Skip to content
Agent90Beta
Back to home
// HOW IT WORKS

One agent. Nine stages. No black box.

Agent Ninety is a pipeline. Each stage has one job and a defined input/output. The point of publishing the architecture is the same as the point of locking the snapshot: you should be able to see how the agent thinks before you decide whether to trust it.

// AGENT PIPELINEPhase 1 · mocked

A pipeline, not a black box.

The agent is a sequence of small, named stages. Each has one job and a defined input and output. In Phase 1, every stage is illustrated with hand-authored mock data — no live model calls, no autonomous browsing, no real learning runtime.

  1. 01

    Data Pack

    Builds a normalized brief for one fixture: form, lineups, rest, venue, recent context.

    Future phase
  2. 02

    Baseline Prediction

    Anchors probabilities so downstream agents can't drift into vibes.

    Future phase
  3. 03

    Analyst

    Drafts the thesis, hypotheses, and what-to-watch from the data pack.

    Future phase
  4. 04

    Skeptic

    Attacks the draft. Surfaces over-confidence, missing factors, weak hypotheses.

    Future phase
  5. 05

    Editor

    Reconciles analyst and skeptic into one voice. Enforces tone, removes hedging slop.

    Future phase
  6. 06

    Snapshot Locker

    Hashes, timestamps, and persists the final pre-match report. Immutable from this point.

    ● Phase 1 demo
  7. 07

    Post-Match Reviewer

    Grades each hypothesis against actual events. Writes what was right, wrong, and learned.

    Future phase
  8. 08

    Memory Curator

    Decides what is worth remembering, in what wording, and what to retire. Curated, not naive.

    Future phase
  9. 09

    Scorecard Update

    Recomputes calibration, hypothesis hit-rate, match-shape accuracy, and the public scorecard delta.

    Future phase
// WHAT “LEARNING” MEANS

The agent learns at the system level, not at the model-weights level. Memory gets updated. Hypotheses get graded. Calibration gets tracked. Mistakes become future context. Bad memories can be retired.

No fine-tuning. No retraining. The next match's data pack is just better-informed than the last one.

// WHAT WE DO NOT DO
  • Edit a snapshot after kickoff. Ever.
  • Skip a review when the agent was wrong.
  • Promise a result. Confidence is calibrated, not sold.
  • Frame any of this as betting advice, value picks, or edge.