AR-WIZ-001·PL-01 — The framework archive. Stylised brand illustration of the laboratory, not a photograph of a physical facility. The charts and figures inside the plate are simulated.

Record AR-PUB-142 · Wizard Framework Archive

Decision frameworks,
not personalities.

This archive studies doctrines. Each profile is an educational framework abstraction assembled from public-interview interpretation — what a widely discussed approach appears to claim about markets, how it decides to enter and exit, what it does when it is wrong, and which conditions make it coherent. No profile is a reconstruction of any individual's private method, and none is attributed to a named trader.

Published 84 profiles Planned library 80–100 Status Preview Functionality Level 1 Updated 2026-07

01Archive DefinitionAR-PUB-142·01

A doctrine is portable.
A person is not.

The public trading literature is largely biographical. It records who made money, in which decade, and how they described the experience afterwards. Almost none of it records the decision rule that produced the outcome, because the decision rule was rarely written down and, where it was, it was not the part the interviewer asked about.

What survives translation is narrower and more useful: a strategy-philosophy model. A statement about what the market is assumed to do, a rule for when that assumption is considered confirmed, a rule for when it is considered dead, and a stance on how much capital the assumption deserves. Those four things can be examined, compared and — carefully — expressed as parameters. The person cannot.

So each entry here is an abstraction, deliberately anonymised. Where a profile says “enter on confirmed continuation rather than anticipation”, that is an interpretation of a position stated repeatedly in public interviews across a period, not a quotation and not a rule anyone published. Where it says the doctrine reduces size in high volatility, that is a summary of a described stance, not a specification with numbers attached.

The archive exists because doctrines fail in patterned ways. Two operators who have never met and who trade different instruments will encounter the same problem if they hold the same view of why price moves — and that shared problem is far more instructive than either of their track records.

The vocabulary this archive uses

Public-interview interpretationThe source layer. Positions a doctrine's proponents have discussed openly, read across many sources and a long period rather than taken from one account.
Educational framework abstractionThe profile itself: the doctrine restated as a decision structure, stripped of identity, dates, instruments and results.
Risk-doctrine presetA stance on sizing, loss control and exposure, expressed as a direction of travel — smaller here, capped there — never as a calibrated number.
Strategy-philosophy modelThe claim about market behaviour underneath the entry rule, and the conditions under which that claim is coherent.
Intraday adaptationWhat happens to the doctrine when its horizon is compressed: what it keeps, what it loses, and whether the remaining edge is still the same edge.

What this archive never claims

It does not claim to have reconstructed anyone's method. It does not attach returns, win rates or track records to any profile. It does not assert that a doctrine described here is the doctrine any particular trader actually executed. A public description of a private process is evidence about the description, not about the process.

02Profile IndexAR-WZ-001…012

Twelve doctrines,
opened in full.

Filter by asset class, strategy style, horizon or mode, or search the full text of every profile — the search reads the expanded record, not just the summary. Each profile opens into its complete record: fourteen fields, none of them truncated, including the ones that are unflattering.

Search the archive

Risk posture

Holding horizon

Decision mode

Showing 84 of 84 published profiles

03Comparison LayerAR-PUB-142·03

Four dimensions,
eighty-four interpretations.

Every cell below is an editorial reading of the written profile, not a measurement of anything. No trade data underlies this grid. It exists so that doctrines which sound different can be seen to share a structure — and so that doctrines which sound similar can be seen to disagree about the thing that matters most.

Figure 1 · Doctrine comparison across four interpreted dimensions Interpretation, not measurement

Each cell carries a score from 1 to 5 on the scale defined beneath the figure. Colour encodes magnitude only — a 5 in horizon length is not better than a 1, it is longer. Hover a cell for the word behind the number. The two columns that most reliably predict how a doctrine behaves in a drawdown are hit-rate expectation and tail dependence, and they move in opposite directions almost everywhere in the archive.

How the four dimensions are scored

04Intraday AdaptationAR-PUB-142·04

What survives
when the horizon collapses.

Most retail adoption of a published doctrine happens at a compressed horizon, because that is the horizon a part-time operator can supervise. The question the archive asks of every profile is therefore not “does this work” but “is it still the same thing at one-tenth of the holding period”.

Two effects dominate. Friction rises as a share of the working stop, because cost is charged in price and the stop shrinks with the timeframe. And the right tail shortens, because an extended continuation needs elapsed time to occur at all. A doctrine that earns its expectancy from rare large outcomes therefore loses the source of its edge before it loses its hit rate — which is why the compressed version can look acceptable for a long time and still be structurally broken.

