Knowledge Base
What is npxG (Non-Penalty xG)?
A single penalty can add 0.76 to a striker's xG tally and make a mediocre performance look elite. Non-penalty expected goals strips that distortion away - and what you find underneath is often a very different story.
The Penalty Problem That Nobody Talks About
Expected goals is the best tool football analysis has for separating good performances from lucky results. But it has a flaw - one that's significant enough to distort player rankings, mislead transfer departments, and misrepresent entire league seasons if you're not careful. The problem is the penalty kick.
A penalty has an xG value of approximately 0.76. That's how often they go in historically, across thousands of penalties across professional football. Which is correct - penalties are converted about 76% of the time. So far, so logical.
The issue is what that 0.76 represents as a slice of a player's total xG. In an average Premier League match, a team generates roughly 1.5 xG from open play. A single penalty adds half of that in one moment, from a situation that tells you almost nothing about a player's attacking quality. The penalty was won - which matters - but the shot itself is a standardised, uncontested kick from 12 yards. It carries no information about movement off the ball, link-up play, finishing technique under pressure, or anything that characterises a striker's open-play contribution.
For teams that win many penalties - or for players who are designated penalty takers - this distortion compounds across an entire season. By the time you're looking at season totals, a prolific penalty winner can have inflated their xG by 3-5 goals worth of expected output that came from the spot, not from the kind of open-play creation that wins you games across a full season.
What npxG Is - and What It Strips Out
Non-penalty expected goals (npxG) is the simplest possible correction to this problem. Take the total xG figure. Remove every penalty. What remains is npxG - a measure of expected goals from open play and non-penalty set pieces only.
That's it. No complex weighting, no black-box adjustment. Just xG minus the contribution of penalty kicks.
The result is a cleaner signal. When you look at a player's npxG, you're seeing the chance quality they've generated from situations that actually reflect their attacking ability: arriving at the right time in the penalty area, making the right run to receive a cut-back, positioning for a headed chance from a cross, capitalising on a defensive error in transition. The calculated, uncontested penalty kick is not in the number.
For teams, the logic is identical. A club's team npxG tells you how well they're creating chances from open play - from their attacking system, their movement, their build-up patterns. Penalties are largely random events in terms of when they're awarded (dependent heavily on referee decisions, opponent fouls, and positional play in a way that doesn't repeat consistently). Stripping them out gives you a cleaner measure of sustainable attacking quality.
Most data providers - StatsBomb, Opta, FBref - now publish npxG alongside standard xG as a matter of course. It has become the default metric of choice for serious attacking analysis.
What About Penalties Won vs. Penalty Kicks Taken?
There's a subtle distinction worth making. npxG removes the xG contribution of the penalty kick itself - the 0.76 expected goal for stepping up and shooting. It doesn't say anything about the value of winning the penalty in the first place.
Some analysts argue that winning a penalty is itself a skill - forwards like Harry Kane and Erling Haaland draw penalties at very different rates through their movement and positioning. That's legitimate. But the correct way to account for it is separately, not by including penalty xG in a metric meant to assess open-play attacking quality. The penalty won tells you about physical presence and intelligent movement in the box. The penalty conversion rate tells you about composure under pressure. Neither of those things belongs in a metric designed to assess open-play chance quality.
MatchAnalyzr Tip
The xG Trend widget in MatchAnalyzr shows your team's expected goals across recent matchdays - and you can toggle between total xG and npxG to see exactly how much of their attacking output is coming from open play versus the penalty spot. A rising npxG trend is a much healthier signal than a total xG spike driven by two converted penalty awards.
Lewandowski vs. Haaland: A Study in How npxG Reframes Narratives
The clearest way to understand why npxG matters is to look at how it changes the story for specific players - particularly those who have been prolific penalty takers at various points in their careers.
Robert Lewandowski, during his peak years at Bayern Munich, was one of the most prolific xG generators in European football. He was also among the most frequent penalty takers, stepping up regularly for a team that drew fouls at a high rate. In multiple seasons, the gap between his total xG and his npxG was meaningful - sometimes representing two or three goals worth of expected output that came from the spot. This doesn't diminish his quality; Lewandowski was genuinely elite in open play. But it did mean that naive xG comparisons slightly overstated his output relative to forwards who generated fewer penalties.
Erling Haaland presents an interesting contrast. His aerial ability and explosive movement create a particular type of threat that generates high-quality chances from open play - headers in central positions, one-on-one situations with goalkeepers after breaking defensive lines. His penalty frequency has fluctuated by club and season. In periods where his penalty count was lower relative to his total output, his npxG and total xG converged, suggesting that his headline numbers genuinely reflected open-play dominance rather than spot-kick accumulation.
The comparison is instructive: two elite strikers, but the relative gap between their npxG and xG tells you something important about the nature of their contributions. Neither is necessarily 'better' - but they're different, and npxG helps you see that difference clearly.
Teams That Live on Penalties - and Teams That Don't
The distortion isn't just a player-level phenomenon. At team level, the gap between xG and npxG can reveal something important about how a club creates its chances.
