Given that Alameda is a *trading* company first and foremost, I thought I’d share some thoughts on how we think about trading (in contrast to Sam’s thread about how we think about investing longer-term). https://twitter.com/SBF_Alameda/status/1314536324381065217

On some level, investing is just trading if you forget about exiting your positions, so there’s certainly a lot of overlap between our mindsets in investing vs. trading. But there are also a lot of differences, which I’ll highlight here.

The gist of all of our activities? Follow the money. If there’s money somewhere, we want to be there. And so we devote a lot of time and energy to 1) figuring out where the money is 2) getting there.

Our trading is not monolithic, and so there aren’t that many highly-specific claims I can make which apply to all of it. Some of our strategies involve doing thousands of trades back and forth, not holding much of a position ever.

Some of our strategies are longer-term -- maybe we’ll have some automated signal which suggests coin X will go up over the next day, so we’ll hold the position for a day. Usually it’s something a bit shorter, but we’ll follow the money anywhere.

And often when markets are quite volatile, there will be a ton of manual intervention and pushing of knows that put on positions we as humans recognize as good -- this is usually quite fast, and based on the intuition and fast thinking which makes it hard to automate me away.

And what kind of inputs do we use? Basically, anything systematic which we can ingest is fair game -- order book data, trade data, blockchain data, yield data, data from non-crypto markets (SPX, etc.), open interest data, etc. etc. etc.

We run studies to determine what sets of inputs give us reliably good information, test the strategies to make sure the liquidity is real, and then scale up. It’s really just the scientific method, applied to a specific set of market games. And that gives us strategies.

To some extent, given a set of lots of these strategies, what our day-to-day trading is can be described as: we have some capital base we can deploy however we want -- how do we decide where to put it? This optimization problem is at the core of what we always do.

Each of our automated strategies might require some working capital all the time to run -- so we need to know how much that will return on any amount of capital in order to decide what to allocate to it.

Some of our strategies are more passive, most of the time, and sort of amount to “leave a bunch of money sitting on some exchange and wait for the 1% of times it’s best to enter a huge position.” Estimating how good and rare those opportunities are dictates how much to allocate.

(That’s how I’d characterize our manual trading, too -- we leave a ton of collateral on the derivatives platforms, for instance, so we as human traders can decide to put on a huge position if we ever want to.)

The same is broadly true for yield farming: at any point, there’s a fixed return on capital for each platform, and we need to decide what to allocate where. Yields can change quickly, of course, but that’s true for the returns from any strategy, so this is basically the same.

Now, the strategies are not *actually* that independent -- often they’re drawing from the same pools of capital, and often their returns are quite correlated. But this model of how our day-to-day works is basically right.

(Our capital base also is not fixed -- many platforms offer leverage and loans and whatnot now, and so you can scale up your capital base if you can accept more risk -- we think about this all the time!)

Another wrinkle to this game is *risk* -- how do we think about risk? What if the money goes somewhere dangerous?

Broadly, we weigh how valuable a given action is vs. how much potential downside that action has (and how likely the downside is), and then if it meets so bar we do the action. And there are all kinds of risks, some of which are really hard to evaluate.

Some risks: platforms defaulting, transfers failing, a position getting liquidated before we can exit, entering into an unhedged but risky position that could go against us, a tech disaster causing us to lose a lot, etc. etc. etc.

There’s a ton going on here, and we often have to make risk adjustment decisions in seconds -- it’s another reason that humans are so important, and it’s an added layer when we’re considering each and every one of our strategies.

On a high level, that’s kind of it -- we learn about the markets, we come up with strategies, we test them, we run them, we think about how risky they are, we count our money. The specifics can get gnarly and really challenging (and fun!), but that’s the gist.

I’d summarize the differences between this and our investing as: different timescales (shorter trades + we also have less time to think), a little more agnostic about “what we’re doing” (sometimes we trust data over e.g. the macro elements of a coin), and more game-like.

(By which I mean: there’s a million buttons we’re pushing all the time and they have lots of sometimes hard-to-predict effects, and they interact with each other.)

There’s also a lot of math in the trading side -- I skimmed over the research process in how we develop strategies (hey, that’s the other word in our name), but we spend a ton of time on it, and we’d be screwed without it.

But trading and investing inform each other in really nice ways, so it’s been a natural fit. Like I said, we follow the money wherever it goes, and lately it’s been going to some pretty weird places :P Can’t wait to see where it goes next!