A Junior Quant's Guide to Event-Driven Trading
The best model in the world can’t beat having better information. And we have it.
When you see a stock with performance like this:
You have to know that it didn’t get that way overnight. And more often than not, it didn’t get that way quietly.
On every step of the way down, companies like this are forced by regulators to publicly share every detail on exactly how business is going and what they’ve got planned.
All you have to do is look for it.
So, that’s exactly what we did.
A Primer on Advanced Event-Driven Trading
For a quantitative trader, you have to know that the “quantitative” part can only go so far.
At the end of the day, markets are driven by information, above all else.
Even for purely quantitative strategies like stat arb, that information component can cloud up returns as it’s really, really hard to hedge against a company that just announced it’s going bankrupt and went down 75% overnight.
By going the event-driven route, you’re still working with data and markets, the only difference is that the event happening is the signal.
Now, none of this might be news to you.
You’re likely already familiar with the most typical market events:
Earnings
Biotech announcements (e.g., FDA drug approvals/rejections)
Fed meetings (e.g., rate cut/hike decisions)
However, this is essentially the tip of the iceberg:
The real edge comes from turning things that happen in the market and making them an event.
To see an example of this, let’s take a look at a understudied, but common market-happening:
Dilution
At a high level, dilution is pretty simple.
A company needs more cash and won’t or can’t take out a loan from the bond market. So, it simply issues new shares and sells them on the open market.
With those proceeds, it can do whatever it pleases; raise exec salary, increase R&D funding, pay off other debts, whatever.
Now, naturally, this isn’t risk-free cash.
If you’re an investor in one of these companies and get told that your 5% stake is now going to be a 3% stake, you’re going to be less-incentivized to hold it, let alone buy more shares.
And you’re not alone; the same dilution happened to the other 3,000 investors and they have the same view too.
When all of those participants have the same view, on the same stock, at the same time, the net result gets pretty predictable:
At this point, you might think to yourself:
“Why not just short companies when they’re diluting their shares?”
Well, you can.
This is the core of the next tranche of event-driven trading:
You take a specific market happening, collect every time it happened, then see what happened after.
If it’s predictable in the way that economically makes sense (i.e., you expect diluted stocks to go down), you dive deeper and see if it can be a real return stream.
So, let’s do exactly that.
The Event IS the Trade
Okay, so we know that certain events lead to certain outcomes, so before building out the strategy, let’s map out which events we want to focus on:
Dilution
Company issues new shares, existing holders get diluted, selling pressure follows.
De-SPACs
A blank check company closes its merger and becomes a real company. The thing is, most of these companies went the SPAC route because they couldn’t IPO the traditional way. So, you’re already starting with an adversely selected pool. From there, it’s a mess of redemptions, float confusion, and brutal valuation discovery. The cohort underperformance is well-documented.
Defaults
Company misses a debt payment, gets delisted, or enters bankruptcy proceedings. By the time this happens, the stock is usually already in freefall, but the event itself tends to accelerate the final leg down.
Once we have them all ordered by date and time, we then want to see what happens if we outright short them for a month.
Now, before moving further, we have to address a very important stage of where event-driven modeling goes wrong: lookahead bias.
Although you can know about the event as soon as it happens, it is very risky to assume that you can or would have traded it exactly when it happened.
If you model on the basis that you traded right as the event happened, it often makes returns look much better than they actually were.
So, going back to the strategy; we’re going to add a full day delay between when the event was logged and the time of trading.
If we got word of the event on Jan 1 at 10 am, we place a trade on Jan 2 at 4 pm.
This comes with drawbacks of its own, but it at least gives us a more realistic baseline that would match practical trade execution.
So, with our strategy and universe defined, let’s run a few mock simulations:
As demonstrated, from a birds-eye view, this approach isn’t all too bad.
Naturally, by the nature of short-selling, since our max gain is 100% (stock falls to 0), and our max loss is virtually infinite (no cap on how high a stock can go), we inevitably hit periods where a stock we’re short goes 3, 4, even 10x higher, in a month.
It’s a rare event, but one that is guaranteed to eventually happen, as we’ve painfully experienced before:
But also coming from experience, we know that diversification is one of the last free “edges”, so by combining each event into a multi-strategy framework, we lessen the impact of the odd adverse event, while still capturing the bulk of the EV.
Now, as with all trading strategies there is always some nuance between backtesting and reality:
For the default events, most of these events happen late-cycle, when the stock is already trading below $1. This isn’t always the case, but it happens relatively often.
Most brokers that allow locates and short-selling have a $1 minimum, making some of these difficult to act upon.
Borrow interest rates vary greatly. For most stocks diluting their shares, it typically isn’t their first round, so there’s often plenty of float and thus low borrow rates (e.g., < 5% for 30 days).
However, before dilution the stock is often already negatively viewed, so if it already has a high short interest, the borrow rates can be higher than usual.Many companies “game” their PR around these events.
For instance, after filing for dilution, they may put out a release stating that they’re pivoting to being a crypto treasury, datacenter, or whatever is in vogue at the moment.
The net goal is to price the final offering at a higher value, so that more cash can be generated. This is often the source of many adverse moves against you.
Closing Thoughts
The beauty of event driven trading is that as long as markets are open, there will constantly be new happenings.
We’ve demonstrated that if you’re willing to dig past the first layer of garden-variety earnings and FOMC meetings, you open up a new world of repeatable, highly predictable trades.
As with all trading strategies, while things aren’t necessarily a money-printing, straight line up, once you know the nuances of the approach, you will at least always know where to look when you’re in need of an edge.
Like usual, the full code for the experiment we ran will be posted below, along with a production file that you can run to get a simple view on the latest actionable events, as they come in real-time:
Code → Alphanume Strategy Lab on GitHub
Where This All Comes From
For this experiment, we used price data from Massive and point-in-time event data from Alphanume.
As we mentioned in our Inside the Quant Galore Research Lab, we had spent years building an internal dataset catalog to research and trade off of.
Alphanume is simply the system we built for ourselves to experiment, take real risks, and to see which ideas survive outside of a code terminal.
Instead of manually parsing filings or stitching datasets together ad hoc, you can go straight to:
What just happened?
When did it happen?
What tends to happen next?
Now that we control the full dataset pipeline ourselves, we can focus on the kinds of edges that are:
too small for large institutions
too messy for traditional data providers
but very real for systematic traders willing to look
This is just one example, but there are dozens more live and sitting in the pipeline.
If you enjoyed this post, these others might be right up your alley:
— A Quant’s Guide to Cross-Section Maxxing [Code Included]
— Behind The Scenes of a Profitable Short Selling Operation
— A Junior Quant’s Guide to Corporate Actions
— A No-BS Look at Quantitative Trading Infrastructure
Good luck, and happy trading. 🫡🫡
 | A guest post by
|