AlmanacDocs
API referenceEvents

Recently scored events

Resolved markets whose predictions have been scored, newest resolution first (by the source's resolvedAt), with each miner's result. pWin and brier are the exact per-event values the leaderboard scores (winning-outcome Brier, scorer rounding); invalid predictions carry nulls and sort last, and predictions without a uid are left out as the scorer leaves them out. The default query (no filters, limit 10, miners 10) is cached for 60s.

GET/v1/events/scored

Query Parameters

miners?number

Miner rows per event, best Brier first, 0..50. Defaults to 10; 0 returns stats only, plus a requested miner's row.

uid?number

Only events scored for the hotkey registered to this UID now (a re-registered UID is a new miner); its row is always included. Not with minerHotkey.

minerHotkey?string

Only events this hotkey was scored on; its row is always included in miners. Not with uid.

cursor?string

Opaque pagination cursor from a prior response's nextCursor. Pass verbatim.

limit?number

Events per page, 1..25. Defaults to 10.

Response Body

application/json

curl -X GET "https://example.com/v1/events/scored"
{  "items": [    {      "eventId": "string",      "source": "polymarket",      "sourceMarketId": "string",      "question": "string",      "resolvedAt": "2019-08-24T14:15:22Z",      "scoredAt": "2019-08-24T14:15:22Z",      "winningOutcome": {        "outcomeId": "string",        "name": "string"      },      "marketPWinAtAdmission": 0,      "stats": {        "scored": 0,        "valid": 0,        "meanBrier": 0,        "calledRate": 0,        "bestUid": 0      },      "miners": [        {          "uid": 0,          "hotkey": "string",          "pWin": 0,          "brier": 0,          "called": true,          "invalid": true,          "predictedAt": "2019-08-24T14:15:22Z"        }      ]    }  ],  "nextCursor": "string",  "meta": {    "computedAt": "2019-08-24T14:15:22Z",    "scorerVersion": "string"  }}