The prediction model processed 700 races over the past seven days, and the headline number is a first-pick win rate of 25.3%. That's comfortably above the baseline you'd expect from random selection in a six-dog field (16.7%) and sits right where the model has been performing over recent months. Consistency matters more than spikes, and this is consistent.
Place rates tell an even stronger story. The top pick placed in 47.4% of races, meaning nearly half the time the model's first choice finished first or second. Broaden it to any of the top three picks and the coverage is remarkable: 64% won at least one race, while 89.1% saw at least one top-three pick finish in the frame. If you're running combination or forecast strategies, those numbers give you genuine coverage.
Day by day, Friday was the busiest with 254 races and produced 61 first-pick winners (24%). Saturday brought 113 races with 33 winners (29.2%), the strongest single-day strike rate of the week. Sunday landed at 25.6% from 156 races. The quieter midweek days showed more variance, as you'd expect from smaller samples: Wednesday had just 12 races and only 2 winners, but that's too few to draw any conclusions from.
Across the three picks, the model found 177 first-pick wins, 148 second-pick wins, and 123 third-pick wins. That descending pattern is exactly what you want to see. It means the model is correctly ranking dogs in order of probability rather than just identifying a pool of contenders without meaningful separation between them.
The combined top-three win count of 448 from 700 races means the winner came from the model's top three nearly two-thirds of the time. For a sport with as much inherent randomness as greyhound racing, where one bump at the first bend can end a dog's chance, that's a strong signal that the underlying ratings are working. You can review every prediction against every result on the today's results page.