Some teams win penalties at a very high rate - through aggressive forward movement, physical forwards who draw fouls, or simply through tactical positioning that gets them into the penalty area frequently. For these teams, a high total xG can partly reflect a penalty dependency that's both strategically deliberate and potentially fragile. Referee decisions vary. Teams adjust tactically to stop conceding fouls. A high-penalty team that faces a more physical defence, or draws a referee who's reluctant to give spot kicks, may see their attacking output drop significantly even when their open-play quality hasn't changed.
By contrast, a team whose xG and npxG are nearly identical is generating almost all of their threat from open play. Their attacking performance is more clearly the product of system, movement, and technical quality - it's more robust to the randomness of penalty decisions and more clearly tied to things the coaching staff can repeat and improve.
This distinction matters enormously for pre-season recruitment. A team that wants to improve its attacking output should look at its npxG gap before deciding whether to sign a target man who draws fouls or a creative midfielder who unlocks defences. The answer depends on what the data actually shows about the source of their current attacking output.
npxG per 90: The Gold Standard for Comparing Strikers
Once you've established the value of npxG as a concept, the next step is to compare players across different clubs, leagues, and amounts of playing time. For that, you need npxG per 90 minutes - the gold standard for striker comparison in modern football analytics.
Per-90 normalisation solves the playing time problem. A striker who generates 12 npxG in 2,700 minutes (30 full games) is performing at 0.40 npxG per 90. A striker generating 8 npxG in 1,350 minutes (15 games) is performing at 0.53 npxG per 90 - a meaningfully higher rate, despite the lower absolute total. Without per-90 normalisation, you'd incorrectly rank the first player as superior based purely on volume.
For strikers specifically, the npxG per 90 threshold that separates elite from excellent is roughly:
- **Above 0.55 npxG/90** - genuinely elite open-play chance generation; top-five striker in Europe
- **0.40-0.55 npxG/90** - strong; quality starter at a Champions League club
- **0.25-0.40 npxG/90** - solid Premier League / top-flight level
- **Below 0.25 npxG/90** - may struggle to maintain a starting place at elite level
These bands shift depending on league and role - a deep-lying forward or a second striker will typically show lower npxG/90 than a true number nine - but as a rough orientation they're useful. The critical thing is always comparing within similar contexts: same league tier, same playing position, similar team strength.
MatchAnalyzr Tip
When tracking strikers on your watchlist in MatchAnalyzr, the player stats widgets display npxG per 90 alongside total xG - so you can immediately see how much of a striker's output is penalty-derived. For pre-season comparison across leagues, npxG per 90 is the number to sort by first.
League-Level Analysis: Which Leagues Award More Penalties?
The penalty frequency isn't consistent across European leagues, and this has a direct impact on how much npxG and total xG diverge at league level.
Historically, the Serie A has been among the higher-penalty European leagues - a combination of attacking movement into the box, defensive fouls at a high rate, and refereeing styles that have at various points been generous with spot kicks. La Liga, particularly in eras dominated by Cristiano Ronaldo and Lionel Messi (both frequent penalty takers), has seen significant penalty-derived xG in its headline team numbers.
The Bundesliga and the Premier League tend to fall somewhere in the middle, though individual seasons can skew significantly based on league-wide refereeing directives around handball rules, which periodically generate unusual penalty spikes.
Why does this matter for cross-league comparison? If you're comparing a Bundesliga striker to a Serie A striker using total xG, you need to account for the fact that the Serie A player may be operating in a higher-penalty environment - even if their open-play quality is comparable. npxG per 90 normalises this away, giving you a fairer cross-league comparison that's less dependent on the penalty culture of the particular competition.
This is an underappreciated benefit of npxG in transfer scouting - especially when clubs are evaluating players across multiple leagues simultaneously.
How npxG Changes Transfer Valuations and Scouting Decisions
The practical stakes of npxG are highest in the transfer market. Recruitment departments at elite clubs now routinely use npxG as a primary metric for striker evaluation - specifically because it identifies players whose goal-scoring quality is genuinely underpinned by open-play threat rather than penalty accumulation.
Consider the transfer calculus: a striker with 20 goals in a season looks spectacular. But if eight of those were penalties, you need to ask two questions. First: will they continue winning penalties at that rate in a new team - different system, different style, different forward partners? Possibly not. Second: when they're not converting penalties, how dangerous are they? The npxG number tells you the answer to the second question directly.
The flip side is equally important. A striker who consistently scores more goals than their npxG suggests - outperforming their expected output from open play - may be an elite finisher, or may be on a finishing run that's unlikely to sustain. npxG gives you the expectation; actual goals give you the outcome. The gap between the two is finishing quality - and sustained outperformance of npxG over multiple seasons is one of the clearest signals of a genuinely clinical striker.
Clubs like Brentford, known for data-driven recruitment, have made this kind of npxG analysis central to their scouting. Rather than chasing goal tallies, they look for players who consistently generate high-quality chances from open play in their npxG numbers - and then evaluate finishing separately. It's a more stable, more predictive approach to identifying attacking talent.
