Churning and Chatting #5: Quant-The Other Side of Your Trade

When it comes to the consideration of becoming a professional trader, I think there are primarily three career verticals.

Prop trading is where you learn to read price and trade right away from day 1. You get a fast feedback loop and full control over your playbook but you’re not going to scale to the 9-10 figure AUM level¹. Pod shops give a portfolio manager and their team the capital, data, and infrastructure to run a proper hedge fund book. You get to do the research but it’ll take years before you’re the guy who puts on the position. It’s hypercompetitive, it scales to the billions, but you’re gone at -5%. Quant trading is you do all the mathy stuff–mine data, test, automate. You are the casino, you make markets, and you basically never lose. It’s the toughest seat to get out of college, and a lot of the work isn’t really trading, it’s just monitoring the systems and taking notes.

All my connections from the past 15 years are concentrated in that first daytrading vertical.² That vertical is shrinking and it’s by far the riskiest proposition to aim for. It’s also the one where, if you fail, you probably don’t build much of a skill set for other industries. So is it a good career choice, I don’t know. Why not consider the other two if you’re still young?

That’s part of why I started this series. I could spend the next ten interviews talking to hyperscalpers and degen short-selling whales, sure. But if you’re a student trying to figure out where you fit, that’s a pretty narrow slice of the business. So when a longtime reader who makes markets at a tier-1 firm reached out, I wanted to know what his job actually looks like. I verified his employment through BrokerCheck. He told me his NDA limits what he can discuss publicly, so we kept the conversation to what he was comfortable sharing and agreed to just call him QUANT. He’s also more open-minded than you might expect from reading quant Twitter, which is a low bar. Bring up retail momentum, technical analysis, or old-school prop trading, and he won’t immediately look at your face with absolute disgust.

We go over the basics. I’m a dum-dum who still doesn’t understand market making entirely despite reading several books about it and I assume my audience is just as dumb. So.. what even is market making? How does it work? Is it really just turning on a free money printer all day as you kick your feet up? If that’s the case, what do new trainees even do? Fetch coffee?

The Makings of a Market Maker

Pete: How did you get into quant trading in the first place? What makes a young child even aspire to be one?

QUANT: My parents were immigrants and we grew up in a rough neighborhood that literally had FBI drug raids near us. They wanted me to stay away from the bad eggs so the message to me was simple: work hard, get educated, stay on the straight and narrow. I remember my childhood was full of games. My dad and I played Monopoly, did math drills, messed around with Rubik’s Cubes. I also practiced martial arts, which I think gave me discipline and self-confidence. There’s this thing when you’re about to break a board where you can feel yourself wanting to pull back. You have to actually believe you can break it and swing through it with commitment. That stuck with me throughout my entire life.

I was very involved in school as well–math club, sports, 4.0 GPA. Many of my friends were those getting into math contests and they were way faster than me. But my thing was that if I got interested in something, I’d put an unreasonable amount of time into it to catch up to everyone else. I heard those guys talk about quant trading all the time and that’s how I went down the rabbit hole. I read Reminiscences of a Stock Operator, listened to every episode of Chat With Traders. I e-mailed every firm to offer to intern for free. I just wanted to get in the room and see what these people did all day.

Pete: Tell me a bit about your career.

QUANT: After graduating, I did an internship for a few months at a Chicago trading firm. I received a full time role as a junior quant trader at one of the top options market making firms. I spent a few years there, but left after I realized that I wasn’t getting paid in line with my PnL. I considered going to the hedge fund side or trading my own account at that point, but I decided to just go work at a rival top options market making firm instead.

Pete: What do you actually do all day at an options market making desk?

QUANT: Before the open, I’m checking that the machines are running, the feeds are live, and the risk looks right. I’ll do some napkin math on the Greeks. How big is my position? What happens if the underlying moves 20%? A desk head will grill a junior trader on that in the morning. You need to have a feel for the answer before you look at the screen.

I’ll check headlines and look for anything that needs special pricing treatment, like a stock split. I have one feed for the biggest trades in the market and another for the biggest trades we’ve done. Once we open, I’m watching executions, checking connections, and looking for anything odd in the stats. If something looks off, I have to work out whether the market changed, our pricing is wrong, or something broke.

