League One Midweek Backlog: Playoff Race Heats Up

Bradford vs Plymouth

Model: leagueone_lr_v1_cal. Probabilities H/D/A = 45% / 17% / 38%.

Closest fixture of the night: 45%/17%/38%. Bradford 4th, Plymouth 7th — both chasing playoff clarity. The tight spread means the model won’t commit, which is honest: fixtures between two sides within six points with something to play for are genuinely hard to call.

Doncaster vs Lincoln

Model: leagueone_lr_v1_cal. Probabilities H/D/A = 26% / 20% / 54%.

Lincoln 54% away — our model’s strongest away call of the round. Lincoln have one foot on the step-up already. Doncaster are well clear of the drop but have nothing obvious to play for. The 26% home number looks generous.

Rotherham vs Luton

Model: leagueone_lr_v1_cal. Probabilities H/D/A = 27% / 21% / 52%.

Luton 52% on the road is the model’s clearest away call of the night. Rotherham are entrenched in the relegation zone and Luton’s recent form has been steady — the probabilities read like a team that should win, facing a team that’s out of escape options.

Stevenage vs Barnsley

Model: leagueone_lr_v1_cal. Probabilities H/D/A = 29% / 19% / 53%.

Model has Barnsley at 53% away, a rare sign of confidence in a Championship-bound side rather than the home team. Stevenage are 6th and need the points to stay there. The 19% draw reflects how many ways this can go when the fixture is eight months late and neither side has built up head-of-steam for it.

Stockport County vs Mansfield Town

Model: leagueone_lr_v1_cal. Probabilities H/D/A = 53% / 19% / 28%.

Stockport 53% at home — the strongest home-win probability in the catch-up slate. Playoff stakes for the hosts, nothing on the line for Mansfield beyond pride. A relatively safe pick if you trust home-team bias on a Tuesday.

Cardiff vs Port Vale

Model: leagueone_lr_v1_cal. Probabilities H/D/A = 69% / 20% / 11%.

Cardiff 69% at home, the most lopsided model call of the week. Port Vale are bottom. Cardiff are chasing automatic promotion. The only real question is whether the scoreline matches the expectation — the model thinks yes, via something comfortable.

About these predictions.

League One predictions come from leagueone_lr_v1_cal — a Platt-calibrated logistic regression with player-quality features, retrained on historical fixtures and evaluated by walk-forward log-loss. The model publishes outcome probabilities (H/D/A), not scorelines; the score predictions in the predictions tab are derived from team goals-per-game and concede rates, tilted by the model’s probabilities.