🐎 Trotting — analysis and ranking
The model compares estimated market probability from win odds with the backing percentage in the pool and ranks over- and under-backed horses in every race.
ℹ️ How the model works
For every race we convert the winning odds into an estimated market probability (1 divided by the odds) and normalizes so all horses together add up to 100%. That gives Model % — a market-based indicator for the ranking, not a known winning chance.
Then we compare against Pick % — how large a share of the bettors in the pool (e.g. V75 or V64) chose the horse. The share Comparison = Model % / Backing % is used for analysis and ranking: above 1 means a relatively lower backing share than the model indicator, below 1 means a relatively higher backing share. It is not a claim about EV+ or future returns.
The tags: BANKER = highly ranked without being heavily overbacked. UPSET 💎 = a lightly backed horse with a higher model indicator relative to its backing share. OVERBACKED ⚠️ = the horse is significantly more backed relative to the model indicator.
Note that the backing percentage moves right up to post time — the analysis is a snapshot, not a forecast of payout.