Optimize My MT5 Indicator – 5 Progressive Versions (Heiken Ashi Removal, Speed Layers, Full Signal Support)

📋 Project Description:

I have a MetaTrader 5 custom indicator named NIKA_Backtest.mq5 , which calculates a smoothed signal using linear regression and EMA smoothing.

This indicator is used in Strategy Tester alongside RunwiseFX, and its logic depends entirely on the crossing of two values:

• The white line → linear regression result ( IndicBuffer[] )

• The orange line → smoothed EMA of the white line ( SmaIndicBuffer[] )

The strategy does not need live tick updates. Signals only matter after a bar closes, so all logic should be optimized to run only on closed bars ( i ≥ 1 ).

🎯 Project Goal:

Deliver 5 progressively optimized versions of the indicator, each with faster execution in Strategy Tester — while always preserving signal accuracy and compatibility with RunwiseFX's CaptureBuffer(...) .

✅ All Versions Must:

• Use only closed-bar values ( i >= 1 , skip i == 0 )

• Keep both white and orange lines fully functional

• Output buffers must be usable in RunwiseFX

• Compile cleanly in MetaEditor (MT5 2024+)

• Avoid ObjectCreate() unless specifically required

🔢 Versions to Deliver:

✅ Version 1 – Inline Heiken Ashi (No Visual Cleanup Yet)

• Remove iCustom(..., "Heiken_Ashi")

• Re-implement Heiken Ashi calculations inline inside OnCalculate()

• Keep all existing chart visuals and rectangles

• Runs once per closed candle only

💡 Why faster: Removes nested iCustom() calls and reduces inter-indicator overhead

✅ Version 2 – Heiken Ashi Inline + Visual Cleanup

• Based on Version 1

• Remove all ObjectCreate() , rectangles, and subwindow visuals

• Only keep the two signal plots: white and orange lines

💡 Why faster: Skips all chart object creation and rendering logic that slows Strategy Tester

✅ Version 3 – Light Optimization

• Based on Version 2

• Replace iMA() with inline EMA formula (no indicator handles)

• Remove redundant array resizing and avoid excessive memory use

• Simplify smoothing logic while maintaining output

💡 Why faster: Eliminates overhead from iMA() handles and reduces RAM operations

✅ Version 4 – Heavy Optimization

• Based on Version 3

• Fully implement prev_calculated logic

• Process only new bars ( start = prev_calculated > 1 ? prev_calculated - 1 : 1 )

• Optimize all loop ranges and minimize floating-point operations

💡 Why faster: Avoids recalculating old bars; dramatically speeds up long backtests

✅ Version 5 – Absolute Max-Speed with Full Signal Logic

• Remove Heiken Ashi completely

• Remove all non-essential buffers and visuals

• Use only:

  • IndicBuffer[] (white)

  • SmaIndicBuffer[] (orange)

• Keep prev_calculated and inlined math

• Use DRAW_NONE if needed to disable visual overhead

• 100% RunwiseFX-compatible

💡 Why fastest: Only the critical logic is calculated, with minimal code per bar and near-zero memory waste — optimized for 100k+ combinations and 128-thread servers


EXTRA: (SEE SCREENSHOT)

PLEASE NOTE THAT VERSION 4 and 5
I want an extra version with different price scale

So 1 version 4. with old scale
1 version with new scale
Same for version 5