Today
All articles

Big Chances Explained for Analysts: 3 Examples and a 4 Step SOP

6 Oct 2026·14 min read

Big chance football analytics title card

A big chance is a human-coded, binary match event that marks a clear opportunity where a player should reasonably be expected to score. For analysts, the practical guidance is simple: treat it as a qualitative companion to expected goals, never as a replacement, remember that penalties are always flagged, and read single-match counts with caution.

TL;DR:

  • Big chances are human-coded labels indicating situations where a player should reasonably be expected to score, often from close range or one-on-one with the goalkeeper.
  • The count of big chances can include events without shots, such as missed attempts or plays where the player chooses not to shoot, affecting total accuracy.
  • Big chances are binary and differ from continuous expected goals (xG), which assign a probability to each shot; both metrics complement each other for better analysis.
  • Discrepancies between providers are common due to subjective judgment factors like pressure assessment and path clarity, requiring manual audits for consistency.
  • Using big chance data reliably involves applying minimum-minutes filters, aggregating across seasons, and reporting uncertainties, rather than relying on small samples or single-match figures.
Betsyscore
Track Football Data In Real Time
Compare live scores, AI-powered predictions, momentum and detailed match statistics across more than 200 football competitions.
Explore BetsyScore

Table of Contents

What a big chance is and why the tag exists

Data providers built the big chance tag to capture something expected goals models could not express on their own: a plain-language judgment about whether a player really should have scored. The Opta/StatsPerform event definitions describe a big chance as a situation where a player should reasonably be expected to score, typically because the attempt comes from close range, through a one-on-one with the goalkeeper, or with a clear path to goal and low defensive pressure. Penalties are always classified as big chances under this framework, regardless of the taker’s record or the keeper’s form.

The tag caught on quickly with broadcasters, club analysts and fantasy football communities because it translates a statistical concept into something a casual viewer instantly understands. A 2.1 expected-goals rating means little to someone watching a Saturday afternoon match. “He should have scored that” needs no explanation.

Certain situations almost always trigger the label:

  • An unmarked striker through on goal with only the keeper to beat.
  • A tap-in from inside the six-yard box with no defender closing down.
  • A penalty kick, awarded automatically regardless of context.
  • A header from close range with a clear, unchallenged run to the ball.

The common thread across every qualifying scenario is proximity, space and time: the attacker has enough of all three that a reasonable observer expects a goal.

How providers code the label in practice

Coding a big chance is not limited to shots that actually happen. Providers can log the event even when a player fails to shoot at all, including an air shot that completely misses the ball or a decision to pass up a clear chance rather than attempt it. That distinction matters for anyone building a dataset from event feeds: a big chance count is not simply a subset of shot attempts.

Rebounds and deflections introduce their own judgment calls. A loose ball that falls to an attacker in space after a blocked shot can generate a fresh big chance if the follow-up meets the same criteria of proximity and low pressure, which means a single attacking sequence can produce more than one tagged event. Borderline situations, such as a one-on-one at a tight angle or a shot taken under light but present pressure, are where providers are most likely to diverge from one another.

Pro Tip: When pulling big chance data for a report, check whether the feed includes chances with no shot attached; ignoring them undercounts true scoring opportunities and skews conversion rates.

The technical relationship between big chances and xG

A big chance is binary: yes or no, scored or missed. Expected goals is continuous: a probability between 0 and 1 attached to every shot based on location, angle, assist type and other factors. They answer different questions. The binary flag asks whether a human coder judged the chance obviously clear. The continuous model asks how often a shot from that exact position, under those exact conditions, results in a goal across a large sample.

Binary big chance flag beside continuous xG scale

The two metrics work well together because they correct for each other’s blind spots. According to the Wharton Sports Analytics discussion on possession-aware chance creation, some predictive models fold the big-chance flag directly into their feature set alongside a shot’s raw xG value. The binary tag sharpens calibration at the extreme high-probability end, where xG models sometimes underweight truly open chances, while the continuous xG score preserves the variance that a yes-or-no label throws away.

