream, labeled Hard):

  • Trivial — Cetvrta, 1.3 pt
  • Trivial — Sum Squared Digits Function, 1.4 pt
  • Easy — Lawn Mower, 1.7 pt
  • Easy — Robot Protection, 2.2 pt
  • Medium — Bitmask, 3.2 pt
  • Medium — Elliptic Curve Point Multiplication, 3.7 pt
  • Hard — Flow Finder, 7.7 pt
  • Hard — Pipe Stream, 7.8 pt

Two things follow from this table. First, difficulty labels and point values rise together in the sample shown, so harder problems are worth more. Second, the spread is wide — a hard problem is worth roughly six times a trivial one — which means a ranklist position reflects both how many problems you solve and how hard they are.

Your total score is the sum of the points from the problems you have solved. To move up, you either solve more problems or solve higher-value ones.

Reading a ranklist position

A score alone doesn't tell you how a user got there. Two users near each other on the list may have very different profiles — one grinding through many easy problems, another clearing a smaller number of hard ones. The homepage doesn't break scores down by difficulty, so the ranklist shows the total only.

The same applies to the affiliation and country tables. A country's score aggregates its users, so a country with many active solvers can outrank one with a few strong ones. The homepage shows the aggregate, not the number of contributors.

Practical implications

  • If you want a higher personal rank, prioritize problems with larger point values once you can handle them; a single 7.7 pt solve is worth several trivial ones.
  • If you're tracking an institution or country, remember the score is a sum across all associated users, so recruiting more solvers raises the total even if individual skill doesn't change.
  • If you're new, the "Suggested problems" table is a reasonable entry point: start at the Trivial and Easy rows, then work upward as the point values climb.

The homepage presents only the top 8 in each category. For anything beyond that, you'd need the full ranklist pages, which the homepage excerpt doesn't cover.