# ProEA Lab: monthly bootstrap arithmetic v1.0

Published: 2026-09-08. Canonical citation: https://proea.app/en/research#monthly-bootstrap
Article: https://proea.app/en/blog/one-system-ten-thousand-years

## Run without a ProEA account or website checkout

Extract the ZIP, open a terminal in its folder and run with Node.js 22 or newer:

    node verify.mjs
    node analysis.mjs

No npm install, external requests, market feed or paid software is required.
The analysis prints its results and creates two SVG figures in ./output.
verify.mjs checks bundled file hashes and every line of expected-output.txt.
Checksums check package integrity against this manifest; they are not independent attestation.

## Exactly what this reproduces

Arithmetic and charts in the linked monthly-bootstrap article, using the 24
already-published rounded gain_pct values, NOT returns recomputed from balances.
Each underlying input is a separate one-month MT5 Strategy Tester simulation
starting from USD 10,000 at 1:100 leverage: Model=0 generated ticks, ForwardMode=0,
January 2024 through December 2025. There are 18 positive and 6 negative months;
the worst is -46.41% and the best +422.38%. This is historical simulated data.
It is not a continuous account, live trading, real-tick verification, or proof of
out-of-sample walk-forward validation. This package cannot re-run those MT5 tests.
The paid EA, broker history and underlying deal ledger are not in this package.

10,000 synthetic 12-month paths draw independently with replacement. The PRNG
is mulberry32, main seed 7, sensitivity seed 8 after excluding the best month.
Percentiles select sorted index floor(q * (n - 1)). Returns are compounded as
if they scale proportionally; the source tests reset their deposit each month.
Drawdowns use month-end values only and omit within-month floating losses.
The thin 24-month sample omits unseen losses, changing regimes, dependence and
execution conditions. Simulation frequencies are not future probabilities.
The original sample is selected and includes an extreme positive outlier;
read the sensitivity result and loss metrics alongside any return statistic.

## Authorship, use and corrections

ProEA Lab is the publishing organization and sells related software, including MTR.
This is our own study, not third-party certification. Contact support@pulltrade.app
with the package version, command and discrepant output for a correction.
When citing, include version 1.0, publication date, canonical URL and the fact that
these are generated-tick monthly simulations, not live results or a forecast.

Permission is granted to use, copy, modify and redistribute the original code,
summary CSV, documentation and generated figures in THIS package with attribution
to ProEA Lab and this notice. Provided as-is, without warranty. This permission
does not license the MTR EA, third-party broker history, or other ProEA products.