We Backtested It: The VP Strategy

The volume profile rules from the Reel, coded and tested on 21 months of US30 — why they lost, what we changed, and how we tested without fooling ourselves. This free lesson covers: Rules precise enough for code, The result: it lost, Work it out, Why: the fills were the wrong days, Which days filled?, How not to fool yourself, Why hide 2026?, What changed, and what each did (2025), The one unseen test, Reading it honestly, What to know before using it, Put it together.

Academy

0 XP

Strategy Playbook · Learn · Step 1 of 12

Rules precise enough for code

"Buy the pullback to the POC in an uptrend" is not testable until every word is defined. We used yesterday's volume profile, and called it an uptrend when the first 1H candle after the 9:30 New York open opened above yesterday's VAH (below VAL for shorts; inside the value area meant no trade).

Entry: a limit order at yesterday's POC. Stop: 5 points beyond the opposite edge of the value area. Targets: half at VAH, the rest at yesterday's high. One trade a day, 2 points of cost on every trade.

Going deeper

More on we backtested it: the vp strategy

All figures come from TradingView's Strategy Tester running the Candle Decoder Pine script on Capital.com's US30 CFD, 1H chart, with 2 points of cost per trade taken off every result in R.

The ground rules were fixed before the design runs. One was adjusted before any variant was tested: the minimum trade count was lowered from 50 to 40, because one year of 1H data produced only 44 trades for the original rules. Changing a rule after seeing which variants pass would have been fitting; changing it before was not.

Three trades from the 2026 test were checked by hand against the plotted levels — entry, stop and exit all matched the rules. That check found one detail worth disclosing: on the 1H chart, end-of-day exits happen at 17:00 New York rather than 16:00, because the script detects the end of the window on the first candle after it.

Common mistakes4

Reporting the best of many variants as if it were the only one tried.
Re-running the unseen data after making a change, which quietly turns it into design data.
Judging a result on fewer trades than the variance allows — under fifty trades, most results mostly measure luck.
Leaving costs out. At 2 points per trade they took a meaningful share of the edge here.

From experience2

Write the rules and the pass/fail criteria down before the first run, and keep the list of every run — including the ones that made things worse.
After a backtest, forward-test on a demo account or at minimum size. It is the only data that cannot have been fitted.

Questions people ask3

Why only 21 months?

That is how much 1H history TradingView loaded on the plan we tested with. A longer test needs exported data or deeper history, and would make the result more convincing.

So does the strategy work?

The original rules did not. The changed version was slightly positive on data it had never seen — +4.2R over 47 trades — which is encouraging but not proof. Forward testing is the next step.

Can I run the backtest myself?

Yes — the rules in the final steps are complete. Code them in any backtester, or ask us for the Pine script, and check the numbers on your own broker's data.