Doctrines that are native to the session — momentum inside a defined window, liquidity reading, scalping — lose nothing, because they never depended on time in the first place. They simply inherit a different problem: cost.

Where expectancy leaks at short horizons →
Figure 2 · Interpreted intraday transferability by doctrine Interpretation, not measurement

An editorial reading of each profile's intraday adaptation field, expressed on a 0–100 scale where 100 means the doctrine is native to the session and 0 means compression removes the mechanism entirely. The dashed line marks the point below which the compressed version should be treated as a different strategy requiring its own evidence.

Intraday adaptation notes for each of the twelve doctrines
DoctrineScoreAdaptation note from the profile

05Parameter Preset PreviewAR-SIM-047·WZ

If this doctrine
had to be settings.

Every profile translated into the 159 controls of the preview engine. Each translation is an interpretation written by the Foundry — the profiles contain no numbers, and none of the traders whose publicly discussed positions informed them has specified a coefficient, a partial percentage or a risk fraction anywhere in the public record.

Level 2 · interactive preview Simulated

Read this section as a demonstration of direction, not of value

The purpose is to show that a philosophy has parameter consequences: a doctrine that refuses to cap its upside implies a low partial percentage and a late trail, whether or not its proponents ever discussed either. The engine then reports what that combination does to expectancy, decline and tail participation inside its own model. It does not report what the doctrine did in any market.

Interactive demonstration — not a historical backtest

The presets disagree about the same control

Partial percentage is set at 10 in the trend preset and 80 in the reversion preset, and both are defensible readings of their doctrine. That single control moves net expectancy and tail participation in opposite directions, which is the clearest available demonstration that a parameter is not good or bad in isolation — it is only consistent or inconsistent with a claim.

A preset is not a starting point for capital

These four configurations were written to express philosophies, not to score well, and several of them score poorly on the engine's own composite. Anyone tempted to copy the numbers should note that the engine reads no market data and that the doctrine they are copying was described after the fact by someone who no longer had to trade it.

06MethodologyAR-PUB-142·06

How a profile is written.

Four stages, each with a different reliability. Knowing which sentence in a profile came from which stage is the difference between using this archive and being misled by it.

STAGE 01 · COLLECTION Read the public record across a period, not a single account A doctrine is only admitted when its proponents have discussed the same position repeatedly across years and across interviewers. A view expressed once, in one conversation, after one exceptional year, is not a doctrine — it is a remark. Each profile records the period its source material spans so a reader can see how much of it is one decade's conditions. Weight: moderate · the source is testimony, not documentation
STAGE 02 · ABSTRACTION Strip identity, instrument, era and outcome The profile keeps the decision structure and discards everything that would make it a biography. Names, firms, dates, headline returns and specific trades are removed deliberately — not for legal caution but because they are the parts most likely to be remembered and least likely to be transferable. What remains is a claim about market behaviour and a stance on risk. Weight: high for structure · zero for attribution
STAGE 03 · COMPLETION Force an answer into every one of the fourteen fields A doctrine that has nothing to say about winner management or psychological demand still gets a field, and the field is filled by reasoning from the rest of the profile rather than left blank. This is the least reliable stage and the most useful one: it makes profiles comparable, and it exposes doctrines whose public description is silent on exactly the decisions that determine an equity curve. Weight: authored inference · flagged in each profile's limitations field
STAGE 04 · TRANSLATION Express the doctrine as engine controls, where it is honest to do so Every one of the eighty-four profiles carries a parameter configuration, because a design position that cannot be make claims specific enough to translate without inventing content. The remaining eight are described in words and left there. A preset is written by asking which control the doctrine's own language implies, then accepting whatever the engine reports — including when the result is poor. Weight: illustrative only · directional relationships, not magnitudes

The fourteen fields

Publication plan

The archive is designed for 80 to 100 profiles. Twelve are published. The structure of a profile is fixed, so growth adds records without rewriting any of them — and the comparison layer, the filters and the identifier scheme are all built for a library an order of magnitude larger than the present one.

Published now84 profiles across six sectors — trend and continuation, mean reversion and fade, breakout and expansion, liquidity and market structure, session and time architecture, and risk, capital and governance. Each carries one of five risk postures and an executable configuration.
Next tranchePlanned coverage of carry, event-driven positioning, relative value, options-structure doctrines and market-making stances — none of which is represented at all today.
Selection ruleA doctrine enters the archive when it is distinct from every published profile in at least two of the four comparison dimensions. Variations on an existing doctrine are recorded inside that profile, not as new records.
Deliberate exclusionsDoctrines whose public description is entirely narrative, and doctrines whose mechanism depends on infrastructure no reader could obtain, are excluded rather than abstracted into something they are not.
Level 3 · expansion is future work

07Limitations of the ArchiveAR-PUB-142·07

The public record
is a biased sample.

