Understanding Football Stats: xG, Possession and What Really Matters
A few years ago, a friend of mine refused to accept that his team had been lucky. They'd won 1 goal to nil, the opponents had hit the post twice and missed a penalty, and he was convinced it was a tactical masterclass. Then someone showed him the expected goals. He went quiet for about a minute and then said, "Well, xG doesn't win you points."
He was right, technically. But he was also missing the point, and that's the problem a lot of fans have with football statistics. We either ignore them completely or use them like weapons in an argument. The truth is somewhere in between. Numbers can't tell you everything about a match. They can tell you a lot more than your gut sometimes wants to admit.
So here's football stats explained in plain language: what the main numbers mean, which ones are overrated and which ones actually help you understand the game.
Expected goals (xG) explained
Expected goals, usually shortened to xG, is the stat that changed how people talk about football. It's also the one that causes the most arguments.
The idea is simple. Every shot is given a value between 0 and 1, showing how likely it is to be scored. That value comes from looking at thousands of similar shots in the past. A tap in from two yards out might be worth 0.8, because shots like that go in around eight times out of ten. A speculative strike from 35 yards might be worth 0.02. A penalty sits at roughly three in four.
The models consider things like distance from goal, the angle, which body part was used, whether it was a header, the type of pass that created it, and sometimes where the defenders and goalkeeper were standing. Add up the value of every shot a team takes and you get its xG for the match.
Why does that matter? Because goals are rare and a bit random. A team can create loads of good chances and lose, or create almost nothing and win. Over one match, luck plays a big role. Over ten or twenty matches, xG tends to give a better picture of which teams are genuinely good than the results alone do.
A few things to keep in mind. xG doesn't measure finishing quality directly, so an elite striker will often score more than his xG suggests. It also misses chances that never became shots, like a cross that flashed across the box with nobody there. And different providers use slightly different models, so the numbers won't always match between websites.
If you want to see xG done well, The Analyst, which uses Opta data, publishes excellent explanations and match breakdowns. It's a good place to build an understanding of what the numbers can and can't tell you.
Possession and passing accuracy
Possession is the stat that shows up on every TV graphic, and it's probably the most misunderstood.
Having the ball more than your opponent sounds like a good thing, and often it is. The best teams of the modern era, like Guardiola's Manchester City, dominated possession for years. But possession on its own says nothing about what a team does with the ball.
Leicester City won the Premier League in 2015/16 with some of the lowest possession figures in the league. They were happy to let opponents have the ball, then hit them on the counter attack. Plenty of relegated teams, meanwhile, have enjoyed decent possession while barely creating anything.
Passing accuracy has a similar problem. A centre back who plays ten short sideways passes will have a lovely percentage. A creative midfielder who tries difficult forward passes will have a lower one. Who's more valuable? Usually the midfielder.
That's why many analysts now prefer stats like progressive passes, which count passes that move the ball significantly closer to the opponent's goal. It's a much better measure of whether a team is actually getting anywhere. You can find those numbers, and many more, on FBref's Premier League stats pages, which are free and incredibly detailed.
Shots, big chances and conversion rate
Shots are an easy stat to understand, but they can be misleading too. Twenty shots from outside the box might be worth less than three shots from inside the six yard area. That's exactly why xG was invented.
A more useful measure is shots on target, because it at least tells you the goalkeeper was tested. Even better is looking at where those shots came from. Teams that consistently get shots inside the penalty area, especially in the central zone, tend to score more over a season.
Then there are big chances. This is a stat recorded by data companies for situations where a player would reasonably be expected to score, such as a one on one with the goalkeeper or a close range effort with the goal gaping. Count how many big chances a team creates and how many it gives away, and you'll often have a decent idea of how a match really went.
Conversion rate is the percentage of shots that become goals. It's fun to look at, but it swings a lot over short periods. A striker who scores five from ten shots in September is very unlikely to keep that up. When a team's conversion rate is far above normal, it usually drops back later in the season. When it's far below, it usually recovers. Statisticians call this regression to the mean. Fans usually call it "our luck turning".
Defensive stats that matter
Defending is the hardest part of football to measure, and this is where raw numbers can really fool you.
Take tackles. You might assume a defender who makes lots of tackles is a great defender. Sometimes he is. But often he's making lots of tackles because he's in a team that gives the ball away constantly, or because he's out of position and has to rescue himself. The best defenders sometimes barely tackle at all because they've already stopped the danger by being in the right place. Paolo Maldini is often quoted as saying that if he had to make a tackle, he'd already made a mistake.
So which defensive stats actually help?
Expected goals against (xGA). The quality of chances a team allows. A side that concedes very little xGA over a season is genuinely hard to break down.
Shots and big chances conceded. Simple, but a very good guide to how well a team protects its goal.
Pressing measures. Stats like PPDA, which stands for passes allowed per defensive action, show how aggressively a team presses. Lower numbers mean more intense pressing.
High turnovers. How often a team wins the ball back close to the opponent's goal. These often lead directly to chances.
Clean sheets are worth watching too, but remember they depend a lot on the goalkeeper and a bit of luck. A team that keeps clean sheets while conceding plenty of good chances is probably living dangerously.
Using stats alongside the league table
This is where things get genuinely useful. The league table tells you what has happened. The stats help you guess what will happen next.
Here's how I tend to do it. First, I look at the table, usually the live standings on Valioliiga, which shows points, goals scored, goals conceded and goal difference at a glance. Then I compare what I see with the underlying numbers.
If a team is high in the table but its xG and xGA are fairly average, I get suspicious. It might be riding a hot streak in front of goal or a goalkeeper having the season of his life. If a team is low in the table but creating lots of good chances and conceding few, I'd expect it to climb.
This season gives a nice example of why you should look at both. After five games, Brighton had scored more goals than anyone in the league and had the best goal difference. Is that sustainable, or is it a brilliant month of finishing? The table alone can't tell you. The shot and chance data can give you a much better idea.
The official Premier League website also has a stats centre with club and player records, which is handy for checking basic numbers like goals, assists and clean sheets before digging deeper.
One warning. Early in the season, all these stats are based on tiny samples. Five matches of xG data is interesting. Fifteen is meaningful. Thirty is reliable. Don't build too much on September.
What really matters
If you only look at a few numbers, I'd choose these: xG and xGA for the big picture, big chances created and conceded for match quality, and goal difference in the table for results. Possession and passing accuracy are fine as background, but I wouldn't lean on them.
Most of all, use stats to ask better questions, not to end arguments. The numbers will tell you that a team was fortunate, or unlucky, or genuinely dominant. They won't tell you about the injury in the warm up, the red card that changed the game or the manager's brave substitution. Watching the match still matters. The stats just help you understand what you saw.
And the next time your friend tells you xG doesn't win you points, you can agree with him. Then gently point out that, over a season, it's usually a pretty good guide to who will.
