Development History
Development Timeline
This is not a company history. It is how a trading system actually gets built — including the parts that did not work.
Every EA Starts With a Question.
- Can this behaviour be quantified?
- Can this risk be controlled?
- Can this logic survive different market conditions?
Then we test. And test again. And rebuild.
From Idea to Algorithm.
Process
The loop we run on every system
-
Idea
A behaviour is noticed in the market. At this point it is nothing more than an observation and a question: could this be written down as a rule?
-
Research
The observation is tested against history, across volatility regimes and sessions, before a single line of trading code is written. Most ideas end here.
-
Prototype
The rule is implemented in its crudest possible form — entries only, no risk layer, no optimisation. The goal is not profit; it is to find out whether the behaviour is measurable at all.
-
Backtest
Tick-level testing over multiple years and market conditions, with spread and commission modelled. The test conditions are recorded at the same time as the result — a number without its conditions is not evidence.
-
Failure
Most prototypes fail here, and this is the most valuable stage in the loop. A model that only worked because it described its own sample is identified and discarded rather than optimised into looking correct.
-
Improvement
The logic is rebuilt around what the failure exposed — usually a definition that was too loose. The risk layer is designed here, before the entry logic is finalised, not bolted on afterwards.
-
Optimisation
Parameters are explored systematically, then judged by the stability of the surrounding region rather than by the single best result. A peak that collapses when a parameter moves one step is not a setting; it is an artefact.
-
Forward Test
The system runs on data it has never been fitted to, under live execution conditions. This is where backtest and reality are allowed to disagree, and where that disagreement gets recorded.
-
Release
Published on MQL5 with its documentation: what it does, what it risks, and the conditions its results were produced under. A release without that documentation is not a release.
-
Update
Live behaviour, user reports and changing market structure feed back into the loop. Every version is traceable to its source, and the loop starts again at research.
Record
Failure is part of the record
A research brand that only shows wins is a marketing brand. Discarded models, rebuilt logic and abandoned assumptions are listed here because they are the reason the surviving systems exist.
Systems we have shipped.
Each Expert Advisor is documented with its concept, logic, risk model and test conditions. No black boxes, no unexplained results.