Evidence before action

Our methodology

A prediction is an estimate, not a bet. FIND THE EDGE keeps the evidence behind a decision separate from the outcome that follows.

Model estimates versus market prices

A model estimates the chance of an outcome from its inputs and assumptions. A market price reflects the offered odds, including the bookmaker’s pricing margin. A likely winner is not automatically a worthwhile bet.

A meaningful comparison needs the same match, market, settlement rules and observation time. All outcomes in that market are needed to estimate its margin. Missing or outdated prices do not establish value.

The comparison can use a margin-adjusted market estimate, but that estimate is not a known true probability. Read how odds and probability differ.

Uncertainty stays part of the decision

Incomplete inputs, changing line-ups, stale prices and model assumptions can all change an estimate. A small apparent difference may be noise rather than a reliable opportunity. Predictions can be wrong even when the underlying process is sound.

BET requires sufficient evidence, a demonstrated estimated edge and explicit validation. WATCH means evidence is incomplete. PASS means no robust value is established. These labels are not promises of a winning outcome.

Model evaluation should examine probability accuracy and calibration across an adequate sample, not just a short winning sequence. No audited model performance is claimed here.

Three records that must not be confused

Technical opportunity
An analytical signal worth checking. It is not an accepted wager, an official bet or evidence of profit.
Paper trading
A tracking exercise using recorded selections and accounting rules. Its balance does not prove money was wagered. A retrospective ticket must remain identified as retrospective.
Official bet
An explicitly validated and published selection with traceable pre-match price evidence. A real-money wager additionally requires evidence that it was placed; an official editorial label alone does not prove acceptance by a bookmaker.

What the public record can establish

The public history contains supplied selections and recorded outcomes, not a live model-performance feed. Publication evidence and stake timing are shown only where supplied. A recorded win alone does not establish model quality or turn a supplied ticket into an official model-backed bet.

Inspect the recorded results and their limitations. No loss should be removed and no selection should be rewritten after the outcome.