An Expert Advisor trouble shooting & optimization needed

Hi, I have developed an MT5 Expert Advisor called SRX which latest version is R24b47, and I am looking for an experienced MQL5/MT5 EA developer to perform a thorough technical investigation, fixing and optimization of the EA.

1. The main problem

SRX is not stable across different market periods.

It can perform extremely well during one period, but the same EA/configuration can perform very poorly or approach Margin Call (MC) during another period.

My objective is therefore not simply to maximize backtest profit.

I want to understand why the EA is unstable and whether the underlying logic can be made significantly more robust across different market conditions.

2. Current EA and testing environment

• EA: SRX R24b47

• Platform: MetaTrader 5

• Symbol: XAUUSD

• Broker: versus

• Starting capital: $100

• MC = Margin Call

3. Current test observations

I tested SRX over Week 1, Week 2, Week 3 and Week 4 of April.

Configuration A

Engulf OFF

Results:

• Week 1: +$1,800+

• Week 2: +$200

• Week 3: almost MC

• Week 4: almost MC

So although Week 1 can be extremely profitable, the EA is not able to maintain that performance across the following weeks.

Configuration B

Engulf ON / CR OFF / MBO OFF

Results:

• Week 1: +$1,100+

• Week 2: survives

• Week 3: survives

• Week 4: almost MC

This configuration appears to improve survival compared with Configuration A, but the EA still becomes highly vulnerable in Week 4.

4. Development history

SRX has already gone through a substantial development process.

I spent approximately 3 months developing the EA with Claude.

After that, I spent approximately 1 additional month working on fixes/improvements with GPT Plus.

Unfortunately, the latest development/fixing work produced no meaningful improvement over the existing EA. The fundamental instability remains.

Therefore, I am now looking for a fresh and independent technical investigation, rather than simply continuing the previous optimization approach.

5. What I want from the developer

I need the developer to first understand and audit the existing SRX R24b47 code and its trading logic.

I want to determine:

• Why SRX performs extremely well in some periods but approaches MC in others.

• What specifically causes the deterioration from one week to another.

• Whether there are weaknesses or conflicts in the trading logic.

• Whether entries are becoming too aggressive under certain market conditions.

• Whether trade management, SL/TP, position sizing, recovery logic, trade stacking or other mechanisms contribute to the problem.

• Whether different components of the EA interact in an unintended way.

• Whether certain filters or conditions are too sensitive to particular market conditions.

• Whether there are programming/implementation issues that affect the intended strategy.

• Whether the EA is effectively over-fitted to certain market behavior.

I am not assuming that there is a bug. I want the developer to investigate objectively and determine the actual cause.

6. Please do not simply curve-fit the EA

This is very important.

I do not want the developer to simply run MT5 optimization and find a set of parameters that produces a spectacular historical result.

A configuration that makes $1,800+ in Week 1 but approaches MC in Week 3 or Week 4 is not the solution.

I would prefer a system that produces more moderate but substantially more consistent results across different periods.

The priority should be:

Robustness → Survival → Consistency → Risk control → Profit

rather than:

Maximum historical profit.

7. Existing strategy

I already have a working EA and its intended trading concept.

I am not asking for a completely new trading strategy at this stage.

The first objective is to understand and improve SRX R24b47.

However, I am open to modifying, replacing or removing specific components if the technical investigation shows that they are responsible for the instability.

8. Testing approach

I would like the improved version to be tested across multiple different periods, not just the period that was used to develop the changes.

The purpose is to determine whether an improvement is genuinely robust or simply curve-fitted.

I am particularly interested in testing periods where the current SRX performs well and periods where it performs poorly.

9. What I can provide

I can provide:

• The complete SRX R24b47 .mq5 source code

• My explanation of the EA and its intended behavior

• The testing periods

• The configurations/settings used

• Any additional information or test results that the developer specifically requests 

I am not a programmer, so I may not be able to provide a technical explanation of the underlying problem. I expect the developer to investigate the code and test results and explain the findings to me.

10. What I expect from the developer

Before making major changes, I would like the developer to explain:

1. What they believe is causing the instability.

2. Which part(s) of the EA they believe are responsible.

3. What changes they propose.

4. Why those changes should improve robustness.

5. How they intend to test whether the changes actually work.

I am looking for a developer who is comfortable diagnosing a complicated existing EA, not just someone who specializes in parameter optimization.

The ultimate goal is to make SRX more robust across different XAUUSD market conditions and historical periods, while preserving its intended strategy as much as reasonably possible.

If additional information, configurations, or test periods are required, I can provide them upon request.