Backtests
Your rules run against years of real bars, with the methodology written down so you can check it.
A backtest is only worth what its assumptions are worth. The number at the end is easy to produce and easy to inflate, so the deliverable here is the number plus everything needed to argue with it.
That means the exact data used and its resolution, how survivorship and lookahead were handled, what happens on gaps and halts, and the fills and costs assumed. Out of sample results are reported separately from in sample. Results that do not survive the split get reported as such.
No performance is promised and none is implied. What is delivered is a reproducible test, the code that produced it, and the honest output including the runs that disappointed.
What you get
- Rules implemented against real historical bars
- In sample and out of sample reported separately
- Lookahead and survivorship handling documented
- Fill and cost assumptions stated
- Reproducible code and written methodology
Backtested results are historical and are not a prediction of future results. This is engineering work, not investment advice.