Hadrek

Protocol / 001 — London · Published every day

You lost money.
Not luck. Process.

> Nobody here tells you what to buy. You get one person's daily work, in writing, at the same time every day.

Get this week's teardown — freeSee the method

No signals. No entries. No performance figures anywhere on this site, including mine.

Regime stateRefreshed 2026-08-17

Read published

High volatility, no trend

Range
Wide
Follow-through
Absent
Depth
Thin
Sample for this read
14 prior instances

Strategy

Should not run

Default state

Flat

A read, not a recommendation. No instrument is named anywhere on this site.

Recomputed 2026-08-17. Method last changed 2026-08-14. Update log · What it reads · How Hermes works

Build logMirrors the repo
2026.08.14Risk gate revised: session loss limit now blocks new entries, not just sizing.
2026.08.11Regime detector v0.4 published to the open repository.
2026.08.07Module 04 rewritten around the gates that fire before the trader does.
2026.08.02Journal schema v1 frozen. Thesis field is now required before an entry saves.
Risk gatesModule 04
  • Thesis written before entry
  • Regime checked and recorded
  • Size derived from stop, not conviction
  • Counter-argument written

Every gate fires before the trade, never after it.

One teardown a week

A real decision, the regime read behind it, and the reasoning — written before the outcome is known.

Get the free teardown
On the desk today—

Journal review

The week's entries read back, including the ones that were wrong.

Next: Sun — Nothing. The schedule does not move with the market, and nothing gets skipped because a week was quiet.

The week, fixedSame every week
  • MonConditions readoutOpenThe seven inputs recomputed and printed, with data completeness stated.
  • TueDaily chartOpenOne chart marked up by hand. What the shape is, and where the sample thins.
  • WedWritten sessionSubscriptionA full work-through: regime classification, gates, sizing arithmetic, critic pass.
  • ThuDaily chartOpenOne chart marked up by hand, dated before the outcome exists.
  • FriTeardownOpenA real decision taken apart in public, reasoning written before it resolves.
  • SatJournal reviewSubscriptionThe week's entries read back, including the ones that were wrong.
  • SunNothingOpenNothing is published. Module 08 applies to the desk as well.
The order, every dayFive steps, same sequence
  1. 01

    Read the conditions before looking at any chart

  2. 02

    Write what would have to be true

  3. 03

    Run the gates in order, stop at the first failure

  4. 04

    Argue the other side in writing

  5. 05

    Record the decision, including doing nothing

You are not meant to improvise this. The sequence is the same on a flat day as it is on a fast one. What the modules cover

0. The instrument deck

The board, marked up the way we teach it

How each input is measured
Conditions board / markup 014Vertical axis is a conditions scale · not a price · no instrument named
A · QUIETB · TRANSITIONC · ELEVATEDD · STRESSED1234BREADTH OF AGREEMENT ACROSS SLEEVES
  1. 1

    Follow-through steady. Conditions inside the tested envelope.

  2. 2

    Dispersion widens before the range breaks. First warning the detector reads.

  3. 3

    Depth thins. The same setup now carries a different risk profile.

  4. 4

    Out of regime. The method's answer here is to stop running the strategy.

Shaded blocks

Regime segments the detector labelled after the fact

Solid rule

The channel the conditions scale held while quiet

Arc fan

Decay geometry, drawn from the first quiet reading

Lower panel

How many sleeves agreed with each other at the time

Realised volatility

30d

Rolling dispersion of returns, 30 sessions

QuietNormalElevatedStressed

Volatility of volatility

30d

How unstable that dispersion itself is

QuietNormalElevatedStressed

Trend persistence

30d

Whether moves follow through or revert

QuietNormalElevatedStressed

Cross-sectional dispersion

30d

How differently components behave

QuietNormalElevatedStressed

Liquidity depth

30d

Size available before the book thins

QuietNormalElevatedStressed

Drawdown state

30d

Distance from the rolling high-water mark

QuietNormalElevatedStressed

No score, no target, no timing

Every series on this board is an illustrative shape for the published method. Nothing here is a recommendation, and no asset is named anywhere in it.

Note 1 · 2026-08-30

Written 07:00 London, every day

Wide range, no follow-through, so the desk stays flat

Same seven inputs, same order, same time. Range is wide, follow-through is absent, depth is thin. That combination has appeared fourteen times in the sample I hold, and in every one of them the strategy I run should not have been running.

