Betting signals (edge)

Model forecast × best market price → actionable, informational edge. Both endpoints cost 18 credits.

Informational only

Signals are decision support — never betting advice and never a guaranteed outcome. Sizing uses fractional Kelly precisely because the model can be wrong.

get-signal

POST /get-signal · edge (18 credits)

Per-fixture betting signal: fuses the Dixon-Coles forecast with the best available market price into a recommendation with edge %, EV and a fractional-Kelly stake.

Body field	Type	Description
event	string	natural fixture description ("Brazil vs Morocco")
event_urn	string	canonical URN (preferred when known)
team / opponent	string	alternative to event
date	string	YYYY-MM-DD, disambiguates rematches
bankroll	string	optional bankroll for stake sizing
kelly_fraction	number	fraction of full Kelly (default 0.25)
min_edge_bps	int	minimum edge to surface, basis points (default 200)
fee_bps	int	market fee in basis points (default 0)
include_reasoning	bool	include the signal's reasoning trace

curl -X POST https://api.machina.gg/world-cup/v1/get-signal \
  -H "X-Api-Token: $KEY" -H "Content-Type: application/json" \
  -d '{"event": "Brazil vs Morocco", "bankroll": "1000", "kelly_fraction": 0.25}'

Response: recommendation (side/skip), edge in bps, expected value, suggested stake, the market used for pricing, and the resolved event_urn.

find-market-edges

POST /find-market-edges · edge (18 credits)

Tournament-wide scan: every open market compared against the model — returns informational edge candidates ranked by gap.

Body field	Type	Description
min_edge_bps	int	minimum edge, basis points (default 200)
limit	int	max candidates (default 10)
status	string	market status filter (e.g. "open")
include_reasoning	bool	per-candidate rationale

curl -X POST https://api.machina.gg/world-cup/v1/find-market-edges \
  -H "X-Api-Token: $KEY" -H "Content-Type: application/json" \
  -d '{"min_edge_bps": 300, "limit": 5}'