Almost nothing in football repeats. Every open-play shot is a one-off: a different angle, a different defender, a different bounce. The penalty is the exception. It is the same ball, from the same spot, against the same lone obstacle, taken thousands of times a season across the world. That constancy is why it is the most studied shot in the game, and why we can say something unusually precise about it: it goes in roughly three times in four across the big leagues, more often than that in the one Premier League season we can count kick by kick (77 of 92 in 2025-26, measured below), and the reasons it ever misses are worth a whole article.
The Premier League penalty conversion rate, measured: 77 of 92 in 2025-26
We counted every Premier League penalty of 2025-26 and every scored penalty since 2003-04 in ESPN’s public match data. The short answer: in 2025-26, 77 of 92 penalties were scored, a conversion rate of 83.7%, and 15 were missed. One season is a small sample, so the 95% interval around that rate runs from 74.8% to 89.9%: above the three-in-four rule of thumb, but not by a margin a single season can prove.
The 92 kicks came in 89 team-matches (three sides took two in a game). 18 of the 20 clubs took at least one; Aston Villa and Tottenham Hotspur took none. Brentford took the most, 10, and scored 8. Eight clubs scored every penalty they took, Crystal Palace’s 8 of 8 the most without a miss; Liverpool scored 1 of 2, Brighton and Wolves 3 of 5 each.
How much of that spread is noise? Almost all of it. If every club converted at the league’s 83.7%, an exact binomial test asks how surprising each club’s count would be. No club falls outside the 95% range; the least likely, Brighton and Wolves, have p = 0.19. Across all 18 clubs the spread gives a chi-square of 14.38 on 17 degrees of freedom (p = 0.64), a little less spread than pure chance would typically give. On five to ten kicks a season a club’s conversion rate says almost nothing about its takers. How many penalties a club is awarded varies more: from 0 to 10, a variance of 6.94 against a mean of 4.6, a ratio of 1.51 where a Poisson would give 1 (chi-square 30.17 on 19 degrees of freedom, p = 0.0496). That is only just past the usual line, so it hints that winning penalties is more of a team trait than scoring them, without proving it.
| Club | Taken | Scored | Missed | Scored % | p |
|---|---|---|---|---|---|
| Brentford | 10 | 8 | 2 | 80% | 0.67 |
| Crystal Palace | 8 | 8 | 0 | 100% | 0.37 |
| Chelsea | 8 | 7 | 1 | 88% | 1.00 |
| Leeds United | 8 | 6 | 2 | 75% | 0.62 |
| Newcastle United | 6 | 6 | 0 | 100% | 0.60 |
| AFC Bournemouth | 6 | 5 | 1 | 83% | 1.00 |
| Manchester United | 6 | 4 | 2 | 67% | 0.25 |
| Fulham | 5 | 5 | 0 | 100% | 1.00 |
| Sunderland | 5 | 4 | 1 | 80% | 0.59 |
| Brighton & Hove Albion | 5 | 3 | 2 | 60% | 0.19 |
| Wolverhampton Wanderers | 5 | 3 | 2 | 60% | 0.19 |
| Arsenal | 4 | 4 | 0 | 100% | 1.00 |
| Manchester City | 4 | 3 | 1 | 75% | 0.51 |
| Nottingham Forest | 3 | 3 | 0 | 100% | 1.00 |
| West Ham United | 3 | 3 | 0 | 100% | 1.00 |
| Burnley | 2 | 2 | 0 | 100% | 1.00 |
| Everton | 2 | 2 | 0 | 100% | 1.00 |
| Liverpool | 2 | 1 | 1 | 50% | 0.30 |
| Aston Villa | 0 | 0 | 0 | – | – |
| Tottenham Hotspur | 0 | 0 | 0 | – | – |
| League | 92 | 77 | 15 | 83.7% | – |
Penalty goals across 23 seasons, and the VAR question. Before 2025-26 the archive records only the penalties that went in, not the misses, so the long series below counts penalty goals per match, not conversion. Video review arrived in the Premier League in 2019-20. In the 16 seasons before it (2003-04 to 2018-19) there were 1,149 penalty goals in 6,071 matches, 0.189 a match; in the 7 VAR seasons, 574 in 2,660, 0.216 a match. That is 14% more (rate ratio 1.14, 95% interval 1.03 to 1.26). Six of the seven VAR seasons sit above the pre-VAR average, and two sit above the highest pre-VAR season (2006-07, 0.230): 2020-21, with 102 penalty goals (0.268 a match), and 2023-24, with 96 (0.253).
