Testing Precious Metals as Predictors of USDCAD

Summary

The article tests whether gold and palladium prices help forecast USDCAD. It explains the economic rationale for comparing the markets, then describes gathering price bars from MetaTrader, aligning observations by shared timestamps, and labeling a future exchange-rate target. The authors report negative correlations of -0.5 for gold with USDCAD and -0.66 for palladium with USDCAD, while the metals correlate at 0.37. Visual plots show no clear separation between outcomes.

Several models are compared using five-fold time-series cross-validation without shuffling. A tuned linear support vector regressor beats its default version on unseen data, but does not outperform a basic linear model; the authors therefore conclude that the additional metal inputs did not improve on using ordinary USDCAD data. They export the model to ONNX and connect it to an MQL5 Expert Advisor. The article provides no detailed out-of-sample performance statistics or evidence of live profitability, and cautions that noisy market data can favor simpler models.

Key ideas

• The study compares USDCAD price data alone with predictors that also include gold and palladium prices.
• The reported correlations are negative between USDCAD and each metal, but the charts reveal no clear predictive separation.
• Model selection uses five-fold time-series cross-validation without random shuffling.
• A tuned linear support vector regressor improves on its default form but loses to the basic linear model.
• The authors conclude that adding precious metals did not improve forecasts over ordinary USDCAD data.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.