Skill

expression-building

Build and validate WorldQuant BRAIN alpha expressions. Use when creating new alpha expressions, applying templates, or debugging expression errors. Triggers on "build expression", "create alpha", "expression template", "alpha formula", "build alpha", or "generate expression".

npx claudepluginhub zxx264547/worldquant-brain-research --plugin alpha-mining

Popularity

Parent stars

Invocation

How this skill is triggered — by the user, by Claude, or both

Slash command

/alpha-mining:expression-building

User invocable

Model invocable

Inline context

Default effort

Context Preview

The summary Claude sees in its skill listing — used to decide when to auto-load this skill

Build valid BRAIN alpha expressions following platform operator rules.

SKILL.md

47 lines · ~448 tokens

Similar Skills

cache-components

139.5k

Guides Next.js Cache Components and Partial Prerendering (PPR): 'use cache' directives, cacheLife(), cacheTag(), revalidateTag() for caching, invalidation, static/dynamic optimization. Auto-activates on cacheComponents: true.

cache-components

docx

135.8k

Creates, reads, edits, and analyzes .docx files using docx-js for new documents, pandoc for text extraction, Python scripts for XML unpacking/validation/changes, and LibreOffice for conversions.

20 files

document-skills

Stats

LanguagePython

Parent stars2

MaintenanceGood

Last CommitMay 17, 2026

Actions

View Source View Plugin View on GitHub View README

Help us improve

Share bugs, ideas, or general feedback.

Stats

Actions

Help us improve

Share bugs, ideas, or general feedback.

Alpha Expression Building

Build valid BRAIN alpha expressions following platform operator rules.

Core Rules

Incremental Complexity

0-op: rank(), zscore() — bare signal

1-op: ts_mean(), ts_delta(), ts_decay() — single time-series transform

2-op+: ts_rank(ts_delta()), signed_power(ts_sum()) — nested transforms

Start at 0-op, only add complexity when baseline shows signal

Time Windows

Only use: 5, 22, 66, 120, 252, 504

Normalization

Fundamental/Volume data MUST be wrapped in rank() first

Price data can use raw values

VECTOR data MUST use vec_min/vec_max/vec_avg/vec_sum (not ts_ operators directly)

VECTOR Data + vec_min/vec_max Same-Direction Rule

Outer ts_max/ts_arg_max → inner vec_max

Outer ts_min/ts_arg_min → inner vec_min

zscore() wrapper can double Sharpe

Batch Rule

Always create 8 alphas per create_multiSim call.

Validated Patterns

Pattern

Best For

Example Sharpe

zscore(-ts_max(vec_max(field), W))

VECTOR lending/risk data

2.18

rank(ts_mean(field, W))

MATRIX fundamental data

0.85

signed_power(zscore(-ts_max(vec_max(field), W)), 10)

Low Prod Corr

1.61

ts_backfill(ts_sum(field, 252), 3)

Quarterly EPS

1.17

Environment

Python: /home/zxx/wq_env/bin/python

Scripts: worldquant_brain/scripts/

Engine: worldquant_brain/engine/backtest_runner.py

Alpha Expression Building

Build valid BRAIN alpha expressions following platform operator rules.

Core Rules

Incremental Complexity

0-op: rank(), zscore() — bare signal
1-op: ts_mean(), ts_delta(), ts_decay() — single time-series transform
2-op+: ts_rank(ts_delta()), signed_power(ts_sum()) — nested transforms
Start at 0-op, only add complexity when baseline shows signal

Time Windows

Only use: 5, 22, 66, 120, 252, 504

Normalization

Fundamental/Volume data MUST be wrapped in rank() first
Price data can use raw values
VECTOR data MUST use vec_min/vec_max/vec_avg/vec_sum (not ts_ operators directly)

VECTOR Data + vec_min/vec_max Same-Direction Rule

Outer ts_max/ts_arg_max → inner vec_max
Outer ts_min/ts_arg_min → inner vec_min
zscore() wrapper can double Sharpe

Batch Rule

Always create 8 alphas per create_multiSim call.

Validated Patterns

Pattern	Best For	Example Sharpe
`zscore(-ts_max(vec_max(field), W))`	VECTOR lending/risk data	2.18
`rank(ts_mean(field, W))`	MATRIX fundamental data	0.85
`signed_power(zscore(-ts_max(vec_max(field), W)), 10)`	Low Prod Corr	1.61
`ts_backfill(ts_sum(field, 252), 3)`	Quarterly EPS	1.17

Environment

Python: /home/zxx/wq_env/bin/python
Scripts: worldquant_brain/scripts/
Engine: worldquant_brain/engine/backtest_runner.py