Stenhousemuir
36%Draw
34%Livingston
31%Most likely scorelines
1-115.2% · 6.60-014.2% · 7.01-011.9% · 8.40-110.7% · 9.32-07.4% · 13.62-17.3% · 13.71-26.7% · 14.90-26.2% · 16.2
Exact scores are low-probability by
nature — the likeliest scoreline here is 15%, and these
eight cover 80% of outcomes between them.
Most likely score by result
If Stenhousemuir 1-0 · 12% · 8.4If draw 1-1 · 15% · 6.6If Livingston 0-1 · 11% · 9.3
How this is modelled. Dixon-Coles Poisson with ridge-penalised
attack and defence ratings, exponential time decay and a low-score correction, fitted on
five seasons of Scottish league results, with a shots-on-target proxy blended into the
expected-goals estimate. Early in a season the fit pools all four Scottish divisions,
because promoted and relegated clubs have no record in their new division; it shifts to a
division-specific fit once the season has its own data. Backtesting on both match result
and exact scoreline showed that schedule beats either fixed choice.
These are model probabilities for information only, not betting advice. In walk-forward testing on a held-out season this model did not beat bookmaker prices — flat-stake betting on its disagreements lost money. It is published as a calibrated forecast, not an edge.