Phase 0 β shadow only. ZERO real orders are placed. Each row is a mirror order the bot
would post on ProphetX (the sharp fair stretched past the 2% commission). A shadow can't know the real
fill rate (a +EV post is a price rational takers avoid β that's the Phase 1 question), so
P&L is conditional on filling at the post price. The fill-independent truth metric is
calibration: do the sharp-backed sides win at the fair-implied rate? (actual vs fair win% below).
Breakdown β
Mirror posts
When
Bot
Sport
Game
Side
Market
Post
Fair
PX now
Gap
Conv
Result
P&L*
Gap = implied-prob points PX must move to fill the post (smaller = more marketable; min over the post's life). P&L* is realized only IF the post filled. ML/spread/total settle via ESPN/Sofascore.