For years, the answer to "where can I find free xG?" was simple: FBref. That changed in January 2026, when Stats Perform (the company behind Opta) ended its data agreement with Sports Reference and FBref removed xG, xA, PSxG and its other advanced metrics. FBref still has deep historical basic stats, but the xG is gone.
So where do you find xG now? Here's a practical guide to the sources that still work in 2026, what each is good at, and how to combine them with the player stats you need for betting.
Disclosure: we build StatPair. StatPair is not an xG provider, so it isn't ranked as one below. We explain where it fits at the end.
TL;DR - where to find xG in 2026
| Site | Price | Best for |
|---|
| Understat | Free | Team and player xG for the big five leagues (plus Russia), with history back to 2014/15 |
| FotMob | Free (with paid extras) | Match xG and shot maps across many leagues, on mobile and web |
| SofaScore | Free (with paid extras) | Match xG during and after games, wide league coverage |
| WhoScored | Free | Player ratings and detailed match centres |
| Opta Analyst | Free | Articles and xG-based tables for major competitions |
| StatsBomb open data | Free | Event-level data for selected competitions, for model builders |
Skip to: which one should you use? →
Quick xG primer (skip if you know it)
Expected Goals (xG) is the probability that a given shot becomes a goal, based on thousands of similar historical shots. It's the foundation of most modern football models. Full xG guide →
Different providers use different xG models. The main inputs:
- Shot location (every provider uses this).
- Shot type (open play, set piece, penalty).
- Body part (foot, head).
- Pre-shot context (cross, through ball, etc.) - some providers, not all.
- Goalkeeper and defender positions - rare in public xG.
That's why two sites can show different xG for the same match. The numbers are usually close, but don't mix providers inside one model or spreadsheet. Pick one source and stay consistent.
What happened to FBref?
On 20 January 2026, Sports Reference announced that its advanced soccer data provider had terminated access and required the data to be deleted. FBref and Stathead removed xG, xA, PSxG, progressive actions, shot-creating actions and related metrics from men's and women's competitions.
What's still there: results, line-ups, goals, assists, cards, minutes and other basic historical stats for over a hundred competitions. FBref is still useful for historical research, just not for xG.
What this means for you: if your spreadsheet or model pulled xG from FBref, you need a new source, and you should rebuild your history from that one source rather than splicing two models together.
Understat - best free xG for the big five
Coverage: Premier League, La Liga, Bundesliga, Serie A, Ligue 1 and the Russian Premier League.
History: seasons back to 2014/15.
Price: free.
Strengths:
- Team and player xG, xA and non-penalty xG.
- Shot-by-shot match data and shot maps.
- Clean tables that are easy to work with.
Weaknesses:
- Only six leagues: no Championship, no lower divisions, no cups outside those leagues.
- No betting workflow.
Verdict: the default free choice if you bet on the big five. Bundesliga vs Premier League for betting →
FotMob - best free xG on mobile
Coverage: match xG and shot maps for many top leagues and competitions.
Price: free, with optional paid extras.
Strengths:
- xG, shot maps and player ratings in one app.
- Fast line-ups and team news, which are essential for player props.
- Good mobile experience.
Weaknesses:
- Built for following matches, not for season-long research or exporting data.
- No hit-rate filters or betting workflow.
Verdict: the quickest way to check a match's xG and confirm line-ups.
SofaScore - best for live xG
Coverage: a very wide range of leagues, with xG for many major competitions.
Price: free, with optional paid extras.
Strengths:
- Live match stats, momentum graphs and xG during games.
- Player heatmaps and ratings.
- Coverage of leagues many other sites ignore.
Weaknesses:
- Data depth varies by league.
- Not designed for historical research or modelling.
Verdict: best for in-play viewing. In-play betting strategy →
WhoScored - best for player ratings
Strengths: famous player ratings, detailed match centres, strengths and weaknesses profiles.
Weaknesses: xG isn't the main focus, and it isn't a betting tool.
