Maker here. I built Turfin because every horse-racing tip site I found worked the same way: a black box hands you numbers, and you either trust it or you don't. Nothing was auditable.
So I made the opposite. Every runner is scored on six criteria and 20 bonuses, and the full breakdown is visible — you can disagree with a score and see exactly where. The analysis is locked at 10am, before the race, so it can never be quietly rewritten afterwards.
Technically it's a Python scoring engine over a database of 43,000 races, with probabilities calibrated against real outcomes.
Happy to answer anything about the model or the data.