Skip to content Opens in a new tab
Football

Premier League Non-Penalty xG vs Goals, 2017/18 to 2024/25

Infographic by Game State (Chris Weatherspoon)

Game State scatter chart titled EPL 2017/18 to 2024/25, Non-Penalty Goals v Non-Penalty xG, plotting player seasons by non-penalty xG per 90 against non-penalty goals per 90 with a dashed diagonal line, over-performers such as Gareth Bale and Yoane Wissa in green and under-performers such as Gabriel Jesus in red
Infographic by Game State (Chris Weatherspoon). View the original.

A penalty is worth a fixed 0.79 xG in Opta's model and 0.76 in Wyscout's, reflecting the share of penalties that have historically been scored. Because a penalty is worth so much, analysts often use non-penalty xG (npxG), which leaves penalties out, to compare how good players are at getting chances in open play.

What is the xG of a penalty?

Every penalty gets the same fixed value in an expected goals model, because the situation is almost identical every time: the same distance, the same angle and only the goalkeeper to beat. Opta, whose data is behind the Game State chart on this page, gives every penalty 0.79 xG, based on the historical conversion rate of penalties [1]. Wyscout fixes its penalty value at 0.76 [2]. Other models use similar figures, generally somewhere between about 0.75 and 0.79.

A penalty’s value is simply the share of penalties that have been scored in the matches a provider’s model is built on, so the figure differs slightly from one provider to another.

Shot Typical xG What it means
Penalty (Opta) 0.79 About four in five are scored [1]
Penalty (Wyscout) 0.76 About three in four are scored [2]
A good chance inside the box 0.1 Scored about once in every ten attempts, Opta’s own example [1]

So a penalty is worth around eight times as much as a decent chance inside the box, which is why it can change a player’s or team’s xG for a match so much.

How often are penalties scored?

Roughly three times in four, and a little more in recent years. Across the big European leagues between 2008/09 and 2021/22, conversion rates ranged from 76.5% in Spain’s top division to 79.8% in the Dutch Eredivisie [3]. Earlier figures were lower: Bundesliga penalties were scored 74.25% of the time between 1963/64 and 2003/04, and in the 2005/06 Premier League season only 57 of 78 penalties went in, about 73% [3]. Shoot-outs are harder still. At World Cups, only 69.4% of shoot-out penalties have been scored [3].

That is why penalty xG sits just under 0.8. It is not that every penalty has a 79% chance; some takers are much better than others and some goalkeepers save far more. It is the average across thousands of penalties, and xG models deliberately do not adjust for who is taking the kick.

What is npxG?

npxG means non-penalty expected goals: a player’s or team’s xG with penalties taken out. You may also see npG (non-penalty goals) and npxG per 90, which is npxG divided by minutes played and multiplied by 90, so players with different amounts of game time can be compared fairly.

Removing penalties matters because they are worth so much. If a striker takes six penalties in a season, those alone add about 4.7 xG to their total on Opta’s figures (6 × 0.79). Two strikers who create exactly the same chances in open play can end up miles apart on xG just because one of them is the team’s penalty taker. Whether a team gets penalties also depends partly on luck, not only on how well it attacks. Taking penalties out gives you a cleaner view of how good a player is at getting into scoring positions, which is what most analysts and recruitment staff want to know.

Penalties do still count, of course. A team’s penalty taker adds real goals, and comparing a player’s penalty goals with their penalty xG is a fair way of judging how good they are from the spot.

How xG is calculated

An xG model looks at a very large number of past shots and works out how often shots of each kind were scored. Opta’s model takes account of things such as where the shot was taken from, the shot type (which foot, a header or a volley), the pattern of play, the type of pass that set it up, and where the goalkeeper was, including whether they were in the box and able to use their hands [1]. A shot gets a value between 0 and 1, and a player’s or team’s xG is the sum of all their shots.

Penalties are left out of the main model and given their fixed value, because nothing about a penalty changes from one to the next except the people involved [1]. If a penalty is saved and someone scores the rebound, the follow-up is a separate shot with its own xG. Wyscout notes that its model puts no cap on shots in the same move, so a penalty plus a rebound can add up to more than 1 xG [2].