Verdict: a useful supplement for player research. Player props analysis →
Opta Analyst - best for xG context and articles
Opta's free editorial site publishes xG-based analysis, tables and previews for major competitions. It's great for context and storylines, less useful if you want raw numbers for every team in every league.
StatsBomb open data - best for model builders
StatsBomb publishes free event-level data for a selection of competitions. It's excellent for learning to build models in Python or R, but it doesn't cover full current seasons of every league, so it won't replace a week-to-week source.
Honourable mentions
- Football-Data.co.uk - historical results and closing odds. No xG, but essential for testing models against real prices.
- Paid providers (Opta, StatsBomb, Wyscout) - professional-grade data, priced for clubs and media rather than individual bettors.
Which one should you use?
You bet on the big five leagues
→ Understat for team and player xG, plus FotMob for line-ups.
You bet on many leagues
→ FotMob or SofaScore for match xG across a wider range of competitions.
You build models in Python, R or Excel
→ Understat for consistent xG history, Football-Data.co.uk for odds, StatsBomb open data for learning. Build a Poisson + xG model in Excel →
You bet in-play
→ SofaScore for live stats and xG.
You bet player props (shots, fouls, cards, corners)
→ xG alone isn't enough. You need per-player match history for the exact stat you're betting on. That's where StatPair fits (see below).
Compare StatPair, FlashScore, SofaScore and FBref in detail →
Where StatPair fits
StatPair doesn't publish xG. It covers the stats that player-prop and team-stat markets are actually settled on: shots, shots on target, fouls committed, fouls won, tackles, cards, corners, passes, saves and offsides, for 35,000+ players across 40+ leagues, with home/away and half-by-half filters, referee stats, a Trends Finder for hit rates and AI projections for shots, fouls, cards and corners props.
A practical stack: Understat or FotMob for xG, StatPair for player and referee stats, and our free xG difference calculator to spot teams over- or underperforming their xG. Start a 7-day free trial →
Free xG sites tell you what happened on the pitch. They don't tell you whether your betting decisions are any good. Closing line value (CLV) does: if you regularly get better odds than the closing price, you're likely making good decisions, whatever your short-term results.
Read the CLV strategy guide → · Use our free CLV calculator →
Related guides: Team form analysis with xG and moving averages · Lower league betting: where the value is · Women's football betting statistics
FAQ - xG websites
Q: Does FBref still have xG?
A: No. FBref removed xG and its other advanced Opta metrics in January 2026 after Stats Perform ended its data agreement. It still has historical basic stats such as goals, assists, cards and minutes.
Q: What is the best free xG website in 2026?
A: For the big five European leagues, Understat. For wider league coverage and mobile use, FotMob or SofaScore.
Q: Why does xG vary between sites for the same match?
A: Each provider trains its model on different data and uses different shot features (some include defender or goalkeeper positions, some don't). Match totals are usually close, but individual shots can differ noticeably.
Q: Can I scrape free xG into my own model?
A: Check each site's Terms of Service first. Many allow personal use but not commercial use or heavy scraping.
Q: Should I use multiple xG providers?
A: Use one provider inside any single model so your numbers are consistent. A second provider is useful for sanity-checking outliers.
Q: Is xG predictive of future results?
A: Yes. Over a season, xG is a better guide to future results than actual goals, because it filters out finishing luck. In our own study of 372,650 player starts, goals were far less repeatable than shots, which is the same reason xG works. Full xG guide →
Q: When is xG NOT useful?
A: Very small samples (under 5 matches), early season after big squad changes, and matches distorted by red cards. 10-match rule explained →
Bottom line
xG is still one of the most useful stats in football, but in 2026 you have to know where to find it. Understat covers the big five, FotMob and SofaScore cover more leagues, and FBref is now a historical database rather than an xG source.
For player props, pair your xG source with per-player stats for the market you're betting on.
Try StatPair free for 7 days →