Using a subset of data from football-data.co.uk.
Create a lookup table with home & away ratings each team.
Seed this table allocating 500 for each team home and away.
Define the last row of training immediately before games we wish to predict.
Add a column showing ABS() difference between rating prediction & actual result.
Add a column indicating whether rating H, D or A pred is correct (1) or not (0)
Use Solver to minimise the total error of the two added columns allowing a change to ratings within the range 300 - 700
I would have attached my example Excel file with the set-up, but it looks as though that's not allowed or possible
A rather poor substitute then are tonight's predictions.
Cardiff
Burnley
4.2
3.4
1.9
-1
A
Hull
Plymouth
1.62
3.6
5.75
1
H
Millwall
Bristol City
2.75
3
2.75
0
D
Preston
Swansea
2.38
3.2
3.1
0
D
Blackpool
Peterboro
2.05
3.6
3.3
0
D
Bolton
Birmingham
3.9
3.25
1.95
0
D
Cambridge
Stevenage
3
3.1
2.38
1
H
Charlton
Barnsley
1.75
3.5
4.75
0
D
Huddersfield
Wrexham
2.55
3.1
2.75
-1
A
Lincoln
Crawley Town
1.71
3.4
4.75
0
D
Mansfield
Wigan
2.15
3.2
3.3
0
D
Northampton
Stockport
4.2
3.1
1.95
0
D
Rotherham
Leyton Orient
2.3
3.3
3.1
0
D
Shrewsbury
Bristol Rvs
1.91
3.25
3.9
0
D
Reading
Exeter
1.73
3.8
4.5
2
H
Wycombe
Burton
1.49
3.9
6
0
D
Accrington
Milton Keynes Dons
2.45
3.3
2.75
0
D
Bradford
Cheltenham
1.62
3.6
5.75
1
H
Bromley
Doncaster
3.5
3.3
2.05
-1
A
Carlisle
Walsall
3.3
3.4
2.1
-1
A
Colchester
Chesterfield
1.95
3.4
3.7
1
H
Crewe
Fleetwood Town
2.3
3.2
3.1
1
H
Grimsby
Tranmere
1.75
3.5
4.75
2
H
Morecambe
AFC Wimbledon
3.4
3.2
2.15
-1
A
Newport County
Gillingham
3.75
3.25
2
1
H
Notts County
Barrow
1.65
3.7
5
1
H
Port Vale
Harrogate
1.83
3.2
4.75
0
D
Swindon
Salford
2
3.25
3.8
0
D