PLLAY Publishes Phase 1: Recorded-Video Outcome Verification, Benchmarked and Sandbox-Ready

Phase 1 verifies game outcomes from recorded video. An offline benchmark of 38 clips across 14 titles is published, and the settlement flow can be tried in the sandbox.

PLLAY today set out Phase 1 of its Outcomes Oracle: verifying the result of a game from recorded video, then settling or holding on that result. Each title's status is published on the site's coverage and status pages, and the settlement flow can be exercised in the sandbox.

The Phase 1 benchmark ran 38 public clips across 14 titles through the full Vision pipeline from scratch, with no fallback. Each clip has a ground-truth outcome checked by a person. All 38 matched the expected result: 31 settled automatically, 4 were held as expected, 1 was sent for confirmation, and 2 were voided. None of the automatic settles was wrong.

The limits matter as much as the result. The clips were chosen by the team, 2 to 5 per title, so this is not a random sample. It is an offline benchmark on recorded video, not an accuracy rate for any title and not a production test. There is no live-video result in the set, and no authenticated production tests are on record yet; those will appear on the status page when a named partner export lands.

Live verification is not part of Phase 1. Where it is offered, it carries a reviewer. Certified automated live verification is not available today and is planned as Phase 2.

Every decision returns a settlement_id with its evidence. A result is settled only when confidence clears the threshold; below it, the event is held or refunded rather than guessed.

Media Contact: press@pllay.io