full image - Repost: Life at a quant HF - the bad, the ugly, and the fireable offenses (from Reddit.com, Life at a quant HF - the bad, the ugly, and the fireable offenses)
Mining:
Exchanges:
Donations:
I put together a thread a while back about a fairly typical day as an execution trader at a quant hedge fund. In the interest of showing both sides of the story, I wanted to share some stories about when things go wrong, and about the less glamorous parts of the job. I’ll tack on a few more especially wild ones if any come to mind. Happy to answer any questions as well.Technology is the heart of our business, and one of the biggest pain points as a result. Hardware failures - fried CPUs, memory faults, dead hard drives - can and do occur. We have an instance of our order management system on additional servers (cold backup in a secondary data center) but failover isn’t immediate and we’re in a business where seconds matter, and we classify ourselves as low-mid frequency so we aren’t even the most sensitive to this. Our hardware teams do their best to maintain our servers and perform preventative maintenance, but we essentially run 24/7 (including backtests/research on weekends) so downtime is hard to come by.Even if your own hardware is performing optimally and in perfect condition, you’re still reliant on someone else’s doing the same. We had a long haul fiber line that was showing random packet loss between two of our data centers, causing delays in position keeping and risk calculations. Turns out some construction company tore partially through a network line in the ground and didn’t even notice. (We invested in redundant connectivity after that particular incident.)Never assume that your data is 100% clean, even if it has been for years. We had an equity trade feed from a well known vendor all of a sudden start printing zeros instead of valid trade prices. We had one PM whose signals interpreted these as real prices and thought it was the perfect time to buy! We have plenty of downstream validations that kept these obviously erroneous orders from leaving the door, but it was still quite the fire drill and our order management system was going crazy with alerts/exceptions. In the midst of building complex systems and extensive logic, it’s always fascinating how we can miss something in our validations as basic as “stocks don’t trade at 0”.The term “systematic” frequently goes out the window when the markets are going wild, and you need to have a thick skin and think on your feet. This was especially true in the last two weeks with the banking crisis. We had some teams taking a bath and screaming at us to unwind longs to stem the bleeding. We had others who were short, firing in manual orders to cover and lock in profits. On a normal day the only time we’re doing manual orders like these is in markets that we can’t trade electronically, or in the event of technical issues where we have to put on hedges or blow out of bad fills. Tensions were high for sure, but at the end of the day we got everything executed properly (including a bunch of crosses on names with opposing internal interest, saving us some transaction costs), and now we’ll be having some discussions with a few desks about oversized factor and sector exposure…In an increasingly automated and low touch world, EQ is still important, and some people have lost sight of that. I maintain strong relationships with our executing and prime brokers, because at the end of the day they are still our conduits to the markets we trade in, and there are always qualitative components of the job that they are invaluable to. One of my (thankfully former) colleagues on the desk was a math whiz but a total asshole when dealing with our brokers, constantly treating them like the help when investigating order rejects/trade breaks, or barking internal lingo/acronyms at them expecting them to understand. It shouldn’t have been a surprise to him when our sales coverage at a broker went above his head to a senior PM who she was a former colleague of to complain about his demeanor, and I strongly believe this factored into him getting fired a few months later for what on paper was a relatively minor trade error. I try my best to meet brokers halfway and stay in their good graces, addressing concerns in a constructive way, and we get rewarded for it (competitive commissions and financing rates, tighter pricing on principal liquidity, seats at industry conferences, first looks to beta test new algo strategies, and so on).
Social Media Icons