Thread by @AlamedaTrabucco

On a plane to Miami, thinking about the topic of the panel i’m on: Is it possible to time the market? It’s a great question, and one my priors were REALLY wrong about.

A thread about being wrong and making money.

My initial training in quant trading came at SIG, trading more traditional assets than crypto (let’s hear it for fixed income ETFs). These are markets that have existed forever, and many large, skilled firms have traded them forever.

In these markets, it’s really hard for there to be large, fairly obvious, predictable moves — if those existed, these big quant firms would have figured them out too and arbed them out (or similar).

What do I mean? Let’s say that tech stocks consistently revert 50% of their post-earnings moves, always over the course of the following day. What would happen over time as the world figured this out?

Once a few firms figure it out, they’ll put the reversion trade for HUGE size first thing in the morning. This will have impact and now the reverting move happens all in the morning ...

until it’s pushed even earlier and now the earnings move is just 50% of its original size.

This is the typical pattern — big effects DO arise, but if they’re predictable and consistent, they start happening sooner once everyone’s trading them, until they’re just totally arbed out.

This is the paradigm I was used to when I started trading crypto.

I remember laughing at the idea that major coins always rally into big events. Surely if that’s true, I thought, someone should just be buying earlier than the rally usually happens ... right?

I said that right before BCH’s 2018 fork on 11/15 ... check out BCH/BTC from then.

Turns out this in fact DID always used to happen — hence the common “buy the news, sell the event” advice. And honestly, it still does USUALLY happen! For all kinds of events: forks, halvings, major exchange listings, etc. all have (decreasingly, in some cases) predictable moves.

I used to fuck this up all the time! I ignored a study that said coins rally between Coinbase announcing a listing and the listing itself — surely the world will know that arb it out. BTC rallies into a specific holiday every year? We’re not the only ones who know that, duh.

This was a really dangerous instance of me leveling myself and deciding that the obvious play is TOO obvious and couldn’t be right. It’s also an example of being way too clingy to one’s priors in the face of lots of evidence.

And it cost me and Alameda a bunch of money.

Eventually I connected the dots — this market worked differently, lots of the sophisticated firms aren’t in the game yet, more $ in the system than I’m used to ISN’T quantitative, etc. But it took me too long, and that was a huge lesson for me.

A profitable lesson, too! It’s let me make great trades around futures expirations, more recent big events, Elon Musk’s tweets, bad news from governments, other exchange listings, USDT, etc. etc. Things DO evolve, just most slowly than I wanted them to.

I pride myself now on updating quickly as new information comes up — I think it’s maybe even my primary value-add, or at least adjacent to it — and that involves abandoning priors much more quickly than I used to. Which is hard and unnatural, but important.

So! My answer to this panel’s question is a fairly nuanced “yes” (I’ve talked about instances of it before), and getting to that yes took me a LONG (but worthwhile) time.

Looking forward to talking about this (and other stuff 😘) in Miami!