Understanding xG and xGI (Without the Maths Degree)
By Geomi George · Published 22 August 2026
If you've spent any time around Fantasy Premier League, you've seen the letters xG and xGI thrown around. They sound technical, but the idea behind them is simple — and once it clicks, you'll make better transfers because you'll spot returns before they arrive rather than after.
xG: expected goals
Expected goals (xG) measures the quality of the chances a player gets. Every shot is given a value between 0 and 1 based on how likely an average player would be to score it — a tap-in from six yards might be worth 0.7 xG, a speculative effort from 30 yards just 0.03.
Add up a player's shots over a run of games and you get their xG: the number of goals they "should" have scored given the chances they had. A striker with 4.0 xG has been getting genuinely good chances, whether or not the goals have gone in yet.
xA and xGI: adding assists
Expected assists (xA) does the same for chances created — the likelihood that each pass a player makes becomes an assist. xGI (expected goal involvements) simply combines the two: xG + xA. It's the single best snapshot of a player's total attacking threat, capturing both their goal chances and their creativity in one number.
Why it predicts points better than goals
Football is noisy. A striker can miss three great chances one week and score a scruffy deflection the next. If you buy players based purely on goals scored, you're often buying after a hot streak — right before it cools off. If you buy based on underlying numbers, you catch players whose process is strong before the goals catch up.
The core insight: over time, players tend to score roughly in line with their xG. So two situations are especially useful to spot:
- Goals below xG: a player creating good chances but not converting is often about to "regress" upward — a buy-low opportunity.
- Goals above xG: a player massively overperforming his chances may be riding luck that won't last — be wary of buying at the peak.
Our xG Regression tool is built around exactly this idea — it highlights players whose goals and xG have drifted apart, flagging potential bargains and potential traps.
How to actually use xGI
- Confirm form is real: a player in "good form" with strong xGI is a solid buy; one in good form with weak xGI may be about to cool off.
- Find differentials early: a low-owned player quietly racking up xGI is a differential before the crowd notices — see finding differentials.
- Compare two players: when choosing between similar options, the one with better underlying numbers is usually the safer long-term pick.
The limits of the numbers
xG isn't magic. A few sensible caveats:
- Sample size matters. Three games of xG is noise; ten games is a signal.
- Penalties inflate it. A penalty is worth around 0.75 xG — a regular taker's numbers are boosted by spot-kicks, which is fine, but know that's where it's coming from.
- Elite finishers can sustainably overperform. The very best strikers beat their xG year after year; it's not always luck.
- It says nothing about clean sheets or minutes. Use it for attacking returns, alongside your other checks.
Treat xGI as your early-warning system. It won't tell you the future, but it will consistently point you toward players whose points are coming — and away from those whose hot streak is running on borrowed time.
See who's over- and under-performing their expected numbers on the xG Regression page.
