Official cards and live odds become pick, structure, or pass guidance—informed by 5.7M+ starter-level records spanning 1990–2026.
1990–2026 partial · multiple datasets across the US, UK/Ireland, Europe, Hong Kong, and Singapore.
Starter-level records analyzed across multiple research datasets. This is not a count of unique horses and these rows are not loaded into every live race read.
A full-year US validation set, not the entire GAMELIN research corpus.
Historical portfolio backtests. The +10.6% result belongs to the top-50% conviction subset across repeated races; it is not a single-race promise or a guarantee of future returns.