How this site works

Scope: this leaderboard includes only Slovak Riftbound events (the underlying data is limited to Slovakia). It is not a global or worldwide rating pool. The rest of the page explains how Elo works here and how games are ordered; a short summary of loaded events and stores is at the end.

What this website is

Riftbound Slovakia Elo is a read-only leaderboard for Slovakia: completed 1v1 games from official data for Slovak events, turned into ratings with standard Elo math. Nothing here can be edited by visitors; the numbers follow the order games are processed in this build and the settings below.

What “Elo” means (short version)

Each player gets a single number, the rating. Roughly: higher means stronger in the pool of players who appear in this data. The scale is not “points” you earn forever; it goes up when you do better than expected and down when you do worse, so it stabilizes once you’ve played enough games.

“Expected” is not a guess about fun or skill in general—it is a precise prediction from the two players’ ratings at the moment before that match, using the standard Elo model.

One match at a time

Ratings are updated one completed game after another in a fixed global order (see below). For each game the app knows the two players and whether it was a win for seat A, a win for seat B, or a draw.

For player A (the first seat in the data row) and player B, let R_a and R_b be their ratings before the game. The model’s expected score for A is a number between 0 and 1 (think “share of the win”):

If A is much higher-rated than B, E_a is close to 1. If they are equal, E_a = 0.5. Player B’s expected score is E_b = 1 − E_a.

The actual score is 1 for a win, 0 for a loss, and 0.5 each for a tie (intentional or unintentional draw, as recorded in the data).

Each player also has a K factor: how many points of rating change one game can cause. After the game, ratings become:

K is the same for both players in a game and is fixed for the whole calculation (all-time vs a season slice each use their own constant K).

Numbers used on this site

• Starting rating for someone who has not appeared yet: 1000.
• K (all-time leaderboard and player history in “All time”): 32.
• Season views recompute Elo from only that season’s events with a higher default K (override with RIFTBOUND_SEASONAL_K or RIFTBOUND_SEASONAL_K_MULTIPLIER).
• The 400 in the expected-score formula is the usual Elo “scale width” (about 400 points ≈ 10× more likely to win).

Which games count, and in what order

Only completed, non-bye, two-player games from this dataset are used. Events are ordered by each event’s scheduled start time (earliest first). Inside an event, tournament rounds keep the order from the official bracket (round number / phase order). Your Δ Elo on player pages uses the same ordering as the leaderboard, so history matches the headline rating.

Data in this build

Ratings use a single frozen export of official event data. Here’s how many events it includes, when that file was last updated on this server, and which stores (venues) show up in it.

• Events: 72
• Data last updated: 2026-04-07 16:01 UTC (when this site’s data file was last written on the server.)

Stores

← Back to leaderboard