MatchAnalyzr Tip
MatchAnalyzr's xG Table widget lets you view league standings ranked by npxG - revealing which clubs are genuinely creating the most open-play danger and which are riding penalty decisions. Compare the npxG table to the actual standings: clubs sitting well above their npxG rank are likely candidates for a second-half regression.
npxG in the xG Table: How League Standings Change
One of the most revealing applications of npxG is recalculating the expected goals league table with penalties removed. The xG table already tends to tell a different story than the actual standings - it shows which clubs are genuinely performing well versus over- or underperforming their underlying quality. The npxG table takes this further.
In a typical Premier League season, swapping total xG for npxG in the table calculation can shift several clubs by a position or two. Teams that have drawn many penalties - particularly those with physical strikers or aggressive box-entry patterns - will drop in the npxG table relative to their total xG ranking. Teams that create most of their threat from open play will remain roughly where they are.
For clubs near the relegation zone, this distinction can be particularly meaningful. A team that looks marginally safer than the drop based on total xG might actually be sitting right in the danger zone when penalties are removed - their 'safety cushion' was partly built on spot kicks that may not recur. Conversely, a team generating good open-play npxG but struggling to win penalties might be better placed than their total xG ranking suggests.
For the title race, the npxG table often gives the most reliable signal of which teams are genuinely creating the most dangerous football - irrespective of referee decisions.
xG or npxG: Which Should You Use?
The answer, as with most good questions in football analytics, is: it depends on what you're trying to understand.
**Use total xG when:**
- You want to assess overall match performance - because penalties are real events that affect results
- You're evaluating a team's total threat in a specific game
- You're looking at short-term result prediction, where penalties are a legitimate source of goals
- You want to understand why a scoreline was what it was
**Use npxG when:**
- You're comparing player quality across different roles, clubs, or leagues
- You're assessing a team's underlying attacking system and open-play quality
- You're doing medium-to-long-term performance assessment where penalty-winning rates may not persist
- You're scouting players for transfer and want a stable signal of open-play attacking contribution
- You're building a league table of underlying performance rather than accounting for all goal sources
The nuanced answer is that sophisticated analysis uses both. Total xG tells you about results and match-level performance. npxG tells you about sustainable quality and fair player comparison. Neither is 'right' in the abstract - they're measuring slightly different things, and both are useful depending on the question you're asking.
For most practical scouting and player comparison purposes, npxG per 90 is the metric that experienced analysts reach for first. It's the number that best answers the question: how dangerous is this player - or team - when the game is actually being played?
Frequently Asked Questions
- What does npxG stand for?
- npxG stands for non-penalty expected goals. It is calculated by taking a player's or team's total expected goals (xG) figure and subtracting the xG contribution of any penalty kicks. In practice, each penalty carries an xG value of approximately 0.76, so npxG = total xG minus (0.76 × number of penalties taken). The result is a measure of expected goals from open play and non-penalty set pieces only.
- Why is npxG considered more accurate than xG for player comparison?
- Standard xG includes penalty kicks, which have a fixed xG value of around 0.76 regardless of who takes them or how well they're playing. This means a striker who takes many penalties will have an inflated total xG that doesn't reflect their open-play quality. npxG removes this distortion, making it a fairer comparison metric across players who take different numbers of penalties. It's particularly important when comparing players across different clubs or leagues where penalty frequencies vary significantly.
- What is a good npxG per 90 for a striker?
- In the Premier League and other top European leagues, elite strikers typically generate above 0.45-0.55 npxG per 90 minutes from open play. A number above 0.55 per 90 places a player among the top strikers in Europe. Figures between 0.25 and 0.40 are solid for a regular top-flight starter. Below 0.20 per 90 over a sustained period usually indicates that a striker is either playing in a limited role or struggling to get into quality positions. Always compare within the same league tier and playing position for meaningful context.
- Does a high gap between xG and npxG mean a player is penalty-dependent?
- A large gap between xG and npxG certainly means that penalties represent a significant portion of a player's expected goal output - but 'penalty-dependent' needs careful interpretation. Winning penalties can itself be a repeatable skill, reflecting good movement in the box and the ability to draw fouls. The question is whether the open-play npxG figure is strong enough to justify the player's role even without the penalty contribution. A striker with elite npxG who also wins many penalties is doubly valuable. A striker whose value primarily comes from penalty conversion is more fragile, especially if they change clubs or face opponents who adapt tactically.
- Do all data providers calculate npxG the same way?
- The core calculation is the same across providers - total xG minus penalty xG - but the underlying xG models vary. StatsBomb, Opta, and other providers use different input variables and historical datasets to calculate the xG value of each shot, which means the npxG figures across providers won't be identical. For this reason, you should always compare npxG numbers within the same data source rather than mixing figures from different providers. The relative rankings between players will generally be consistent even when absolute numbers differ slightly.
