Thread by @HangukQuant

1\ Risk management is probably my primary focus on top of alpha research. Am always thinking about how to better combine alpha strats. One of the safest bet, especially in lower dimensional portfolios is a static equal weighted basket.

2\ There are also different levels of risk/portfolio management, and some techniques can apply to multiple layers. For example, you can apply risk controls to individual assets, timeframes, strategy level and portfolio level et cetera.

3\ Academic literature on this is pretty extensive. Recently a tweet (can’t seem to place it) by @choffstein on literature of portfolio management. Definitely one of the my favourite narrators on mathematical models for trading.

4\ Imo, it is definitely possible to perform superior to what is suggested by academic literature and is what I consider my competitive advantage. I also extensively use some of the techniques mentioned in different layers of my strat controls, on top of my own prop stuff.

5\ For example, readers of my weekly reports know that vol-targetting is one of the layers. For those who want to learn more about multi-start quant stuff, I suggest -

6\
i) understanding the multivariate portfolio theory and attacking its assumptions (read quoted tweet)

https://twitter.com/HangukQuant/status/1446913123319877633?s=20

ii) diving into literature by Corey and tinkering with the techniques

7\ Generally it is advisable that you should get intimate with your alpha strats (add lags, change params, new datasets) etc and also identify its statistical properties.

Is the return distribution symmetrical or skewed?

8\ Some risk-management strats may not be compatible. For instance, if you are running a right tail strat with strategy activation on a lagged profitability curve, you only get to enter after the “big meat” has already occured and your alpha harvesting has been destroyed.

9\ At this point I would like to add that you should probably pay attention to ideas floating around in alternative content spaces including Twitter. I also once saw a great manumatic approach guide by @goodalexander, but also can’t find it😓

10\ What I am trying to say is, it is very important to keep your eyes out because when people talk, it is often their "best work" and usually no prop stuff comes out but you can see what they are focusing on and give you an idea where to start looking.

11\ Like AI agents, you want to reduce your search space for reliable content on trading, particularly since there is so much noise in the space of trading advise.

12\ For example, I thought that one of the most overlooked spaces were CTAs. Sounds less cool than “hedge funds” but these have been around for a long time, most implementing trend following systems which are by nature right tail.

Meaning most of the time they bleed money.

13\ So they bleed money and still have decent performances; so you should probably pay attention to what they do. Another example are some vol funds who trade negative EV strats but their risk management is so good that they can come out profitable.

14\ By the way, great follow @moritzseibert, and his podcast is also good, reliable stuff that works. Even if you don’t do vol stuff or trend following these people always have tricks up their sleeves that you can extend to your own portfolio management.

15\ It’s all about tweaking and hacking stuff that most people overlook until it works and then making them even better.

FIN.