Faultline H4: what it does, what it refuses to do, and what the backtests do not show
THIS IS AN H4 STRATEGY. It will run on any chart you attach it to, and since version 1.11 it reads H4 bars wherever it is attached, so the logic does not change. But it was designed, tuned and validated on H4 bars only. H4 is the timeframe this strategy works on, and the only one the published results describe.
I have published Faultline H4 on the Market. Rather than write a launch post about results, I want to describe the mechanism, the settings you control, and the things the published tests cannot tell you. If you have been on this Market for a while you have seen the other kind of launch post, and you already know how much it is worth.
WHAT THE PROGRAM DOES
Faultline H4 trades the break of a recent support or resistance level on 28 forex pairs, from a single file. It works on completed bars only. Once per new H4 bar it locates a recent swing high and swing low over that pair's own lookback, and checks whether the previous bar broke one of them.
What happens next depends on the pair. On a pair configured to fade the break, a bar that pierces the level and closes back inside it is a reversal signal and the entry is taken against the break. On a pair configured to follow the break, a bar that closes beyond the level is a continuation signal and the entry is taken with the break. Twenty of the 28 pairs fade on both sides, two follow on both sides, and six fade on one side and follow on the other. That split is not a design preference. It is what each pair's own validation supported, and I am publishing it because a seller who cannot tell you whether their product is a reversal system or a breakout system has not read it.
A candidate signal then has to clear that pair's own confirmation filters before an order is sent. Which filters apply, and the trading hours and weekday restrictions, also vary per pair.
HOW A TRADE ENDS
A stop loss and a take profit are attached when the order is opened. The stop is set at a multiple of Average True Range for that pair; the target sits at a fixed reward-to-risk ratio from the entry.
Neither is modified for the life of the trade. No trailing stop, no move to break even, no partial close, no averaging in, no time-based exit. There is no position-modify call in the program at all. A trade ends at its stop or at its target.
Position size comes from account balance and that trade's stop distance, so the money risked is intended to be the same percentage on every trade regardless of pair or stop width.
WHAT IT DOES NOT DO, AND WHY THAT IS THE ACTUAL PITCH
No martingale. A loss does not increase the next position size.
No grid, no basket, no averaging down, no recovery mode. There is one order path in the code.
No trailing, no break even, no partial closes. The bracket set at entry is the bracket the trade ends on.
No hedging within a chart. One position per chart at a time, in one direction.
No DLL calls, no web requests, no external licensing, no file writing.
No arbitrage, no tick scalping, no latency dependence. Entries are evaluated once per completed H4 bar.
Every one of those is checkable in the free Strategy Tester demo in about five minutes. That is the point of listing them. Most product pages ask you to believe something. This list asks you to go and check.
WHAT YOU CONTROL
The strategy parameters are compiled in. Per-pair levels, stop multiples, target ratios, filters and session rules are not exposed, so they cannot be changed by accident and the version you run is the version that was tested.
What is exposed: RiskPercent, MinLots, three alert switches, and the account-level exposure gate (UseExposureCap, MaxOpenPositions, MaxHeatPct, MaxPerDirection, RestrictToOwnMagic, the magic-number window, and its logging switches).
RiskPercent is the only lever that scales drawdown, and every published figure was produced at its shipped default. I have not published tests at any other setting, so if you scale it, treat the resulting expectation as your own estimate rather than as a validated result.
One default deserves calling out rather than burying: RestrictToOwnMagic ships as false, which means the exposure gate counts every position open on the account, including your own manual trades and any other Expert Advisor you run. If you trade the same account by hand, set it to true, or the gate can quietly stop this program from opening anything.
THE RESULTS, AND THE MISSING NUMBER
The listing publishes all 28 pairs individually, with profit factor, trade count, drawdown and win rate for each, from MetaTrader 5 Strategy Tester runs on real tick data over 2022-01-01 to 2026-07-28, which is 4.6 years. Each pair was tested separately, on its own 1,000 USD account at 1:100 leverage, at the shipped default risk setting, on one broker's tick history, with no commission configured.
Across those tests, 26 of the 28 pairs finished with a profit factor above 1.30, and none finished below 1.00. Per-pair maximum drawdown ranged from 9.1 percent to 35.1 percent, averaging 21.4 percent, and nine of the 28 exceeded 24 percent. Those are per-pair figures from tests run in isolation, not a portfolio drawdown.
You will notice there is no total profit figure anywhere on the page. There will not be one. Those 28 tests each ran on their own separate account, so adding them together would describe 28 accounts rather than one, and a combined figure produced from tests run that way would not describe any real account. The exposure gate that limits total open positions and total open risk cannot bind in a single-symbol test either, so the tests are also more permissive than the live configuration they represent.
WHAT THE BACKTESTS DO NOT SHOW
The 28 equity curves are correlated. Several of these pairs share a currency and move together, so simultaneous drawdowns across charts can be deeper than any single pair's figure suggests.
The results come from one broker's tick history and one spread environment, with no commission configured. Your spreads, commission, swap and execution will differ, and this strategy holds positions for days, so swap is a real cost on some pairs.
The tested window is not a full market cycle. It contains particular regimes and not others.
Parameters were selected using this same historical window, with walk-forward validation across sub-periods. That reduces the risk that some of the fit is specific to the period. It does not remove it, and I am not going to claim otherwise.
There is no live track record published with this release. Everything on the page is simulation.
AVAILABILITY
Once more, because it is the one setup detail that matters most: this is an H4 strategy. It will run elsewhere, but it was only ever validated on H4. Put it on an H4 chart.
Faultline H4 is on the Market at 149 USD, or rentable at 49 USD for a month if you would rather test it on your own broker's data first. There is also a free demo you can run in the Strategy Tester, and I would rather you ran it before buying than after.
Questions in the comments are welcome, including hostile ones. I will answer what I can. I will not publish the parameter values, and I will not tell you which specific pairs fade the break and which follow it.
RISK DISCLAIMER
Trading foreign exchange on margin carries a high level of risk and may not be suitable for all investors. The high degree of leverage can work against you as well as for you. You could sustain a loss of some or all of your initial investment.
All results referenced above are from historical simulation in the MetaTrader 5 Strategy Tester. Past performance, whether historical or simulated, does not indicate future results. A backtest describes what a set of rules would have done on one broker's recorded history, not what they will do in future markets.