Non-penalty xG vs goals in the Premier League

Comparing non-penalty goals with npxG shows who is finishing better or worse than the chances they get. The chart by Chris Weatherspoon of Game State [7] plots Premier League player seasons from 2017/18 to the early part of 2024/25 this way, using Opta data via FBref, with players above the diagonal line scoring more than their npxG and players below it scoring fewer.

A few examples stand out. Gareth Bale’s loan season at Tottenham in 2020/21 is one of the biggest over-performances: he scored 11 Premier League goals in 20 appearances [4], which the chart puts at more than one non-penalty goal per 90 minutes and far above his npxG. Erling Haaland’s seasons sit at the far right of the chart because nobody gets more or better chances. In 2022/23, when he set the Premier League record of 36 goals in a season [5], the chart shows he also scored more than his npxG. At the other end, Gabriel Jesus in 2018/19 had a high npxG per 90 but scored far fewer than expected.

Two things are worth bearing in mind when you read this kind of comparison:

  • Sample size. The chart’s qualifying bar for 2024/25 was just 300 minutes and 10 shots, against 900 minutes and 30 shots for earlier seasons, so the 2024/25 dots (Yoane Wissa and Liam Delap among the over-performers) rest on a few weeks of games. Wissa did go on to have his best season, with 20 goals in 39 appearances for Brentford in 2024/25 [6].
  • Luck and skill. Over-performing xG can mean excellent finishing, but over a single season it can also be partly luck, so a player who beats their xG year after year is better evidence of good finishing than one who does it once. Opta also points out that xG measures the quality of chances, not the result: goals still win matches [1].

For more ways of comparing attackers across leagues, see our page on football player statistics normalised across top leagues.

Questions about penalty xG

How much xG is a penalty?

0.79 xG in Opta’s model [1] and 0.76 in Wyscout’s [2]. Most providers use a fixed value between about 0.75 and 0.79.

Why is a penalty 0.79 xG?

Because historically about 79% of penalties in Opta’s data have been scored. Penalties are given a constant value that reflects that conversion rate [1].

What does npxG mean?

Non-penalty expected goals: xG with penalties removed. It shows how good a player is at getting chances in open play and from set pieces, without the boost from taking penalties.

What does 0.3 xG mean?

That a shot of that kind would be scored about 30% of the time, or roughly three times in ten. It describes the chance, not what happened to that particular shot.

What percentage of penalty kicks are scored?

Around 75% to 80% in top European leagues. Between 2008/09 and 2021/22, rates ranged from 76.5% in Spain to 79.8% in the Netherlands [3].

Does a rebound from a saved penalty count in xG?

Yes. The rebound is a separate shot with its own xG, added on top of the penalty’s value [2].

Sources

  1. Opta Analyst, What Are Expected Goals (xG)?
  2. Wyscout, Glossary: xG.
  3. Wikipedia, Penalty kick (association football), statistics section.
  4. Wikipedia, Gareth Bale, career statistics.
  5. Wikipedia, Erling Haaland.
  6. Wikipedia, Yoane Wissa.
  7. Chris Weatherspoon, Game State. The maker of the chart.

Last reviewed 8 November 2026

Add as a preferred source on Google

Embed this infographic

Use this code only if the creator has permitted your intended use. Providing embed code does not grant a licence. Keep the creator credit and the link to the text version and sources. How you can use these infographics.

Made this infographic, or spotted a mistake? Tell us and we will put it right.

About this page

You can rely on this page: reviewed 8 November 2026, with the figures checked against their sources and every source listed.

How we review infographics
Creator

Game State

3 infographics here

Game State is a Substack publication by Chris Weatherspoon about the money, numbers and data behind sport.

See all by Game StateVisit gamestate.substack.com
Topic

Football

More in this topic
For creators

Get your infographic listed

A permanent page like this one for your graphic, with your name and a link to your site. $29, paid once.

Submit an infographic

Made an infographic?

Give it a permanent page where people looking for your subject find it, with your name and a link to your site. Content marketers, agencies, designers and researchers all list here.