What's Up, Wratings?

1. Why don’t you use Classic Elo, Glicko, or WHR?

Wratings uses OpenSkill’s Plackett–Luce model and explicitly identifies disconnected groups of players as Ponds. This makes the limits of the available connection evidence visible instead of treating every displayed player as directly comparable. The site retains a simple Elo-style curve only as a 1D explanatory comparison in its PvP tools; it is not the Wratings rating engine.

2. Why am I not in Pond 1?

Ponds are disjoint groups of players defined by mathematical connectivity, not by skill. If you are not in Pond 1, neither you nor anyone connected to you through eligible recorded Sets is connected to the largest component in that selected calculation. A valid Set that links the two networks creates a Bridge.

3. Why isn’t my local tournament showing up in the calculations?

Wratings uses Event-level eligibility, including a minimum of 8 entrants by default. An eligible Set must also have a deterministic recorded result and fall inside the selected date and connection scope. Normal collection waits until a candidate is more than 24 hours past its recorded Start.gg tournament start, rather than treating a scheduled or still-settling bracket as final evidence. Recently incomplete published brackets are rechecked during the normal 14-day audit path.

4. Will you support games other than Super Smash Bros.?

The system currently supports Super Smash Bros. Ultimate and Super Smash Bros. Melee. Its durable data model is designed to support other competitive games when reliable source data and a suitable game-specific implementation are available.

5. Why do some players appear as (id: 123456) instead of a name?

These are UGTs (Unidentified Gaming Tags). The recorded Set supplies an entrant or player identity, but Wratings does not currently have a usable public gamer tag or profile for that identity.

This can happen when a bracket entrant is not linked to a public Start.gg account, when an account’s public information is unavailable, or for another source-data reason. Wratings does not assume which explanation applies to an individual ID.

Why are they on the leaderboard?

Removing an otherwise eligible identity would remove recorded opposition from the Set network and distort the calculation for other players. Wratings therefore preserves eligible recorded match evidence even when a display name is unavailable.

If you can help identify a public gamer tag, email thisiswhatsup@wratings.gg.

6. How do I interpret the leaderboard controls?

The controls have two different roles:

• Calculation controls (Game, Connection, and Time Window): Each combination selects a separately calculated leaderboard. Changing any of these choices changes the evidence and rating calculation.
• View filters (Country and Pond): These choices narrow the displayed players within the current calculation. They do not recalculate ratings. The Country filter uses recorded event-location metadata where available. It is not a filter for a player’s residence or nationality.

In short: changing a calculation control asks for a different rating calculation. Changing a view filter focuses the current one.

7. How do I read the three displays on the PvP Probabilities page?

The PvP dashboard shows three complementary views:

• 1. 1D Elo Expectation: A simple Elo-style curve based only on the difference between the two displayed Wratings. It is an explanatory comparison, not the Wratings rating engine.
• 2. 2D OpenSkill Projection: A model-based probability display using the selected players’ raw OpenSkill skill and uncertainty values within the selected calculation context.
• 3. Recorded Head-to-Head: Eligible recorded Set wins between the two selected players inside the selected scope.

8. How do I read the odds on the PvP Probabilities page?

Both American Odds and Decimal Odds are displayed side by side, for example +217 | 3.16x.

• American Odds (+ / −): Centered around a $100 reference. A minus indicates the model favorite and shows the stake needed to win $100. A plus indicates the model underdog and shows the profit from a $100 stake.
• Decimal Odds (x): The total model-implied return, including stake, for every $1 wagered.

These are mathematical conversions of the displayed probabilities, not betting advice or an offer to place a wager.

9. What do the Evolution Graph windows mean?

A graph point is the Wrating calculated for one player, one Game and Connection choice, and one stable selected evidence series on that snapshot date. Maximum Available follows the declared available-evidence start and grows with the actual calculation date. It is not an unbounded all-time claim. Standard choices use nominal months × 30 + 14 days for 1, 3, and 6 months (44, 104, and 194 days), or nominal years × 365 + 181 days (546 days for 1 year, 911 days for 2 years, and so on), only where that full span of data is available. They are separate calculations, not merely different views of one ranking.

Shorter windows react more quickly to recent eligible results. Longer windows preserve more evidence and generally change more gradually. The visible labels disclose the actual inclusive date range and output as-of date; an older output is not presented as if it were calculated today. If maximum exactly matches a standard total, one choice serves both roles. Older saved numeric windows may be unavailable: the site informs you and selects Maximum Available rather than silently reinterpreting the old choice.

The legacy public Evolution Graph history begins on 2026-09-12 and is retained separately. The buffered policy has its own stable Maximum Available and standard-day series; it begins with accepted buffered snapshots and lengthens as additional snapshots are recorded. Earlier legacy values are not relabeled or spliced into the new series.

10. What does the “Updated” timestamp mean? Is my tournament included?

Updated is the newest eligible input-load timestamp contributing to the selected leaderboard calculation. It is specific to that game, connection, and time window. It is not the time that the web page changed.

The timestamp alone cannot prove that a particular tournament is included. A tournament must satisfy the completed-results, Event-entrant, connection, and selected-window requirements before its eligible recorded Sets can affect a later calculation.

A recently completed event may therefore appear in a later refresh rather than immediately. If a published bracket remains incomplete, the normal audit path can revisit it during its recent audit period.

11. Why does a player’s graph sometimes show a sharp drop right before the line ends?

Wratings uses a rolling evidence window. When older Sets leave that window, the player’s remaining evidence can change sharply. Recalculation can then change both the conservative public score and whether the player has enough current evidence to appear in that selected calculation.

Wratings does not smooth, interpolate, or carry forward a prior value. Each visible point is the output of the relevant daily calculation.

12. I know I’m worse than another player, but my Wrating is higher than theirs. How is that possible?

Wratings is based on eligible recorded results, not a community consensus. A score reflects the selected Set network, results, and the model’s uncertainty at that time. It is not a complete statement about a player’s worth or future performance.

Two players can have different scores even if their win–loss totals look similar. They may have faced different connected opponents, played in different parts of the selected window, or have different amounts of reliable evidence. More uncertainty lowers the conservative public ordinal used for the Wrating.

Event size is an eligibility threshold, not a weighting system. Wratings does not give a Set extra rating weight merely because the Event had more entrants.