This week's Pundit Showdown: Gaffer Gary argues that Expected Goals (xG) is destroying the beautiful game. Stats Sarah disagrees. Obviously.

Gary's Case: "Football Was Fine Without Algebra"

Right. Expected Goals. xG. The single worst thing to happen to football since they stopped letting goalkeepers pick up back-passes.

Let me tell you what xG is. It's a number that tells you a goal should have been scored when it wasn't, or shouldn't have been scored when it was. In other words, it's a number that disagrees with what actually happened. On a football pitch. In real life.

Why would I want that? Why would anyone want that? A striker scores a 30-yard volley into the top corner and some lad on Twitter says "low xG, not sustainable." Not sustainable? It's in the net! It's the most sustained thing in football! It's a goal!

The problem with xG is it gives people who've never kicked a ball the confidence to tell people who have that they don't understand football. "Actually, your team overperformed their xG by 0.7 this season, suggesting regression." Regression? I'll show you regression. Watch a team managed by someone who learned football from a spreadsheet instead of a pitch.

In my day, we had a metric for whether a striker was good. It was called goals. He scores goals? Good striker. Doesn't score goals? Not a good striker. Simple. Beautiful. No algebra required.

Sarah's Response: "Gary, That's Not How Evidence Works"

I want to address Gary's points in order, because dismantling them efficiently brings me genuine professional satisfaction.

First: xG doesn't "disagree with what happened." It contextualises what happened. If a striker scores five goals from chances that historically convert at 2%, that's remarkable. It's also, statistically, unlikely to continue. That's not an opinion. That's probability. Probability predates football, Gary. It predates Stoke.

Second: the "lad on Twitter" argument. Yes, some people misuse xG. Some people also misuse hammers. We don't ban hammers. We teach people to use them properly. xG is a tool. It tells you the quality of chances a team creates and concedes. If you're conceding high-xG chances regularly, you have a defensive problem. If your striker is finishing low-xG chances at an elite rate, you have a very good striker. Either way, you have information. And information is better than vibes.

Third: Gary says "in his day" they just counted goals. Fine. But goals over a small sample size are enormously noisy. A team can score 5 goals in one match and create 0.8 xG in the next three. Goals tell you what happened. xG tells you what's likely to happen next. And if you're a manager, or a recruitment team, or a fan trying to understand whether your team's September form is real or a mirage, "what's likely to happen next" matters quite a lot.

Gary doesn't hate xG because it's wrong. He hates it because it occasionally proves him wrong. And that, for a man who's been confidently incorrect about football for 62 years, is a terrifying prospect.

Community vote: Who won this debate? Vote on our social channels. Results published Friday.