End-of-milestone HTML report: explain each tested strategy (+ graph), show & evaluate results #142

New Issue

2026-06-25T12:32:02+02:00

Brummel commented

2026-06-25 12:32:02 +02:00

What

A single self-contained HTML overview, produced at the end of the edge-research
milestone, that for each strategy candidate screened during the milestone:

explains the strategy in prose — the signal mechanic, the params, the
hypothesis it tests (e.g. SMA-cross momentum; Donchian channel breakout);
shows its graph — the strategy DAG, ideally via the existing aura graph
HTML rendering (a node/edge diagram), so the reader sees the wiring, not just prose;
presents the results — the Stage-1 R metrics across the cross-symbol +
temporal-OOS screen (the E[R] / sqn / sqn_normalized cross-tabs, per symbol and
per IS/OOS split), as tables and/or charts;
evaluates / scores it — the honest verdict (refuted / survives / promoted to
Stage 2), with the reasoning (cross-symbol generalization, statistical
significance, OOS robustness).

Why

The milestone's output is a truthful per-candidate verdict (#137). This report is
the human-readable capstone that makes those verdicts legible in one place — what was
tried, how it was wired, what the data said, and why each candidate lived or died.
Requested by the user as the milestone deliverable.

Notes / constraints

Web-from-disk, each result rendered natively (not a terminal/stdout mirror) — the
project's visual-face direction: recorders persist traces to disk, a web frontend
charts them. aura graph already renders a DAG to self-contained HTML; aura chart
renders run curves; both are reuse candidates.
Self-contained static page, served on the dev host (the user is usually remote).
Depends on the candidates being screened first — this is the aggregation /
presentation layer over the per-candidate screens (momentum refuted on #141;
breakout in progress under #137/#0071; further candidates may follow).
Much of the per-strategy content already exists as the #137 / #141 decision log
(cross-tabs, verdicts) — the report formalizes and visualizes it.

refs #137

## What A single self-contained HTML overview, produced at the **end of the edge-research milestone**, that for each strategy candidate screened during the milestone: - **explains the strategy** in prose — the signal mechanic, the params, the hypothesis it tests (e.g. SMA-cross momentum; Donchian channel breakout); - **shows its graph** — the strategy DAG, ideally via the existing `aura graph` HTML rendering (a node/edge diagram), so the reader sees the wiring, not just prose; - **presents the results** — the Stage-1 R metrics across the cross-symbol + temporal-OOS screen (the E[R] / sqn / sqn_normalized cross-tabs, per symbol and per IS/OOS split), as tables and/or charts; - **evaluates / scores it** — the honest verdict (refuted / survives / promoted to Stage 2), with the reasoning (cross-symbol generalization, statistical significance, OOS robustness). ## Why The milestone's output is a *truthful per-candidate verdict* (#137). This report is the human-readable capstone that makes those verdicts legible in one place — what was tried, how it was wired, what the data said, and why each candidate lived or died. Requested by the user as the milestone deliverable. ## Notes / constraints - Web-from-disk, each result rendered natively (not a terminal/stdout mirror) — the project's visual-face direction: recorders persist traces to disk, a web frontend charts them. `aura graph` already renders a DAG to self-contained HTML; `aura chart` renders run curves; both are reuse candidates. - Self-contained static page, served on the dev host (the user is usually remote). - Depends on the candidates being screened first — this is the *aggregation / presentation* layer over the per-candidate screens (momentum refuted on #141; breakout in progress under #137/#0071; further candidates may follow). - Much of the per-strategy content already exists as the #137 / #141 decision log (cross-tabs, verdicts) — the report formalizes and visualizes it. refs #137

Brummel added this to the Edge research: prove a real tradeable edge milestone 2026-06-25 12:32:02 +02:00

Brummel added the feature label 2026-06-25 12:32:02 +02:00

Brummel referenced this issue

2026-06-25 15:29:19 +02:00

Strategy research: find a Stage-1 positive-E[R] edge (honest OOS) #137

Brummel commented

2026-06-25 15:29:20 +02:00

Closing: the premise is gone. This was an end-of-milestone overview of the edge research; that milestone is being wound down rather than completed (the simple-signal thesis was honestly disproved — see #137; finding an edge by hand was a means-end inversion). The verdicts on #137 are the honest record of what was tested and why it failed. If a write-up still has value later — reframed as a DEMO of Aura's lie-detector rather than an edge report — reopen.

Brummel closed this issue

2026-06-25 15:29:20 +02:00

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: Brummel/Aura#142