Inkwell & Algorithm Back to Matchweek 7 · After kick-off

The breakdown · Sunday 18 October, 14:00

Nothing to split Leeds and Man United

The model would have called Leeds win, 2–1, with a 39% chance (bookmakers 40%). Here is everything that went into it.

Home

Leeds

7th · 9 pts

Called after kick-off · confidence: too close to call

2–1

Home win

Away

Man United

13th · 6 pts

Us against the market

Our model

39%

28%

33%

Bookmakers (average, margin removed)

40%

26%

33%

Leeds winDrawMan United win

Most likely scores

1–1

13%

2–1

1–2

0–0

1–0

How we got to 39%

We start from an average home side against an average visitor, then add or take away for each factor. Hover a bar for the detail.

Pushes towards our call Pushes against it Starting point

Average home side v average visitor

40%

Man United's attack

−6

Man United's defence

Leeds' defence

−2

Leeds' attack

Our call: Leeds win

39%

0%30%60%

The reasons

1 · Chances created and allowed

Little between them on chances

Expected goals (xG) measures the quality of chances. Over their last 5 home games Leeds created 1.5 and allowed 1.2 a game. Over their last 5 away games Man United created 1.5 and allowed 1.2.

xG per game, last 5 at this venue

Leeds created (home)

1.5

Leeds allowed (home)

1.2

Man United created (away)

1.5

Man United allowed (away)

1.2

2 · Form

Similar recent form

Last five league games: Leeds took 6 points, Man United took 6.

Leeds

DDWDL

Man United

WDLDD

Last 5 league games, most recent on the right

3 · Head to head

Not much history to go on

All of the last 2 meetings were drawn or won by one goal. Past meetings count for little in the model.

Apr 2026Man United 1-2 Leeds
Jan 2026Leeds 1-1 Man United

What would change our mind

If Man United's defence were only average, this would swing to an away win (37%).

Verdict

Leeds 2–1 Man United, home win, too close to call