backtest-determinism-and-reproducibility SKILL.md skillA

backtest-determinism-and-reproducibility from himanshuj16/algo-trading-skills is an agent skill: a SKILL.md folder an AI agent loads when a task calls for it.

Install it, pinned
mdr add himanshuj16/algo-trading-skills/backtest-determinism-and-reproducibility@v2.0.0

Pins v2.0.0 in mdr.lock. Pin the fingerprint instead.

Audit grade
A, 17 of 17 checks passed
Kind
SKILL.md skill
Repository
himanshuj16/algo-trading-skills
Path
skills/backtest-determinism-and-reproducibility/SKILL.md
Versions recorded
5
Latest version
v2.0.0, committed Sep 3, 2026
First version recorded
Jul 30, 2026
Size
7,790 bytes
Repository stars
50
License
Apache-2.0

Use when identical backtest code and data produce different P&L or Sharpe across runs, making optimisation untrustworthy: deterministic seeding, strict input sort order, a simulated clock replacing wall-clock reads, and run-to-run divergence detection.

Collected from public GitHub. Every version is saved exactly as it was and never changes. Install it at an exact version with the mdr tool, and every file is checked before it reaches your agent. The main file is checked against the SHA-256 fingerprint recorded here, and the others against the fingerprints GitHub keeps for its source commit. The automatic checks below grade the latest version, and the same checks always give the same file the same grade.

What the file says

## When to Use

Invoke this skill when building backtest simulation engines or CI regression suites. If backtest runs produce varying P&L curves or Sharpe ratios across identical code and data, parameter optimization becomes impossible to trust — you cannot tell an improvement from noise. Non-determinism comes from unseeded RNGs, set iteration order, unordered event streams, wall-clock reads, and floating-point accumulation order.

## What This Can and Cannot Guarantee

**It cannot guarantee bit-identical results in general, and no tool can.** PyTorch's reproducibility guidance states plainly that "completely reproducible results are not guaranteed across PyTorch releases, individual commits, or different platforms," and that results may differ between CPU and GPU "even when using identical seeds." The same holds for BLAS thread counts, SIMD code paths, and library versions.

Determinism is a property of a **pinned environment** — same platform, same library versions, same thread configuration. This skill does two things within that envelope:

1. **Eliminates the controllable sources**: seeding, event ordering, clock reads.
…

Read the whole file at its exact version.

How to install

The exact contents, by fingerprint
mdr add himanshuj16/algo-trading-skills/backtest-determinism-and-reproducibility@sha256:18abbeaf1c3a9a9a

The line at the top of this page pins the label v2.0.0, which is what most people want. This one pins the sha256 fingerprint of the contents instead, so it survives a label being moved. Either way the exact version is written to mdr.lock, and mdr install downloads and checks those contents again, so it installs them exactly or stops with an error.

Badge

mdr badge

[![mdr](https://markdownregistry.com/badge/art_gknvpcvflepgbd3b.svg)](https://markdownregistry.com/a/art_gknvpcvflepgbd3b)

2 badge image fetches in 30 days. What the badge shows.

Versions

The registry has recorded 5 versions of this file, between Jul 30, 2026 and Sep 3, 2026, about one every 9 days.

versioncommittedcommitsizeaudit
v2.0.0 latest2026-09-03 8527a51 7,790 BA view · diff
v2.0.02026-09-03 64e8b01 7,761 BA view · diff
v2.0.02026-08-13 097e217 7,720 BA view · diff
v1.0.02026-08-07 89d1436 3,236 BA view · diff
v1.02026-07-30 dfc60c7 3,234 BA view

Audit of the latest version

A  17 of 17 checks passed. Fixed automatic checks with no AI involved, so the same checks give the same answer every run.
  • pass: Frontmatter block present
  • pass: Frontmatter declares a name
  • pass: Frontmatter declares a description
  • pass: Size between 200 bytes and 200 KB (7790 bytes)
  • pass: No zero-width or bidi control characters
  • pass: No instruction hidden inside an HTML comment
  • pass: No link to an exfiltration or paste host
  • pass: No credential-shaped string
  • pass: No instruction to send local credentials anywhere
  • pass: No text hidden with inline styles
  • pass: No prompt-injection phrasing
  • pass: No curl or wget piped into a shell
  • pass: No recursive delete of root, home or parent
  • pass: No instruction to read or print local credentials
  • pass: No base64 blob over 200 characters
  • pass: No link to a raw IP address
  • pass: No script tag

