Episode 05

The Problem This Episode Solves

 

Most traders believe they are collecting data.

 

In reality, they are collecting descriptions.

 

When variables are undefined and allowable values are not clearly constrained, identical examples get logged differently. Different examples get logged identically. Results become narrative instead of measurement.

 

This episode exists to correct that problem.

 

The objective here is to understand why undefined variables corrupt trading data, to differentiate features from variables and states, to learn a structured process for defining both, and to formalize the variable system used in my own strategy.

 

By the end of this stage, you should be able to collect trading data impartially and build datasets that reflect actual probability — not interpretation.

 

We are still within the hypothesis stage of the quant pipeline. Anchor events, features, market regimes, and binary outcomes should already be defined. What remains is determining what exactly must be measured in relation to that binary hypothesis.

 

A simple “yes” or “no” outcome is not sufficient. Measurement requires structure beneath it.

​

Why Features Are Not Measurable

​

In earlier episodes, features were defined as descriptive tools used to make price action interpretable. But description is not measurement.

 

Markets are not measured by naming phenomena. They are measured by instrumenting price action through variables.

 

Consider a simple analogy.

 

If “weather” is the feature influencing the decision to go to the beach, weather itself is not measurable. You cannot measure “weather” directly. You must break it down into measurable components such as temperature, wind speed, and rainfall.

 

Only once those measurable components are defined can the influence of the feature be evaluated objectively.

 

The same applies to trading.

 

If liquidity or inefficiencies are features, they must be decomposed into variables before their influence on outcomes can be measured.

​

Feature Buckets: Structural Organisation

​

A feature bucket is an organizational layer used to group related concepts and maintain structural clarity.

 

In my strategy, there are two primary feature buckets:

 

1. Initiation Liquidity

2. Supplementary Confluences

​

Feature buckets are not mathematically required. They are organizational tools. Their purpose is to prevent logical sprawl and maintain consistency.

 

Initiation liquidity answers:

Where was the event anchor initiated?

 

Supplementary confluences answer:

What supporting behavior existed around the initiation?

 

From these buckets, variables are defined.

 

The Variable Selection Process

 

Defining variables requires discipline. The process follows five rules.

 

First, start from the scenario — not the feature. Ask: what differences could plausibly change the outcome of this specific scenario? Variables must be behaviorally relevant to the anchor event, not descriptive abstractions detached from it.

 

Second, variables must be recordable in every example. This does not mean they must be present every time. It means their presence or absence can always be logged objectively.

 

Third, variables must be mutually exclusive. A single example cannot fall into multiple categories simultaneously. If it can, probability cannot be isolated correctly.

 

Fourth, variables must be identifiable live. Anything that can only be determined retrospectively measures outcome, not condition.

 

Fifth, complexity must earn its place. If a variable does not create plausible behavioral separation, it fragments data unnecessarily and reduces sample reliability.

​

Understanding States

​

Variables sometimes require states.

 

A state is a predefined allowed value a variable can take. States prevent mixing different behaviors inside a single variable and improve comparability.

 

For example, temperature as a variable could be divided into five-degree increments. Without states, temperature becomes too granular and fragments data excessively.

 

However, states are not always required.

 

If a variable is already binary or categorical, and behavioral separation already exists at the variable level, additional states only introduce fragmentation without adding clarity.

 

In my strategy, states are not required because variables are already categorical and mutually exclusive.

​

My Variable Structure: The Pyramid

​

My system follows a simple hierarchy:

 

Feature Bucket

→ Feature Type

→ Variable

→ Final Classification

​

Initiation Liquidity

 

Feature Bucket: Initiation Liquidity

Feature Types:

 

• Swing Liquidity

• Low Resistance Liquidity

​

Variables: Timeframe-based distinctions beginning at the 15-minute timeframe (since the event anchor exists there).

 

Examples:

 

• 15-minute swing liquidity

• 15-minute low resistance liquidity

• 1-hour swing liquidity

• 4-hour swing liquidity

 

Timeframe matters because higher timeframes plausibly hold greater behavioral significance. This satisfies the behavioral relevance requirement.

 

Each variable is:

 

• Recordable

• Mutually exclusive (highest timeframe identifier used)

• Identifiable live

• Behaviorally hypothesized

​

Therefore, valid.

 

No states are added because timeframe separation already creates the necessary behavioral segmentation.

​

Supplementary Confluences

 

Feature Bucket: Supplementary Confluences

Feature Types:

 

• Fair Value Gaps

• Inverse Fair Value Gaps

​

Variables:

 

• Already Filled

• Unfilled

 

The behavioral logic is straightforward.

 

If a fair value gap has already been filled, the inefficiency has theoretically been rebalanced. An unfilled inefficiency may hold greater behavioral significance.

 

Again, each variable:

 

• Is recordable

• Is mutually exclusive

• Is identifiable live

• Has behavioral logic

​

Therefore, valid.

 

Preventing Fragmentation: The Six-Class System

​

Supplementary confluences could theoretically fragment into infinite combinations.

 

To prevent double counting and preserve behavioral meaning, I group them into six classifications:

 

1. Fair Value Gaps only

2. Inverse Fair Value Gaps only

3. Already Filled FVGs only

4. Already Filled IFVGs only

5. Conglomerate (any mixed combination)

6. None (no confluences respected)

​

Every scenario falls into one of these six classes.

 

This prevents infinite subdivision while preserving structural clarity.

 

Objective Identification: The 5-Minute Rule

​

Determining whether a confluence was respected requires consistency.

 

The 15-minute anchor is broken into 5-minute substructures. If price closes beyond a confluence on the 5-minute timeframe, that confluence is invalidated.

 

This prevents subjective interpretation and ensures only respected confluences are logged.

 

It is mechanical, repeatable, and live-applicable.

 

Market Structure Variables

 

From Episode 03, market structure context is already defined:

 

• Lack of weekly and below

• Lack of 4-hour and below

• Lack of hourly and below

• In favor

• Against

 

These remain categorical and require no additional states.

 

Full Scenario Application

​

Once market structure, initiation liquidity, and supplementary confluences are defined, the full scenario becomes measurable.

 

Example classification:

 

Initiation Liquidity: 15-minute low resistance buyside liquidity

Market Structure: Lack of weekly and below

Supplementary Confluence: Inverse FVG

Objective: Nearest opposing confluence

 

Outcome is then measured relative to the predefined binary criteria.

 

This process is repeated identically for every example.

 

No discretion. No reinterpretation.

 

Only measurement.

​

Why This Stage Matters

 

Without defined variables:

 

• Data fragments

• Behaviors mix

• Probabilities distort

 

With defined variables:

 

• Behavioral separation becomes testable

• Subsets remain large enough for statistical relevance

• Strategy components become measurable

 

This is the difference between logging trades and building a model.

​

Closing Thoughts

 

By now, you should understand:

 

• Why undefined variables corrupt data

• The difference between features, variables, and states

• How to define valid variables

• How to prevent fragmentation

• How the full variable system operates within my strategy

 

The next stage is conditioning — understanding how these variables interact and influence probability.

 

Join me in Episode 06.