Forecasts & intelligence
The model layer: Dixon-Coles forecasts, accuracy backtests, grounded market-move explanations and LLM-authored briefs. Intelligence endpoints cost 12 credits; the social context block costs 8.
get-match-forecast
POST /get-match-forecast · intelligence (12 credits)
Dixon-Coles model forecast for one fixture — 1X2 probabilities plus model-vs-market gaps.
| Body field | Type | Description |
|---|---|---|
event_urn | string | required — get it from /get-schedule |
include_reasoning | bool | include the model's reasoning trace |
min_gap_bps | int | only report model-vs-market gaps above this (basis points) |
curl -X POST https://api.machina.gg/world-cup/v1/get-match-forecast \
-H "X-Api-Token: $KEY" -H "Content-Type: application/json" \
-d '{"event_urn": "urn:machina:sport:soccer:event:brazil-vs-morocco:20260613:wor"}'backtest-forecasts
POST /backtest-forecasts · intelligence (12 credits)
Historical accuracy report of the model over played fixtures — calibration plus a CLV (closing-line value) ledger. Use it to decide how much to trust the forecasts.
| Body field | Type | Description |
|---|---|---|
league | string | competition filter (default: World Cup) |
season | string | season filter |
explain-market-move
POST /explain-market-move · intelligence (12 credits)
Grounded explanation of why a market moved: price history fused with live research, source-cited.
| Body field | Type | Description |
|---|---|---|
market_id | string | required — from search-markets or market-movers |
window_hours | int | move window to analyze (default 24) |
min_move_bps | int | ignore moves smaller than this |
include_reasoning | bool | include the reasoning trace |
generate-market-brief
POST /generate-market-brief · intelligence (12 credits)
LLM-authored market brief for a fixture or theme — grounded in live markets, optional social pulse.
| Body field | Type | Description |
|---|---|---|
event_urn | string | scope to one fixture |
query | string | free-text theme when no fixture |
market_ids | array | specific markets to cover |
depth | string | brief depth/length |
include_social_pulse | bool | blend in fan sentiment |
fan-sentiment-context
POST /fan-sentiment-context · social (8 credits)
The social-sentiment building block behind the Fan Pulse card: live X/web search (xAI Grok) summarized with citations — without the full card authoring.
| Body field | Type | Description |
|---|---|---|
event_urn | string | scope to one fixture |
query | string | free-text subject |
lookback_hours | int | search window |
force_live | bool | bypass cache |
Block vs card
Need a publishable sentiment piece? Use /skills/fan-pulse (40 credits). Need the raw sentiment context to compose yourself? This endpoint at 8.