Provisional rating
Provisional rating
Definition
A provisional rating is a numerical chess rating that is considered temporary or statistically uncertain because the player has not yet played enough rated games for the system to be confident in their true strength. Many systems signal this uncertainty with a suffix (for example, US Chess uses a “P” with a game count like 1520P12), or a question mark “?” on online platforms.
How it’s used in chess
Provisional ratings help organizers and pairing systems place newcomers and infrequent players roughly where they belong, while allowing rapid adjustment as more results are recorded. Typical uses include:
- Pairings: Swiss-system pairings may treat a provisional rating like any other numerical rating for pairing purposes, but arbiters often keep in mind that the number is less reliable.
- Sections and prizes: Some events require an established rating (i.e., not provisional) for certain class sections or prize eligibility to discourage exploitation and to keep sections fair.
- Online pools: Servers mark ratings as provisional (often with “?”) until enough games are played, and they adjust such ratings quickly after each result to converge on an accurate number.
System differences and calculation
Different organizations implement provisional status in different ways:
- US Chess (USCF): A player’s rating is provisional for their first several dozen rated games (commonly shown as something like 1475P18). The formula gives more weight to recent opponents’ ratings and allows larger swings, accelerating convergence. After enough games (historically 26), the rating becomes “established.”
- FIDE: FIDE does not label ratings as provisional. Players are unrated until they meet the criteria for an initial FIDE rating, after which the rating is treated as established. However, newly rated players may still experience larger changes due to higher K-factors at lower ratings or early-career stages.
- Online platforms: Systems based on Glicko or Glicko-2 track a “rating deviation” (RD), a measure of confidence. A rating may appear with a “?” while RD is high; as you play more games, RD shrinks, the “?” disappears, and rating changes become smaller.
In short: during the provisional phase, the system’s uncertainty is high, so each game can move your number a lot more than it will later.
Strategic and practical significance
- Volatility: Expect big jumps. A single upset win or loss can move a provisional rating by hundreds of points.
- Opponent assessment: A low-rated provisional opponent might actually be much stronger (or weaker) than their number suggests. Prepare seriously regardless of rating.
- Event planning: If you rely on a rating floor/ceiling for section entry, know that provisional status may affect eligibility or how a TD seeds you.
- Psychology: Don’t overreact to swings in your first events or sessions. The goal is to reach an accurate rating, not to protect a temporary number.
Examples
- US Chess tournament example: A player enters their first weekend Swiss as 1200P05. They score 4/5 against mostly 1400–1500 opponents. Because their rating is provisional, the post-event increase is large, reflecting the strong opposition faced. After a few more events, swings naturally get smaller as the rating stabilizes.
- Online example: A new blitz account starts at a default 1200? and wins the first three games against 1300–1500 opponents. The system quickly raises the rating (e.g., near 1400–1500) and reduces the uncertainty marker “?”. After 20–30 games, changes per game are much smaller.
-
Upset on the board (illustrative PGN): A provisional player can cause a large rating swing with an early tactical shot. For example:
Even a very short win like this, while not common in serious play, demonstrates how a single result can disproportionately influence a provisional rating.
Historical notes
When Elo ratings were adopted widely in the late 20th century, federations grappled with how to onboard new players fairly. US Chess explicitly marked early ratings as provisional, allowing rapid correction. FIDE historically set high entry thresholds for ratings (later progressively lowered), and today issues an initial rating once published-list criteria are met. Online servers accelerated this idea with statistical models that show uncertainty directly (via “?” or RD).
Tips for players in the provisional phase
- Play more rated games to reduce uncertainty quickly; consistency matters more than any single result.
- Study openings you actually reach; accuracy in the first 10–15 moves prevents quick collapses that skew early results.
- Focus on blunder reduction (checks, captures, threats on each move) to stabilize performance.
- Enter appropriate sections; if allowed, choose based on your training strength rather than a transient number.
Interesting facts and anecdotes
- Notation: US Chess appends “P” with a count of games (e.g., 1680P12), while many servers display a “?” until the uncertainty falls below a threshold.
- Biggest early swings: It’s normal for a brand-new account to jump 200–400 points in a handful of games online, then move only 5–15 points per game after the rating settles.
- TD practices: Some organizers restrict class prizes to established ratings to deter sandbagging—see Sandbagging.
- Related math: Provisional treatment is essentially about higher K-factors or higher rating deviation: the system is “willing to learn quickly.” See Elo rating and Glicko for the underlying models.