Dennis, hope you are well!
During your talk at the Quantifi Annual Conference (watch here), you have noted that a turkey on November 1 may not be able to correctly predict its fate based on historical data. Well, Thanksgiving has arrived in March and finally everyone understands (again) that market regimes do change. This puts us in a great position to deliver our message. We'll report back in a bit on how BondDroid did in March and what we can do to improve its performance during the rest of the crisis and its tail.
In the meantime, I wanted to share this bit from Marcos Lopez de Prado. I read the paper and just listened to his webinar. His way of thinking is very similar to ours. Would love to know what you think about the paper and if you are familiar with his work. I know him as a practitioner when he ran his bond portfolio at Guggenheim.
Here's the paper: Three Quant Lessons from COVID-19 and a slide:
Notes from the webinar:
- Forecasting is dead, nowcasting is in, but requires a solid technological base.
- Quants didn't do well in March because they use backtesting as a source of trade ideas, not to disprove their hypotheses.
- One could have "predicted" the COVID crisis by means of nowcasting (looking at the supply coming from China, prices of copper, which all indicated a change in regime back in early February).
- He was asked if any long-term strategy is dead and his response was that it only works if it adjusts to a regime on an ongoing basis.
Best
K
7 Chord, Inc. | Dock 72, Brooklyn Navy Yard | 917-930-1974 | kristina.fan@7-chord.com | www.7-chord.com
BondDroid is a Customizable AI engine for Bond Traders and Investors: Book a Demo
We Are Hiring: AngelList | Stack Overflow
