kalmanGrid
Every correctness invariant holds on every market tested.
Headline
| Tested on | 7 markets × 3 years (2023-07-11 → 2026-09) |
| Median return | 0.3% |
| Median buy-and-hold, same window | 39.2% |
| Median edge vs buy-and-hold | -38.5% |
| Median max drawdown | 0.3% |
| Time in market | 0.1% |
| Profitable on | 4 of 7 markets |
| Beat buy-and-hold on | 2 of 7 markets |
| Total orders placed | 236 |
| Correctness invariants | 5 / 5 pass |
| Presentation surfaces uniform | 4 / 5 |
| Issues found and fixed | 4 |
What this does and does not mean
"ENGINEERING COMPLETE" means the code ran clean on every market, its accounting reconciles, it drew no chart mark for a trade that did not happen, and it places the same number of orders whether fees are 0% or 0.10%. That is a correctness result.
It is not a claim that the strategy is profitable — on this window it is not, and not advice to run it unattended. The performance report states the window, the market count and the sample size for exactly that reason.
Every number on this page is read from the run record that produced it — the audit pipeline writes the page and the record together, so the two cannot drift apart.