More than ten years ago I wrote a forex strategy, ran it for a while, and eventually retired it. Not because it exploded — because I could no longer honestly answer the question every trader must answer: does this still work, or did I just get lucky during a period that flattered it? I never answered it, so I stopped trading it. That question has been sitting in a drawer ever since.

Two things changed. First, my tooling did: proper tick data, a real backtesting pipeline, and the scripting experience to automate what used to be weekend-long manual exercises. Second, and more importantly, I did. A decade of building production software teaches you that a system you cannot test is a system you cannot trust — and I had never once tested this strategy to a standard I would accept in code review.

The process so far is deliberately boring: pull clean historical tick data, rebuild the strategy as an expert advisor so nothing depends on memory or feel, and log every assumption in writing. Most of the work so far has been making the test honest — spread and slippage modeling, data quality checks, and refusing to peek at results until the harness is done. It turns out the boring part is the whole game.

I will write up what the validation shows once it actually shows something — including if the answer is no. An honest retirement with evidence beats a decade of wondering. Until then: no signals, no promises, and no getting ahead of the data.