Three cautions keep that from being a clean VAR effect. First, 2020-21 carries much of it: the season was played mostly behind closed doors and under a handball interpretation the league softened the following summer. Without it, the VAR seasons average 0.207 a match, a ratio of 1.09 whose interval (0.98 to 1.22) includes no change. Second, scoring rose too, from 2.67 goals a match to 2.86, so penalties’ share of all goals moved only from 7.1% to 7.5% (7.2% without 2020-21). Third, seasons swing on their own: before VAR the season totals averaged 71.8 with a standard deviation of 9.4, against 8.5 for pure Poisson chance, and they ran from 56 (2017-18, 0.147 a match) to 87 (2006-07). A single season of 69 or 96 says little; the seven-season average says a little more.
| Seasons | Matches | Penalty goals | Per match | Share of all goals | All goals per match |
|---|---|---|---|---|---|
| 2003-04 to 2018-19 (16, before VAR) | 6,071 | 1,149 | 0.189 | 7.1% | 2.67 |
| 2019-20 to 2025-26 (7, VAR) | 2,660 | 574 | 0.216 | 7.5% | 2.86 |
| VAR seasons without 2020-21 (6) | 2,280 | 472 | 0.207 | 7.2% | 2.89 |
Method and limits. Attempts come from ESPN’s per-team “penalty kick shots” box-score field for the 380 matches of 2025-26; goals from the penalty flag on each goal in ESPN’s goal list for the same matches. We did not use ESPN’s own “penalty kick goals” field: it totals 110 against 92 attempts, because in 32 team-matches it shows exactly twice the penalty goals the goal list flags. We cannot split the 15 misses into saves and shots off target, we do not know how ESPN counts a retaken kick, and shoot-out kicks are not league penalties at all (for those, see the shootout discount, every 2022 World Cup penalty). The season series assumes ESPN flagged penalty goals as completely in 2005 as in 2025; the flags start in April 2003, so 2002-03 (11 flagged) is left out, and 12 older matches without a usable goal list are dropped.
The cleanest chance in football
An expected-goals model assigns a probability to a shot from its features: distance, angle, body part, the positions of defenders and the keeper. For almost every shot those features vary. For a penalty they are fixed by the laws of the game: the ball sits on a marked spot a set distance from goal, struck without obstruction, with no outfield defender allowed between it and the line. That is why a penalty carries a near-constant value in every model (commonly quoted in the region of 0.76 to 0.79 xG) and why it converts at a remarkably stable rate. The widely cited finding, across the big European leagues and over long samples, is that penalties are scored something like three times in every four.
Treat the xG value and the three-in-four rate as well-established ranges rather than exact constants (the precise rate drifts a little by league, era, and how you count retakes and shootouts), but the shape of the answer is not in doubt. Our own count for the 2025-26 Premier League, 77 of 92 (83.7%), sits above three in four, with a 95% interval of 74.8% to 89.9%. The penalty is, by a distance, the highest-value standard chance in football. The interesting question is not why it is converted so often, but why a shot this favourable is still missed somewhere between one time in four and one time in six.
The geometry, and the time the ball takes to arrive
The answer starts with a race the taker is heavily favoured to win but does not always win cleanly. From twelve yards a struck ball reaches the goal in a fraction of a second: less time than a goalkeeper needs to see the ball, decide which way to go, and physically travel from a central position to the corner of a goal that is wider than he can cover by standing still. That asymmetry is the whole reason the penalty is a good chance: the keeper essentially cannot react to a well-placed shot, so he must anticipate, committing before or as the ball is struck.
The corollary defines the target. The corners of the goal (high and wide, where the keeper cannot reach even with a correct guess) are effectively unsaveable if the taker hits them. The geometry hands the penalty taker a region of the goal that, struck accurately, beats any goalkeeper. The catch is in the word accurately: aiming at the unsaveable corners brings the post and the bar into play, and that tension is the source of every miss.
Placement versus power
That tension is the central trade-off of penalty taking, and it is best described qualitatively because the exact numbers vary with sample and definition. At one extreme is the placed penalty: aimed close to a corner, prioritising a spot the keeper cannot reach. It maximises the reward of a correct strike but shrinks the margin for error: miss your spot by a little and the ball is wide or against the woodwork. At the other extreme is the powered penalty: hit hard and a touch more centrally, trading some of the keeper-proof placement for a larger target and less chance of missing the goal entirely, while giving the keeper a marginally better hope if he guesses right and stays big.
There is no single correct answer, which is exactly why takers differ. Some live in the corners and accept the occasional ball into the stand; others trust power and a slightly safer line. The one robust generalisation is the failure mode of each: the placed penalty’s enemy is the post and the missed target, the powered penalty’s enemy is a keeper who reads it and gets a strong hand to a more reachable ball. Both strategies are converted at high rates (that is what a conversion rate of three in four or better is made of), and both miss in their own characteristic way.
The goalkeeper’s dilemma and the line
The keeper’s side of the duel is a genuine decision problem, and the rules shape it tightly. Because he cannot reliably react to a well-struck shot, he is forced into anticipation: commit early to a side and you might reach a corner you could never have reached by waiting, but commit early and guess wrong and you have removed yourself from the shot the taker actually chose. Wait to read the ball and you keep your options but sacrifice the head start the geometry says you need. Some keepers try to delay the decision to the last possible instant; some try to influence the taker’s choice by their stance or by feinting; some simply pick a side and dive.
