Monolith Research
The FX review started with an uncomfortable answer: the NAS-style intraday edge did not travel cleanly. Session breakout, raw momentum transfer, and broad FX trend tests struggled once costs were made less forgiving. That matters. A model that only works when moved unchanged into another market is usually not a model; it is a coincidence waiting for a name.
The better result came from a daily currency-strength rotation that allowed XAU into the basket. The best stressed configuration returned 7.63%, annualised at 2.56%, with daily Sharpe 0.63, maximum drawdown of -4.60%, profit factor 1.12, and 221 trades. By contribution, the useful part was plain: XAU carried most of the edge, while the pure FX basket remained weak.
The separate XAU trend branch was stronger as a standalone research candidate. Its best stressed result returned 21.27%, with daily Sharpe around 0.70, maximum drawdown near -5.72%, and 75 trades. That is enough to stay in the research programme, but not enough to blend carelessly into the main book.
The conclusion is deliberately narrow. FX/XAU strength is a separate paper candidate, not a replacement for the index or OU work. The value of the pass was in refusing the broad story and keeping only the part that survived.