Efficient Capital Allocation - A🧵
A lot of traders are looking to deploy multiple systems with opposing styles (trend following + vol selling) on same capital to enhance returns. The assumption is that with sufficient buffer, idle capital can take care of losses in one sys. 1/n
While some results may be generated from the other system. The risk from this comes in form of systems becoming highly correlated and leading to more than perceived DD from back tests. The traders can be better off if they perform following: 2/n
1- Understand if your returns are coming from different aspects of markets. This could be cross asset class or from different style (Trend/Mean rev/Vol selling etc) 3/n
2 - Perform a rolling correlation of your returns from both systems. Use different lookbacks in your analysis to identify when does your system come under stress and correlation either shoots up or breaks down. 4/n
3 - You're assuming future correlations to hold for your combined system to make money. Stress this assumption. A Monte Carlo simulation is easiest and quickest way to do so. 5/n
Lastly - appreciate the fact that backtesting is one data point from historical possibilities. Go in with expecting pretty different result than your backtest shows. Be aware of Ergodicity in your returns. n/n
I meant - lack of Ergodicity in expected returns by being aware of it.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
A brief history of #BTC crashes. As a #BTC HODLer, I've been mockingly asked the ques of - how is your portfolio doing several times in the past during BTC crashes. What is happening this month is no diff 4m several other instances in the past. 🧵 on 2 such past events: 1/n
#BTC went from sub $100 to making a top at $1163 in Oct'13. Over the next 14 months it retraced and fell around 87% to make a bottom at $152.4 in Jan'15. I did not have any long term exposure during this period. BTC at $250-350 was considered "EXPENSIVE" for a long period. 2/n
Over the next 3 years it went to a high of $19666 in Dec'17. A 129x increase in price from the bottom and 17x from previous top. Previous "Crash" was a blip on chart and $250-$350 entrants were considered lucky now. 3/n
Weekend learning session with @alok_dharia and @VohiCapital. Muchas Gracias 🙏
Had an opportunity to discuss a wide range of topics from vol trading to global macro. Learnt handful of new things and got reinforced on some others.
A brief summary of our discussion: 1/n
1. All of their and our systems are automated. It is not only important to get confidence on our edge via backtesting, but also opens up possibilities to explore more opportunities as current algos/models trade Live.
2. Volatility Trading has more to do with understanding the dynamics of volatility itself, rather than rampant/random volatility selling. While later may work in high vol environments such as 2020, to get long term performance, spend time on modeling/understanding vol features