The laws constrain all of this with the line rule. The goalkeeper must stay on or in front of the goal line until the ball is kicked, and the modern requirement is that he keep at least part of one foot on or in line with the line at the moment of the kick: a slight relaxation introduced to stop keepers being penalised for a toe drifting a centimetre forward, while still forbidding the old trick of charging out to close the angle. He may move along the line and dance to distract, but he may not advance off it early. That single constraint is a large part of why the penalty stays a three-in-four chance rather than a coin flip: the keeper is pinned to his line until it is almost too late to matter.
VAR-era encroachment and retakes
Video review sharpened the enforcement of rules that referees had long policed loosely, and the penalty was an obvious target. Two infringements now get scrutinised frame by frame. The first is the keeper coming off his line before the ball is struck; the second is encroachment: attackers or defenders entering the penalty area before the kick to attack a possible rebound. In the VAR era these are flagged far more consistently than they once were.
The consequence is a small but real shift in the probabilities, and crucially it is asymmetric. If the keeper saves a penalty but was clearly off his line early, the kick can be retaken, handing the taker a fresh attempt at a roughly three-in-four chance. If an attacker encroaches and the penalty is scored, the goal can be disallowed and the kick retaken; if a defender encroaches and the penalty is missed or saved, it can be retaken in the attacker’s favour. Net, tighter enforcement of the line and of encroachment nudges value toward the taker on saves and toward the defence on scored kicks with attacking encroachment: a reshuffling at the margins of an already taker-friendly event. The base rate did not move much; the bookkeeping around the edges did. Our archive cannot test that for conversion, because it records misses only from 2025-26, but it can count the penalties that went in: penalty goals per Premier League match were 14% higher in the seven VAR seasons than in the sixteen before, with 2020-21 carrying much of the rise (see the measured section above).
Why pressure changes the picture, and a word on shootouts
Everything above describes the mechanics of the kick. It says nothing about the state of mind of the person taking it, and that is where the in-game penalty and the shootout penalty quietly diverge. A penalty in open play is one moment in a flowing match: take it, and the game continues whatever happens. A shootout penalty is isolated, sequential, and decisive: each kick taken in turn, in silence, with the tournament hanging on it, and often after 120 minutes of fatigue. The physical shot is identical; the context is not.
The honest analytical position is that pressure and fatigue plausibly depress conversion relative to the clean in-game baseline, and that shootouts have their own dynamics: order effects, the weight of going behind, the keeper’s growing information about a team’s habits. But the size of those effects is contested, the samples are smaller and noisier than league-play penalties, and it would be a fabrication to attach a precise penalty to “pressure.” So we will not. For one tournament’s measured gap between in-game and shoot-out kicks, with its small-sample caveats, see the shootout discount. The safe summary is the one the data supports: the penalty is the most constant shot in football, worth around three-quarters of a goal and converted about three times in four (77 of 92 in the 2025-26 Premier League), and the moment you take it out of the run of play and stand a tired player over it with everything on the line, you are no longer measuring quite the same thing. For how single events like a won penalty move a match, see how win probability models work; for the broader family of dead-ball chances around it, see set-piece xG explained.
Try it: shootout chances from a conversion rate
Set a per-kick conversion rate and the tool computes, exactly from the binomial, how a standard best-of-five shootout (plus sudden death) plays out between two equally-skilled sides. The surprise it surfaces: at a higher rate, more shootouts stay level after five kicks (because both sides keep scoring) so the real edge is out-converting the opponent, not the absolute rate.
This calculator needs JavaScript. The method: each of five kicks is an independent Bernoulli trial with probability p, so made kicks follow Binomial(5, p); compare two such sides and break any remaining tie 50/50 for sudden death.
Open this calculator on the tools page →
Reproduce it
Every figure in the Premier League section (added November 5, 2026) comes from our analysis script. It reads data_layer/epl_2025_26_boxscores.json (ESPN’s public match summaries for all 380 matches of 2025-26, retrieved 7 September 2026) and data_layer/epl_goal_times_2002_2027.json (ESPN’s public Premier League scoreboard with every goal’s minute and penalty flag, pulled 5 October 2026) through our archive loader, which drops the 357 duplicated 2009-10 rows. Attempts are the box-score penaltyKickShots field; scored penalties are the goals with pen set. The script runs the binomial and chi-square checks, writes every number to data/stats/the-penalty-conversion-rate-epl.json and draws the chart. No network, nothing hand-entered.
The data behind this piece
- ESPN Premier League goal archive, 2002-03 to 2026-27: every goal’s minute with penalty and own-goal flags, as served on this site (pulled 5 October 2026).
- ESPN Premier League box scores, 2025-26: 28 per-team statistics for all 380 matches, including penalty kicks taken (retrieved 7 September 2026).
- StatsBomb open data: shot-level event data, including penalties, for computing conversion rates and xG yourself.
- StatsBomb: methodology on how penalties are valued within an xG framework.
- FBref: penalty and shot data (via Opta) across many competitions, useful for comparing eras and leagues.
- Understat: season-level shot and xG data for the major European leagues.
Published by SoccerAnalytics from public data and the sources named in this piece. How this site is made.