Celtic’s players will decide this title, not some box of microchips.

Two articles surfaced tonight which you would almost swear were designed to drive me mad. Both were about Celtic. Both were on a subject I absolutely loathe, because it is usually stupid. And both were extraordinarily negative.

Now, negative stories are fine when they are grounded in something real. We are not perfect this season. We deserve criticism. Plenty of it.

But this particular subject is where I draw the line.

I am going to assume both pieces are broadly about the same thing, although they mention different organisations. One refers to Opta. The other refers to a site called Football Rankings, which presumably uses statistical data to make predictions.

Both apparently have Celtic finishing third in the table. In at least one of the articles, there is talk of “supercomputers” making predictions, as if that alone gives the whole thing some aura of authority. It does not.

There is no magic here. There is no genius. Nothing mystical is going on. This is probability analysis. It takes previous results, runs them through a model and produces likely outcomes based on that data. In Celtic’s case, those previous league results against Hearts and the Ibrox club are obviously going to count against us.

We have not beaten either in the league this season.

Of course, we have beaten the Ibrox club in both cup competitions. But because those wins came outside the league, and in one case after extra time and in the other on penalties, they apparently do not count in this particular framing.

That is the problem with these things. They sound sophisticated, but they only reflect the assumptions and data behind them. Anyone can run a version of this with the right information. You take previous results, feed them into a model and ask what future results look most likely.

That is not prophecy.

It is statistical analysis wearing a cloak and a pointy hat.

From that, the Record builds an apocalyptic headline about Celtic already losing £40 million, as though the outcome is settled. It is not settled. It is a probability analysis.

That is what drives me nuts about all this “supercomputer predicts” stuff.

It is as lazy as bad journalism gets. The media loves it because it sounds authoritative and modern. It looks clean and technical. It looks as though someone has taken human bias out of the equation and replaced it with something cold, rational and scientific.

But football is not that tidy.

The media used to love the FiveThirtyEight site Nate Silver was involved in building which was all based on statistical probability. Every time it had Celtic projected to finish behind the Ibrox club, the papers could not wait to splash it around. It sounded impressive, even definitive. It sounded like the machine had spoken.

And how often did it actually get the final outcome right?

By my reckoning, maybe once in a decade. That was the Covid season, and you did not need a supercomputer to see where that was going once Celtic started playing the way we did.

Probability analysis is not useless. I am not saying that. It has its place. Bookmakers use it. Analysts use it. Clubs use it. It can offer a guide to likelihoods and trends. But it is not destiny, and it should never be treated as such.

If you had asked a betting model at the start of extra time at the weekend what the chances were of Celtic winning that match by four goals, the number would have been tiny. If you had put money on it, you would have been able to build an extension on your house with the winnings. Because of course, against the numbers, Celtic went and did it.

That is football. That is the point.

The supercomputer will not decide whether Celtic win this title. The players and the manager will. The performances after the split will.

Celtic will win this or not based on the way we handle pressure will. On how we convert chances. Injuries, confidence, momentum, refereeing decisions, tactical choices, individual moments of brilliance and individual moments of madness will all matter.

A model can try to measure some of that.

It cannot capture all of it.

And if you ask that same model again after this weekend’s fixtures, there is every chance it will give you a different answer. That alone should tell people how careful they ought to be before treating these predictions like verdicts.

This is one of the lowest-common-denominator types of football story. We see them all the time now. “Supercomputer predicts final table.” “The Supercomputer gives title verdict.” “Supercomputer says Celtic will finish third.”

The Supercomputer, by the way, gave us no chance our last two times at Ibrox. Know how I know that? Because those articles were doing the rounds then too.

This is clickbait dressed up as science, and outlets use it to chase the most dramatic headlines possible, just as The Record does tonight.

It tells supporters the fight is already lost. It tells them the money has already moved and now sits in the coffers at Ibrox. This is meant to make us think the table is already set. It tries to smother hope under a sheet of technical-looking nonsense.

But football lives on grass, not in a box filled with microchips.

This is not artificial intelligence. It is not some grand machine mind peering into the future. A model takes available data and produces a forecast. The next match can change that forecast. The week after can change it again.

Football has its own logic. It has its own variables. It has numbers you can measure and moments you cannot. Football has panic, belief, pressure, momentum and sheer daftness. It has players who miss sitters one week and score worldies the next, and it has teams who collapse when they look safe and teams who suddenly remember who they are when everyone has written them off.

No supercomputer captures that properly.

So yes, let the media have its headline. Let them tell us what the model says. Let them pretend this is something more than it is.

But do not ask Celtic supporters to treat it as gospel.

The title race is not finished because a website says so. Celtic have not left £40 million on the table because a model currently projects it. We will find out what Celtic are, and what this season becomes, on the pitch.

And if we do this, when the supercomputer finally decides it got this wrong and spits out a new set of data at those who have trumpeted this? Will they tell us? My guess is no. This is the last from the supercomputer that we’re going to hear, and it should be.