Status: This is a current working protocol, not a mature strategy declaration. It is being tested in live practice and will change when evidence shows that it should.

The public version describes method and high-level review. Detailed trades, positions, and account data remain private. Nothing here is financial advice.

1. Purpose and scope

The protocol is designed for low-frequency, months-to-years investing. Its purpose is to make research and decisions more explainable, comparable, and revisable while keeping investing within a sustainable work-and-life boundary.

The framework is evaluated by process questions rather than short-term returns alone:

  • Were events and narratives translated into the right variables?
  • Were the main operating drivers and dependencies identified?
  • Were scenarios meaningfully different rather than cosmetic variations?
  • Was uncertainty stated before the outcome became known?
  • Did new evidence update the structure in a disciplined way?

2. Current constraints

These are working constraints intended to limit avoidable risk while the framework is still being tested:

  • no leverage and no high-frequency trading;
  • low action when expected return or uncertainty cannot be explained;
  • no change to a long-term thesis solely because of short-term price movement;
  • technical signals may inform execution, but do not create the fundamental thesis;
  • complexity must earn its place by improving a decision or exposing an assumption;
  • taking no position is always an admissible outcome.

These constraints are defaults, not evidence that the process has been validated.

3. Research architecture

The current architecture has three connected layers. A fuller public summary is available in Framework.

3.1 Translation

Fundamental analysis begins before valuation. An event such as deteriorating competition is not yet a model input; it must be translated into variables such as pricing pressure, market-share change, retention, margin compression, capital intensity, or a lower valuation multiple.

The first question is therefore not “which formula should I use?” but “what would have to change in the business for this narrative to be true?”

3.2 Risk-Reward scenarios

Business model, operating drivers, risks, and valuation are compressed into Bull, Base, and Bear scenarios. Each scenario should state its distinguishing assumptions, the implied operating path, a valuation range, and evidence that would move probability toward or away from it.

Scenarios are decision objects, not predictions presented with certainty.

3.3 Agentic R-D-G

Agentic R-D-G organizes events, operating drivers, dependencies, risk paths, and scenario prices into a structure that can be updated as evidence arrives. The aim is to make the path from qualitative research to probability, distribution, risk exposure, and decision visible enough to inspect.

It is not an automatic trading system. AI may help gather evidence, challenge assumptions, and maintain structure, but responsibility for interpretation and risk decisions remains human.

4. Research and decision cycle

For a new or revised thesis:

  1. Define the question and the decision it could affect.
  2. Map the business model, key drivers, dependencies, and risk paths.
  3. Translate important narratives into observable variables.
  4. Build Bull, Base, and Bear scenarios with valuation ranges.
  5. Record confidence, a review date, verification sources, and falsification conditions.
  6. Compare the opportunity with realistic alternatives, including near-riskless yield and no action.
  7. If action is justified, keep implementation consistent with uncertainty.
  8. Review the thesis when scheduled evidence or a defined trigger arrives.

The public record emphasizes steps 1-6 and 8. Step 7 and account-level implementation remain private.

5. Uncertainty review

Substantive forecasts should be written so that later review is possible. A useful private record includes a clear proposition, a confidence level chosen before resolution, a deadline or review date, a verification source, and observations that would count as disconfirmation.

The public site keeps this at the level of method and framework changes. Detailed forecast logs remain local.

6. Updating rules

New evidence can change different parts of the structure:

  • an event may alter a driver without invalidating the business model;
  • a driver may change the scenario range;
  • a risk path may change scenario probabilities;
  • a price change may alter prospective return without changing the business thesis;
  • repeated review error may require changing the framework itself.

Updates should identify which layer changed and why. A new opinion is not enough; the changed variable, relationship, probability, or decision threshold should be visible.

7. Attention boundary

Research is done in bounded windows rather than through always-on monitoring. Review frequency should follow the information cadence of the thesis, not the emotional cadence of the market.

If investing begins to crowd out broader work, learning, health, or relationships, reduce its frequency and scope. Attention is part of risk management.

8. Public and private records

The public lab contains framework summaries, methodology notes, and high-level review structure. Exact positions, trade records, forecast logs, cost bases, orders, and account-level exposure remain private.

The older Log, Lessons, and Monthly Reviews are retained as historical records. They may reflect earlier versions of the framework, including rules and assumptions that are no longer current. They should be read as dated evidence of iteration, not as presently binding instructions.

9. Version note

This protocol replaces the earlier “playbook as constitution” framing. The current emphasis is on translation, scenario construction, explicit uncertainty, review discipline, and structured updating. Version changes will be recorded when live testing produces a meaningful change in method, not merely when wording is edited.