So it is not running. That is the whole of today's decision and it took four minutes to reach. If you came here for something more exciting than that, you have misread what this is.

The part worth your attention is not the read. It is that the read was written down before anything resolved, with the sample size next to it, so that in a month you can check whether I was describing conditions or describing my mood.

The run

3

Consecutive daily notes since 2025-07-14. Published at 07:00 London, every day, including the days when the answer is that nothing should be done.

PreviouslyNothing is deleted
  • 2026-08-29 · 411Yesterday's classification was too confident and I have said so
  • 2026-08-28 · 410Doing nothing, for the fourth session running

I. The strike wall

Everything this market says, corrected

01

10x this week

Here is what your position size should have been, and why you never worked it out.

02

My private group is up massively this month

No performance figures appear anywhere on this site, including mine.

03

Entry, target, stop. Copy me.

You will never receive an entry, a target or a stop here. That is the whole product.

04

Financial freedom by 25

Most retail traders lose money. Buying this changes nothing about that base rate.

05

The bot prints while you sleep

Module three is about knowing when your strategy should not be running at all.

The method, mechanically

How discipline is actually taught here

Discipline is not a mood. It is a sequence you run in the same order every time, where each step leaves a written artefact behind. Nothing in this sequence tells you what to hold, when to act, or what anything is worth. It governs how you decide, and how you check yourself afterwards.

  1. 01

    Describe the conditions

    Write down what kind of market you are looking at before you form any view about acting. Conditions get named in words, not numbers you can trade off.

    Produces — A dated regime note

  2. 02

    Write the thesis first

    The reasoning goes in the journal while the outcome is still unknown, in a fixed schema, so it cannot be quietly rewritten afterwards.

    Produces — A locked journal entry

  3. 03

    Set the gates

    Rules that fire before you do: how much of the account is at risk, how often you may act, what stops you acting at all. Written once, applied without negotiation.

    Produces — A written rule set

  4. 04

    Argue against yourself

    Build the strongest version of the opposite case in your own words. If you cannot write it, you do not understand the decision well enough to make it.

    Produces — A counter-argument

  5. 05

    Check the sample

    Replay whether this kind of situation has genuinely occurred before, and how thin the evidence is once you stop selecting for the times it worked.

    Produces — An honest sample count

  6. 06

    Review in writing

    Read your own entries back after the fact and grade the process, not the result. Good process with a bad result stays good process.

    Produces — A written review

Standing behind it

Who runs thisReal name, real record of work

This is written by one person, under their own name, with a working history in building and reviewing systems rather than selling access to them. The build log on this site mirrors a public repository, and the journal schema is open source, so the work can be read directly instead of taken on trust.

Identity

Named

No anonymous persona, no stock photography.

Code

Open

Journal schema and regime detector published openly.

Journal

Unedited

Entries stand as written, including the wrong ones.

What is deliberately absentBy design
  • No instruction to buy, sell, hold or avoid anything
  • No entries, exits, targets or forecasts
  • No signals, alerts, copy-trading or tips channel
  • No performance record, past or projected, mine or anyone else's
  • No member counts, testimonials or logos offered as proof

Education about trading process. Not investment advice. Nothing here is a recommendation to buy or sell anything. Most retail traders lose money.

The full statement

II. The method

01

Should you be doing this at all

The base rates, including the ones against buying this.

02

Why you actually lost

Position sizing and regime, not bad picks.

03

Regime

Knowing when your strategy should not be running.

04

Sizing and risk gates

The rules that fire before you do.

05

The critic

Arguing against your own trade before you take it.

06

The journal

Writing the thesis before you know the outcome.

07

Replay

Checking whether this setup has genuinely happened before.

08

Doing nothing

The hardest module and the shortest.

The full breakdown

III. Not for everyone

  • Told what to buy
  • A bot that prints
  • One signal away

If that is you, this will feel slow. Don't subscribe. Most retail traders lose money, and nothing here claims to change that for you.

01 — Public build log

A dated changelog that mirrors the repository.

02 — Open journal

Written before outcomes exist, including the reads that were wrong.

03 — Free teardown

Ungated beyond an email. Better than most paid work in this market.

IV. Start here

One teardown a week. No positions in it.

A real decision, the regime read behind it, and the reasoning, written before the outcome is known.

Get the free teardown