01
Source integrity
The exact Pine version is frozen and identified before measurement.
Private validation · $99 pilot
Know whether to deploy, revise, or drop it.
One frozen Pine strategy goes through the same public rulebook as the catalog: 18 out-of-sample market/timeframe cells, fees on every fill, bias checks first, and performance against buy-and-hold.
Price
$99 one time
Delivery target
1 business day
Privacy
Private by default
Forward record
30 days included
The delivery target starts after we confirm the source, scope and payment. If the script is outside the pilot’s fixed scope, no payment is taken.
Evidence stays separate, so one good number cannot hide a failed guard.
02
18-cell OOS grid
Six markets × three timeframes, using windows not selected by the strategy.
03
Trading frictions
A 0.05% fee is applied per side on every fill; assumptions remain visible.
04
Bias checks
Repaint and look-ahead surfaces are reported before performance is interpreted.
05
Decision memo
Deploy, forward-test, revise or drop — with the evidence for that call separated.
06
Files you keep
A versioned private report and CSV-level cell results, not a screenshot alone.
Real report shape · public catalog example
See the decision before you submit.
This is measured catalog data, not an invented success story. Your report uses the same separation of integrity, edge, benchmark and tracking, while remaining private.
Audit WA-001 · public sample
Golden Cross (SMA 50/200)
version frozen · 18 cells · out of sample
Decision
47/100 evidence
Do not deploy broadly. Forward-test only where the edge appeared.
Positive risk-adjusted return in 16 of 18 cells — strongest on SPX500 · 1H (Sharpe 1.84); 2 cells lose after the 0.05% per-side fee. Median cell trails buy-and-hold by 16.9pp; 5 of 18 cells beat it.
Inspect the complete public row →Code integrity
Pass
OOS Sharpe
0.78
Positive cells
16 / 18
vs buy & hold (median cell)
-16.9pp
Fill model
Documented
Forward record
Day 0
Repaint / look-aheadStatic checks passed for this reference implementation
Market stability16 of 18 cells produced positive risk-adjusted return
Benchmark5 of 18 cells beat buy-and-hold; the median cell trailed it by 16.9pp
After delivery · 30 days included
The backtest stops moving. The forward record starts.
- 01
Freeze
The audited source receives a version ID. Any later code change is a new version, not a rewrite of history.
- 02
Observe
Future bars are recorded after the audit date and never mixed into the backtest window.
- 03
Compare
At day 30, the report shows backtested expectation beside the untouched forward result.
Not ready for a full audit?
Run the six-market quick check first. It is free, browser-side and intentionally does not pretend to be an audited verdict.
Start free quick check →Not investment advice. Audit results are research measurements, not recommendations to trade.