Wednesday's card across 158 races gave the predictions a productive workout. The first pick won 35 times — a 22.2% strike rate that sits above the long-run average for the model, representing a solid day of directional accuracy when measured against the inherent unpredictability of greyhound racing. There are no certainties in this sport, but finding the winner in better than one race in five, consistently over large samples, is the foundation on which any edge is built.
The place count adds further texture. First picks placed — finishing first or second — in 65 races out of 158, a place rate of 41.1%. That is the metric that matters most for the model's underlying quality: a dog rated first that ends up placed consistently confirms the ratings are separating quality from the field even when winning margins are narrow or interference plays a role in the finish.
The broader picture across all three top predictions is similarly encouraging. The model's picks accounted for 93 winners from 158 races — a coverage rate of 58.9%. That means in nearly six races out of ten yesterday, at least one of the top three predictions crossed the line first. The any-top-3 place figure of 134 takes that coverage to 84.8% of races where a rated selection reached the frame.
These numbers don't translate automatically into profit — markets absorb model quality, and SP varies enormously dog to dog and race to race — but they confirm the platform is reading the form correctly and ordering the field in the right sequence more often than not. Wednesday was one of those days that quietly does the work. The full breakdown by track and grade is available on the historic results page.