What are expected points (xPts) in football

Expected points, or xPts, are the points a team would have earned if results followed the quality of chances. Each match gets a win, draw and loss probability from the xG of both sides. Then xPts is three times the win chance plus one times the draw chance. A team that creates a lot and concedes little collects a high xPts even after a 0–0.

How it’s built

Start with xG for both teams in one match. From those two numbers a model estimates how likely every scoreline was, and adds up the chances of a home win, a draw and an away win. Say it comes out at 55% win, 25% draw, 20% loss. Then xPts is 3 × 0.55 + 1 × 0.25, which is 1.9 points. The real result gave the team 3, 1 or 0. The 1.9 is what the chances were worth.

Sum this over a season and you get an “expected table”. Put it next to the real one and the gaps jump out.

How to read the gap

Points above xPts mean the team took more than its chances justified. Finishing ran hot, the keeper stood on his head, a couple of deflections went in. Points below xPts mean the opposite.

The point of the gap is sustainability. A team sitting eight points over its xPts after ten rounds is living on margins that usually shrink. A team eight under is playing better than its table suggests. Neither is a guarantee about the next match. It’s a signal about how solid the current position is.

Where xPts misleads

It inherits every flaw of xG. If the xG model misses something, such as a long-range specialist or a side that plays for a lead, xPts misses it too. Two providers give two different tables.

Match state is baked in. A team that goes 1–0 up in the tenth minute sits back and stops creating. Its xG stays low, and xPts calls that a poor performance when it was a plan.

Small samples are noise. After three matches the gap tells you almost nothing. It starts to mean something after ten or more, the same window as team form.

It doesn’t see who played. A red card, a missing striker or a rotated squad all change the match, but xPts only reads shots.

xPts next to xG

xG grades one shot. xPts rolls the whole match into the one unit the league table actually uses: points. That makes it the bridge between the chance-level numbers and the standings. It answers a different question than xG does: not “how good was the attack”, but “what would the table look like if luck evened out”.

How it enters a match analysis

Example of an analysis card. Figures are illustrative.

Sharkline doesn’t publish an expected table. The chance-quality data behind xPts comes from a sports data API, next to the other metrics in how to read football match stats. What you get is a read: the probability of each outcome, a confidence percentage and a rating of the data behind it.

The full chain is on how Sharkline works.

Frequently asked questions

What does xPts mean in football? Expected points: the points a team’s chances deserved in a match or a season. Calculated from the win and draw probabilities that xG implies, as 3 × win + 1 × draw.

How is xPts calculated? A model turns the xG of both teams into probabilities of a win, a draw and a loss. Then xPts = 3 × P(win) + 1 × P(draw). Summed over matches, it gives an expected league table.

Does a team above its xPts always fall back? Not always, but the gap usually narrows. Finishing runs above the norm for a while, then drifts back. A team with a real finisher can hold a small positive gap for years.

Is xPts the same as a prediction? No. It looks backwards at matches already played. It says what the chances were worth, not what happens next.

How many matches before xPts means anything? About ten. Below that, one lucky night moves the gap more than the team’s quality does.