I’d rather lose the upsell than hand you a number you’ll trade on

Why this is the whole business

Almost every broken backtest is broken in the strategy’s favour. That is not a coincidence — it is selection. Nobody keeps debugging an engine that shows a loss; they keep debugging until it shows a profit, then stop. The bugs that flatter the result are the ones that survive.

So the useful skill is not writing the backtest. It is knowing the specific ways a backtest lies, and going looking for them on purpose: pivots that were not yet confirmed, limit orders that filled at prices no broker would have given you, spread quietly set to zero, and parameters tuned on exactly the data used to score them.

I check for all four on every engagement and I write down what I find, including when what I find is that the edge is not there. A client who hears that from me is annoyed for an afternoon. A client who does not hear it finds out later, with money on the line.

What you actually get

Commented source code you can read and modify without me. Every assumption I made, written out in plain English. A results section, and next to it a section titled “what this does not prove”. If you want to see exactly what that looks like before you spend anything, the Dealing Range Theory case study is the real thing, published in full — limitations included.