Welcome to The Athletic’s Premier League’s predictions challenge, where the subscribers, like Arsenal, are inching closer.

It was a tense weekend for our latest guest subscriber, Aston Villa fan Jack, but Manchester City’s third goal against Brentford, in stoppage time, gave the readers a precious spot-on prediction to extend their lead at the top.

The subscribers went into last weekend just five points clear at the top of our table. That lead now stands at six points with only 21 games left to be played.

Each week since the season began in August, four of us — an algorithm, a guest subscriber on rotation, six-year-old Wilfred and I — have been predicting the Premier League results with varying degrees of success.

We are awarding three points for a correct scoreline and one point for a correct result. We are also awarding a bonus point for any “unique” correct prediction, so, for example, that 3-0 Manchester City win over Brentford earned the subscribers a precious four-pointer.

I felt I was building up a dramatic late challenge after I got four points for backing Brighton & Hove Albion to beat Wolverhampton Wanderers 3-0, again thanks to a stoppage-time goal.

Then on Sunday, I briefly had the prospect of another two spot-on predictions when Aston Villa and Everton were 2-1 up at Burnley and Crystal Palace respectively, but both of those games ended up 2-2 which, along with Nottingham Forest’s late equaliser against Newcastle United, pretty much killed off my fleeting remontada hopes.

Wilfred is still in serious contention, though. There were no correct scorelines for the boy wonder this week but a steady stream of correct results kept the pressure on the subscribers.

But the real star of this week was the algorithm, which won a remarkable — and record-breaking — 18 points.

Its season-long obsession with 1-1 draws was finally vindicated by Liverpool vs Chelsea on Saturday, Forest vs Newcastle on Sunday and Tottenham vs Leeds on Monday. All three of those yielded four points.

There were another five correct results, plus a bonus point for the draw between Sunderland and Manchester United. What a relief that it didn’t finish 1-1 at the Stadium of Light.

All of that leaves the subscribers with a slender lead over Wilfred at the top and the algorithm looking a little more respectable at the bottom. In fact, having dared to imagine I might make a late surge to the top, I’m now just hoping there isn’t another glut of 1-1 draws for the algorithm to dump me to the bottom of the table.

Anyway, there’s just one midweek match to preview, so Jack is staying on to give a prediction for Manchester City vs Crystal Palace. Good luck, Jack.

Manchester City v Crystal Palace

Oli says: “A few people in the comments section last week asked why there is often an assumption that Manchester City will breeze to victories while every Arsenal win is predicted to be tight and nervous.

“It’s a fair question. After all, City have only scored four more goals so far. This season and last have been their lowest-scoring seasons in a decade under Pep Guardiola but personally, I feel City have the greater capacity to win comfortably while Arsenal are better at winning tight games.

“I initially wondered about a big City victory here, given a) their need to attack a goal-difference deficit and b) Palace’s preoccupation with the Conference League final, but City play the FA Cup final on Saturday. They could win comfortably but perhaps not emphatically.”

Manchester City 3-1 Crystal Palace