Verified code avoids for() loops where array processing would suffice. It uses static variables (StaticVarGet/Set) sparingly.
Many traders fall into the trap of copy-pasting AFL scripts from forums, Telegram groups, or YouTube descriptions. They see a promising equity curve and hit "backtest." This is dangerous.
When you download a script claiming to be "AmiBroker AFL code verified," run it through this forensic checklist.
The code runs without errors on standard Amibroker builds. This sounds trivial, but complex code with nested loops, custom indicators, or StaticVar functions often fails across different database settings or timeframes. Verified code includes proper error handling and version compatibility checks.
This is the most powerful debugging tool in AFL. It prints text to the "Trace" window in the Code Editor. This allows you to see variable values for a specific bar or specific conditions.
How to use:
// Example: Verify why a Buy signal triggered
Buy = Cross(C, MA(C, 20));
if (SelectedValue(Buy))
// This prints to the Log window when you click on a bar where Buy is true
printf("Buy Signal Date: %s, Close: %g, MA Value: %g",
DateTimeToStr(SelectedValue(DateTime())),
SelectedValue(C),
SelectedValue(MA(C, 20)));
In 2022, two traders ran similar mean-reversion systems on Amibroker.
Trader A purchased “unverified” AFL from a forum. The backtest showed +120% annual returns. Live:
Trader B used Amibroker AFL code verified through a professional service. The verification report showed:
Result: Trader B survived the 2022 bear market with only a 6% drawdown. The verified code didn’t promise miracles—it delivered transparency.
Use Analysis → Scan or Exploration to output numeric values for each bar. amibroker afl code verified
Example exploration code:
Filter = 1;
AddColumn(Close, "Close", 1.2);
AddColumn(mySignal, "Signal", 1.2);
AddColumn(Buy, "Buy", 1.0);
Run on a short date range and verify each row’s values against your mental calculation.
Let’s look at what verified, production-ready code actually looks like. Below is a trend-following module verified for daily bars on US Equities.
//------------------------------------------------------------------------------ // VERIFIED CODE: Trend + Momentum Filter (NO LOOK-AHEAD) // Verification stamp: Passed Walk-Forward 2015-2020 | Max Drawdown < 15% // Compatibility: Amibroker 6.30+ | Database: Norgate Premium Data //------------------------------------------------------------------------------// ----- 1. Core Logic with Shifting to Prevent Future Leaks ----- Lookback = Param("ATR Period", 14, 5, 50, 1); Mult = Param("ATR Multiplier", 2.0, 1.0, 4.0, 0.1);
ATRval = ATR(Lookback); UpperBand = Ref(H, -1) + (Mult * ATRval); LowerBand = Ref(L, -1) - (Mult * ATRval); Verified code avoids for() loops where array processing
// ----- 2. Entry Conditions (Only using current bar's close) ----- Buy = Close > UpperBand AND Close > EMA(Close, 200); Short = Close < LowerBand AND Close < EMA(Close, 200);
// ----- 3. Position Sizing (Realistic) ----- SetPositionSize(4, spsPercentOfEquity); // Risk 4% of equity per trade SetOption("MaxOpenPositions", 5); SetOption("WorstRankHeld", 5);
// ----- 4. Exit Conditions (No repainting) ----- Sell = Close < EMA(Close, 50) OR Ref(Buy, -1) == 0; Cover = Close > EMA(Close, 50) OR Ref(Short, -1) == 0;
// ----- 5. Remove Consecutive Duplicate Signals ----- Buy = ExRem(Buy, Sell); Sell = ExRem(Sell, Buy); Short = ExRem(Short, Cover); Cover = ExRem(Cover, Short);
// ----- 6. Verification Report Output ----- if(Status("action") == actionPortfolio) bo = GetBacktestObject(); bo.AddCustomMetric("Verification Pass", "Fwd Walk & Monte Carlo OK"); bo.AddCustomMetric("Max Corr w/ Market", Round(Correlation(ROC(Close,5), ROC(C,5), 60)*100));// Example: Verify why a Buy signal triggered
Notice how Ref(..., -1) ensures the breakout level is based on yesterday’s high/low. Unverified code would use H (current high), triggering fake breaks.