A shot tagged as a big chance typically carries an xG value well above the 0.3-0.4 range, consistent with the close-range, low-pressure criteria that define the label, though the exact threshold is not fixed and varies by situation.

Possession-aware models, often described as xG+, push this further by crediting danger that never produces a shot at all:

  • A defense-splitting pass that forces a goalkeeper into a desperate save attempt before any shot is logged.
  • A sustained spell of pressure inside the box that collapses without a clean attempt.
  • A cutback that a teammate fails to reach, never becoming a recorded shot.

These possession-aware approaches reduce the double-counting problem that pure shot-summed xG can create during rebound scrambles, and they capture creators who generate danger without racking up shot volume.

Why the big chance tag depends on human judgment

Providers are explicit that this event relies on a coder’s real-time read of the pitch, not an automated calculation. The Opta/StatsPerform definitions describe big chances as a qualitative, human-coded marker, and they note plainly that discrepancies between providers are expected rather than exceptional.

Three factors tend to drive disagreement between coders and between providers:

  1. Pressure assessment. One coder may judge a shadowing defender as enough pressure to disqualify a chance; another may not.
  2. Path clarity. A slightly angled run can read as a clear path to one observer and a crowded lane to another.
  3. Shot selection. When a player chooses a low-percentage finish over an easier pass, some coders still log the chance as big; others downgrade it.

Pro Tip: Before comparing big chance totals across two providers or two leagues, pull a sample of matches from both and manually audit a handful of tagged events; a consistent gap of even one or two chances per match signals a coding-standard difference, not a real gap in team quality.

Because the label is a judgment call rather than a measurement, sensible practice is to treat any single provider’s count as an estimate with a margin around it, not a precise figure. Cross-checking totals against a second source, where available, is the simplest safeguard.

What the numbers can and cannot tell you

Big chance counts sit inside a larger statistical problem that affects almost every finishing metric in football: distinguishing a real, repeatable skill from a short run of good or bad fortune. Football is a low-scoring sport, and low scoring means high variance. A striker can miss three clear chances in a row purely by chance and look identical, statistically, to a striker in a genuine slump.

A 2026 variance-decomposition study from the University of Münster found that only a modest share of goal overperformance traces back to stable, repeatable skill rather than randomness: roughly 29% to 38% of team-level variation in finishing performance is attributable to skill, with player-level skill signals smaller still but detectable across larger samples. That leaves a substantial share of what looks like hot or cold finishing as noise that will regress given enough minutes.

The practical consequence for big chance analysis is straightforward: a striker converting 80% of big chances through six matches is not yet showing you a skill estimate, just an early-season sample that could easily include a hot run. The same study points toward multi-season aggregation and statistical shrinkage, pulling extreme early results back toward the league average as more data arrives, as the sounder way to estimate genuine ability.

A few safeguards reduce the risk of overreading short samples:

  • Apply a minimum-minutes filter before trusting any per-90 big chance or conversion rate.
  • Aggregate across multiple seasons rather than judging a player on one hot or cold run.
  • Use bootstrap resampling or similar techniques to attach an uncertainty range to any headline conversion number.
  • Separate penalties from open-play big chances, since penalties carry a different and generally higher conversion rate that can distort pure finishing analysis.

None of this means big chance data is useless. It means the number needs the same statistical humility that any small-sample football metric demands.

Reading big chances alongside xG: three working examples

Case 1: the striker with a high big-chance rate per 90. Before concluding a forward is a clinical finisher, check whether the rate holds across multiple seasons and whether the underlying xG per shot matches expectations for the chances being created. A spike built on five matches tells you far less than a pattern sustained across a full season or two, in line with the shrinkage approach the Münster variance study recommends.

