How DaBoard reads the market

We pull live prices from roughly a dozen sportsbooks and store every capture. Nothing on the board is a projection — it's the market itself, recorded, timestamped, and shown as-is (a sparse feed renders sparse; liveness is never faked).

Every posted price has the book's cut baked in. We remove it — de-vig — to get the fair line: what the consensus market really implies once the house edge is taken out. We label it “fair,” never “model” — it's market math, not a forecast.

Prices move, so we keep the full trail — open to the latest stored tick — and let you scrub it. But movement is a fact we report, not a signal to chase: our own backtest measured no edge in following intraday moves, so you'll never see a “steam, bet it” cue here.

The same bet pays differently at different books. We surface the best available number and name the book offering it — and flag when a book still sits off the consensus line, a line-shop gap in our latest stored read. The product is the best number, not just any number.

The edge is the gap between the de-vig fair line and the best price. We qualify it by how many books priced it — a thin market is easy to misprice — so a thin edge is cautioned and an outlier edge reads amber, never green. There is no “PLAY” badge.

The honest scoreboard is CLV: did the price you got beat the closing line? We measure it on observed closes only, never estimated (estimating would flatter us). That proof lives in the Ledger — with losing strategies left on the books.