Non-Penalty xG vs Total xG: What to Compare
Total xG includes modelled penalty attempts, while npxG isolates their contribution. Learn when the comparison helps and why definitions can vary.
Why npxG and xG appear side by side
Total xG includes modelled estimates for the shots in a sample, including penalties when the data source classifies them that way. Non-penalty xG, or npxG, excludes the penalty estimate so a reader can inspect the contribution from open-play and other non-penalty chances separately. The exact calculation may vary by provider, so check its definitions and event treatment.
In a simplified teaching example, a team records 1.20 total xG, of which 0.75 comes from one penalty estimate; npxG is 0.45. These invented values demonstrate arithmetic and are not a standard value for every penalty or match. The model determines the estimate. Do not carry a constant from one source into another. The xG and xGA guide covers broader model limitations.
What question does each metric answer?
Total xG asks: how much modelled shot quality did the team accumulate across all recorded attempts? npxG asks: how much came from shots other than penalties? If a team earned several spot kicks, the difference helps keep that contribution distinct from chances created in open play.
Removing penalties does not make npxG automatically “better.” For the complete match profile, penalties are part of the events that occurred. If the question is about open-play chance creation, a separate view can help. For a particular game, also check whether a penalty was retaken, overturned, saved or recorded according to the data source and the bookmaker’s settlement rules.
Compare matching samples
Record total xG, npxG, penalty count and match count. Do not compare one team’s season total with another’s per-match average. Check date range, competition, provider, game states and treatment of extra time. Beside npxG, consider total xG and average xG per attempt, because the same sum can come from very different shot volumes.
Short samples are sensitive to one event. A single penalty can materially change the gap when only a few matches are included. Show both the values and the number of games, and avoid rounding before making the comparison.
Do not turn the split into a direct bet
npxG can describe open-play chance creation, but it does not guarantee the next match will have no penalty. Total xG is not a ready-made probability that a team will score in its next fixture either. Compare the metric with the current market price, lineup, opponent and exact market. Without a separately validated link to price, the honest conclusion is a description of past games rather than a betting recommendation.
Read next
Sportexa AI
Apply this to a real match
Choose a match in Sportexa AI and ask the assistant how the available sports data and current odds relate to the idea in this article.
Open the assistant