Project specifications
- Market
- GBPUSD
- TF
- M30
- Platform
- MT5
- Client type
- Professional trader
- Delivery window
- 7 weeks
- Stack
- MQL5, Python scripts, a heat map
- Strategy
- Parameter sweep + stability map
- Modules
- Two-parameter map + warning threshold
- Testing
- Grid sweep on held-out data
Who this was for
Leila Sadr
- Role
- Strategy researcher
- City
- Isfahan
- Project reference
- TBT-1410-36
The client's idea
Leila Sadr works GBPUSD on M30; the opening brief was: I accept the optimum, but tell me how far the result falls if the parameters shift a little.
Initial challenge
Optimisation produced a single good point with no way to tell a real peak from a random bump in the data.
Solution
A grid sweep over two key parameters, a stability map, and a report of how much the result falls for a small shift in each parameter.
Why this design
The parameters were sorted by interdependence so the map carries a readable meaning. It matches how GBPUSD is actually traded and can be tested against the data already available.
Safeguards and risk control
No parameter was optimized on the whole set: part of the data was held back for out-of-sample testing.
Our next-step proposal
The next step is running the selected build on demo and reviewing weekly deviation from the test.
Portfolio
Measured results
These are the figures the test produced — not a promise about future performance.
- 48 Configurable inputs
- 8 scenarios Scenario test coverage
- 4 years Test window
- 97 % Runtime stability in test
Portfolio
How the delivery ran
Every project moves through these five phases, with the window for each agreed in the specification.
-
Discovery and problem framing
CompletedReviewing the client's current test method and the quality of the available GBPUSD data.
Day 2–6
-
Written specification
CompletedFixing the assumptions: spread, commission, slippage and the tick model used.
Day 7–15
-
Implementation
CompletedPreparing the test environment, the run scripts and results that can be compared.
Day 16–33
-
Testing and validation
CompletedControlled optimization, out-of-sample testing and forward testing on fresh data.
Day 34–43
-
Handover and support
CompletedA report separating the optimized part from the robust part, plus the remaining risks.
Day 44–47