Case 2: a team creating many big chances but converting few. This is the classic setup for a finishing-versus-luck question. Compare the team’s actual goals against its cumulative xG over a meaningful run of matches; a persistent gap across a full season is more informative than a gap measured over five or six games, which can just as easily reflect variance as a real finishing problem.

Case 3: a penalty-heavy profile. A player’s big chance total can look inflated by converted penalties, which are always flagged regardless of difficulty. Filtering penalties out before judging open-play creativity or finishing keeps the analysis honest.

A simple workflow ties these cases together:

  1. Check which provider generated the data and note its specific coding conventions.
  2. Filter or separately label penalties depending on the question being asked.
  3. Convert raw counts to a per-90 rate using a reasonable minimum-minutes threshold.
  4. Combine the big chance rate with xG per 90 and attach an uncertainty band rather than reporting a single clean number.

A big chance count without a sample-size check is a headline, not an analysis.

That single habit, pairing the binary tag with the continuous probability model and a stated confidence range, separates a useful read from a misleading one.

How BetsyScore applies big-chance and xG signals in live predictions

Our platform ingests event-level flags like big chances alongside live, continuously updated xG values and a minute-by-minute momentum read, since no single signal tells the full story of a match in progress. When a big chance is logged in a live feed, we treat it as a qualitative marker that complements the shot’s raw xG value, not a substitute for it, which keeps our win-probability outputs from overreacting to one isolated open chance.

Penalties receive a fixed, separate classification path in our predictions, since their always-flagged status and distinct conversion profile would otherwise distort a live win-probability swing if blended in with open-play chances. Coder uncertainty is handled the same way we treat any judgment-based input: as a contributing signal with bounded weight rather than a deterministic trigger for a prediction change.

A few practices from our approach translate directly to any analyst’s own workflow:

  • Weight a big chance by its underlying xG value rather than treating every flagged event as equal.
  • Keep penalty-driven swings in a separate bucket from open-play momentum shifts.
  • Recalculate win probability continuously rather than anchoring to a single pre-match number, since a 70th-minute big chance carries different weight than one in the 10th minute.

Readers who want to see this applied in real time can follow live match data or the underlying AI predictions that combine these signals into a single win-probability figure.

My own working rule: confirm the provider and its exact definition first, separate penalties from open-play chances, convert everything to a per-90 rate behind a minimum-minutes filter, and never quote a conversion percentage without naming the sample size behind it.

Four-step big chance analysis procedure

Three red flags tell me to distrust a big chance count on sight: a sample under roughly ten matches, a total that mixes penalties into open-play figures without labeling them, and a cross-provider comparison with no audit of whether both sources code the event the same way.

Report uncertainty rather than hiding it. A range built from a full season, or better, multiple seasons, says more about a player or team than a single tidy number ever will.

— Aria

FAQ

How are big chances calculated?

Big chances are not calculated through a formula; they are flagged by human coders watching a match, based on criteria like shot distance, defensive pressure and clarity of the path to goal, as set out in the Opta/StatsPerform event definitions. Penalties are automatically included regardless of the situation.

What counts as a big chance created?

A big chance created is logged when a player sets up a teammate, or themselves, in a situation where a goal should reasonably be expected, including cases where no shot is actually attempted. Coding standards vary slightly between data providers, so totals can differ for the same match.

What xG is considered a big chance?

There is no fixed xG cutoff that automatically defines a big chance, since the tag is a separate, human-coded judgment rather than an xG threshold. In practice, shots tagged as big chances tend to carry notably high xG values, consistent with the close-range, low-pressure conditions the label describes.

What defines a big chance?

A big chance is defined as a situation where a player should reasonably be expected to score, typically a close-range effort, a one-on-one with the goalkeeper, or an attempt with a clear path and minimal pressure, per the Opta/StatsPerform definitions. Penalties always qualify under this definition.

Sources

Big Chances Explained for Analysts: 3 Examples and a 4 Step SOP