Five problems apply to every profile here, including the ones that read most convincingly. They are not caveats added at the end; they are properties of the source material, and a reader who does not hold them in mind will overread the entire archive.

Nobody is interviewed about the doctrine that quietly failed. The population of publicly discussed approaches is selected on outcome, and outcome over a career is a function of the doctrine, the era, the capital base, the cost structure, the institutional support and a large quantity of luck — with no way to separate the contributions.

The consequence is specific and uncomfortable: for every doctrine in this archive there is an unknown number of operators who held the same view, executed it comparably, and are not in any book. Their absence is not evidence that they were worse. It means the archive describes the surviving tail of a distribution and says nothing about the width of that distribution.

This is why the archive attaches no returns to any profile. A doctrine's presence here is evidence that it was articulated well by someone who did well. It is not evidence about the doctrine's expectancy.

An interview is a reconstruction. The trader is explaining, years later, why a set of decisions made sense — and human memory reliably tidies a messy process into a coherent one. Rules that were discovered halfway through become rules that were held from the start. Positions that were exited in discomfort become positions that were exited by design.

The specific distortion this creates is that published doctrines are almost always more consistent than the behaviour that produced the record. A profile therefore describes an idealised version of a practice, and an operator who fails to match its consistency is not necessarily falling short of what the original did — only of what the original later described.

Each profile records the period its source material spans, and those periods are not interchangeable. A trend doctrine articulated across a stretch of unusually persistent macro direction is a claim that was true of that stretch; the doctrine does not carry a marker saying which of its properties were structural and which were the decade.

Cost structure moves the same way. Doctrines formed when spreads were wide and commissions were high were selected for tolerance of friction, and doctrines formed afterwards were not. A scalping doctrine from a period of favourable microstructure is not portable to a retail cost base at all, which is why the archive includes one explicitly for contrast rather than as a model.

The honest treatment is to read the period line at the top of every profile as part of the doctrine, not as metadata.

The distance between what a trader says they do and what their fills show is the most reliably underestimated quantity in this field. A doctrine that says “never chase” is consistent with a record containing a great many chased entries; a doctrine that says “let winners run” is consistent with a record of early exits during the one drawdown that mattered.

Nothing in the public record allows the two to be separated, because the fills are not public. This is not a criticism of any particular account — it is a structural property of testimony. It also means the profiles describe intent, and intent is the easiest part of a system to adopt and the least consequential.

The practical implication is that adopting a doctrine from this archive without the diagnostics that would reveal a gap between stated and executed behaviour reproduces the weakest part of the source material rather than the strongest.

Three of the fourteen fields — winner management, psychological doctrine and intraday adaptation — are frequently absent from the source material and are completed by reasoning from the rest of the profile. That reasoning is the Foundry's, and it will carry the Foundry's assumptions about how systems fail.

The comparison grid and the parameter presets amplify this. A doctrine's position on the tail-dependence axis is an editorial judgement, and a preset's partial percentage is an editorial judgement expressed as a number, which makes it look more objective than it is. Both are labelled as interpretations wherever they appear, and both would move if a different author wrote them.

Twelve is a thin sample of a large space

Entire families of professional practice — carry, relative value, event positioning, options structures, market making — have no representation here. The archive's coverage looks broad only because retail literature is itself concentrated on directional discretion.

The filters imply more precision than exists

Asset class, style, horizon and mode are labels the Foundry assigned. Several profiles sit awkwardly in their category: the risk-doctrine profile has no signal logic at all and is filed under a style that is not a style.

Nothing here tells you what to trade

The archive contains no ranking, no recommendation and no route to one. Its only defensible use is comparative: to recognise which doctrine your own approach already resembles, and therefore which failure mode you have already inherited.

A doctrine explains the intent.
The parameters decide the outcome.

Every profile here ends at the same boundary: a philosophy stated clearly enough to be argued with, and no numbers behind it. The gap between the two is where systems are actually won and lost, and it is the only part of this that can be examined directly.

Translate a doctrine into parameters Framework library Risk and drawdown Research archive

84 published profiles · 6 sectors · 5 risk postures · every profile executable in the laboratory