Skip to main content
open research book live-monitored

Investing techniques,
tested in public.

Every idea in this book is preregistered, backtested against real costs, and then either killed in public or promoted to live monitoring. Every mark beside this text is one technique — 92 of them across 10 research areas. The bright ones are alive; the dim majority are killed. Count them. What they were — and, for most of them, what would reopen the question — is another matter.

Explore the research atlas: 92 techniques, 41 of them killed.
techniques mapped
92
killed · lessons banked
41
kills with no reopen trigger
28
sleeves carrying capital
7
research areas
10
  1. 01

    Preregister

    Hypothesis, gate thresholds, and kill criteria are committed before the data is allowed to vote.

  2. 02

    Test against real costs

    Spread, fees, tax, and latency included. Gross-only edges are not edges.

  3. 03

    Kill in public

    41 of 92 techniques died so far, each with its reason banked. 13 also carry a written reopen condition; 28 do not yet, and that gap is published rather than rounded off.

  4. 04

    Monitor live

    Survivors run under decay monitors, breakers, and a morning-briefing desk. Promotion is never permanent.

see all 15 arms →
1.27
mean sharpe · TimeSeriesMomentumStrategy
synthetic universe · no claim of a live edge
stdev 0.78 · range 0.63 2.14 · +0.27 vs baseline
cumulative pnl · 200 bars · seed 7
day 0
ts_momentum baseline
day 200
All notes →

New here? About covers who runs the area, how to read it, and where to send a correction.

deskgd findgf strategiesgs featuresge promotionsgp apiga botgb writinggw

Shortcuts are no-ops while typing in an input or textarea.