Source

GitHub

himanshuj16/algo-trading-skills · 50 stars · license Apache-2.0 · pushed 2026-09-14 · branch main

API

GET https://markdownregistry.com/api/v1/artifacts/art_gknvpcvflepgbd3b
GET https://markdownregistry.com/api/v1/resolve?ref=himanshuj16/algo-trading-skills/backtest-determinism-and-reproducibility
GET https://markdownregistry.com/api/v1/blob/18abbeaf1c3a9a9ad8d5c156136e3fa6c64cf0cf908e63d34134c334c4150980

Your AI assistant does the legwork. You hear about the deals worth your word. Hand yours the standing instructions at modelranch.com and it joins the network that reads files like this one.

Guides: the SKILL.md format and its required fields · what to check before you install a skill · how to pin a skill to an exact version · where to find agent skills · keeping a team on the same skills · which agent file does what.

More from himanshuj16/algo-trading-skills

algo-trading-skills Cursor rule
himanshuj16/algo-trading-skills · .cursor/rules/algo-trading-skills.mdc · Automatically discover and load algorithmic trading skills from the library in skills/
git:20260903.8527a51 · audit A · 50 stars
AGENTS.md AGENTS.md file
himanshuj16/algo-trading-skills · AGENTS.md
git:20260903.327fc0d · audit A · 50 stars
CLAUDE.md CLAUDE.md file
himanshuj16/algo-trading-skills · CLAUDE.md
git:20260907.26c257b · audit A · 50 stars
llms-full.txt llms.txt file
himanshuj16/algo-trading-skills · llms-full.txt
git:20260903.327fc0d · audit A · 50 stars
llms.txt llms.txt file
himanshuj16/algo-trading-skills · llms.txt
git:20260903.8527a51 · audit A · 50 stars
adaptive-batch-size-tuning-under-load SKILL.md skill
himanshuj16/algo-trading-skills · skills/adaptive-batch-size-tuning-under-load/SKILL.md · Use when writing high-volume ticks or order logs into TimescaleDB, ClickHouse or Kafka and one static batch size either…
v2.0.0 · audit A · 50 stars
adaptive-execution-under-volatility-spikes SKILL.md skill
himanshuj16/algo-trading-skills · skills/adaptive-execution-under-volatility-spikes/SKILL.md · Use before each child-order decision in a TWAP, VWAP or POV schedule when a volatility spike should bound…
v1.2.0 · audit A · 50 stars
adaptive-sampling-under-extreme-tick-rates SKILL.md skill
himanshuj16/algo-trading-skills · skills/adaptive-sampling-under-extreme-tick-rates/SKILL.md · Use when a trade-tick stream exceeds a consumer's measured processing ceiling and you must emit 1:N samples that still…
v1.2.0 · audit A · 50 stars
adjusted-vs-unadjusted-price-series-pitfalls SKILL.md skill
himanshuj16/algo-trading-skills · skills/adjusted-vs-unadjusted-price-series-pitfalls/SKILL.md · Use before loading historical OHLCV into a backtest, to declare and verify the adjustment mode; separates raw…
v1.3.0 · audit A · 50 stars
adversarial-robustness-of-trading-signals SKILL.md skill
himanshuj16/algo-trading-skills · skills/adversarial-robustness-of-trading-signals/SKILL.md · Use before promoting an ML trading signal, to measure how often epsilon-bounded feature perturbations flip its output…
v1.3.0 · audit A · 50 stars
adverse-selection-measurement-for-passive-orders SKILL.md skill
himanshuj16/algo-trading-skills · skills/adverse-selection-measurement-for-passive-orders/SKILL.md · Use when passive fills keep happening just before the market moves against you, measuring forward markouts in basis…
v1.3.0 · audit A · 50 stars
air-gapped-signing-workflow-for-cold-storage SKILL.md skill
himanshuj16/algo-trading-skills · skills/air-gapped-signing-workflow-for-cold-storage/SKILL.md · Use when moving crypto out of cold storage and the signing key must never touch an internet-connected machine; models…
v2.0.0 · audit A · 50 stars

Every file in himanshuj16/algo-trading-skills

Browse by kind, by grade A, or by owner.