Hello folks – just a quick note that we're making progress behind the scenes on dgm-0.7 and so far, the backtest performance looks great compared with dgm-0.6. While dgm-0.5/0.6 did great over the summer, we didn't miss the fact it has been somewhat disappointing since the November market melt-up.
We're confident that the dgm-0.7 outperformance is due to better optimization and feature selection, and that it doesn't mean we're overfitting the model. The tests against multiple time periods (25/50/100/200 days) should help keep that in check. Time will be the ultimate test.
We're on track with testing and should be rolling out soon, so please stay tuned.
For any questions: [email protected] (or just reply to this notification)