At the end of the day, there are exercise decisions and conversations with the prime broker. When the market closes, we all gather to discuss new and ongoing projects. Projects are prioritized by PnL potential. If a project involves pricing improvement in liquid TSLA options, it’ll get more consideration than illiquid names.

Pete: What happens to market makers when their markets get really wild?

QUANT: There can be a bit of overload. Sometimes an exchange feed goes stale and everybody has a problem. Spreads blow out³, message traffic spikes, and some firms’ systems fall behind. During the tariff volatility, I saw firms have to stop trading because their systems couldn’t keep up.

There’s also a pricing problem. Take the events that occurred around the Iran War and the Strait of Hormuz–we saw a lot of outlier moves. Brent crude hit $120 and fell below $90 later that same day. In its recap that morning, Cboe reported that one-month oil implied vol had jumped nearly 40 points the prior week to 104%, the highest since 2020. You may need to price in event risk and revisit pricing parameters across a huge number of names at once. A big firm has the infrastructure to cover a lot of symbols. A smaller firm focused on one name might get to the right price in that name faster.

And the people still have bad days. I remember a desk head putting his keyboard through his monitor. The monitor was replaced within the hour.

Pete: When you send a quote, how much of that price comes from your model and how much comes from the market?

QUANT: Both matter. The model gives you a fair value. The market tells you where people are actually willing to trade. You also have to look across the chain. Two quotes can each look reasonable on their own and still give someone an arbitrage when they trade them together. Your quoting system needs to catch that before the orders go out.

Then there’s inventory. Say I think an apple is worth a dollar, and I’m quoting 50 cents bid and $1.50 offered. I buy ten apples. If my inventory adjustment is one cent per apple, I move the whole quote down to 40 cents bid and $1.40 offered. My bid is less attractive to the next seller and my offer is more attractive to the next buyer. That’s what I mean by leaning my quotes. I still think the apple is worth a dollar. I’ve changed the prices at which I’m willing to buy or sell the next one.

How a Market Maker Thinks

Pete: How do you price earnings vol when you only get a handful of observations a year?

QUANT: I’ll start with a comparable company. Is another name in the sector reporting that week? How do those two names usually price relative to each other? I’ll also look at how big an earnings-day move tends to be versus a normal day in that sector. Those are ways to get some bearings with a small sample.

Then I ask what’s different this quarter. Do the fundamentals give me a reason to expect a bigger or smaller move? If one option looks out of line, I want to know how it got there. Does someone trade that same option every earnings cycle? Is there a recurring order pushing the price around, or is the market reacting to something I’ve missed? You can see the same price in both cases and want to trade it very differently.

Pete: How do you avoid getting picked off (adverse selection) when you’re quoting electronically?

QUANT: A lot of it is table selection, like poker. Find the fish and play against them. Which names attract retail flow? Where do people regularly cross wide spreads? Where can I actually price well enough to compete?

Learn from the sharks who play against you and get feedback from them too. I want to learn from the people who consistently trade well against me. If they keep picking me off in the same spot, there’s probably something about that spot worth understanding.

Once you’re trading, look at your markouts. That’s what happens to the value of a trade after you get filled. Break it down. Did the stock move against you, or did implied vol move against you? Does it happen in certain products, with certain types of flow, or when a particular trader is running the book? A bad average tells you there’s a problem. The breakdown can tell you which quote to change. You can decide to quote certain expirations or delta ranges and leave the rest of the surface alone.

Pete: How much can you tell about who’s on the other side?

QUANT: We categorize our flow into 3 categories: market maker, customer, and professional customer⁴. I want to know which bucket most of my fills are coming from and how those fills behave afterward. If other market makers keep trading against me and the markouts are consistently bad, I’m going to look closely at my pricing there. The buckets also help you pick out customer flow that shows up over and over. You’re basically getting feedback from the market itself.

Note: Quant says they do not know anything about a customer besides his number tag. They don’t know it’s Ken Griffin buying, they don’t know it’s a random piker at a 3rd prop shop who’s always complaining they’re stopping him out.

Pete: How do you actually spot repeat flow⁵?

QUANT: The manual version starts with a print that’s big enough to catch your eye. You ask whether you’ve seen a similar trade before. Then you look for a third and a fourth. Does it keep showing up in the same part of the chain? Is there some event that explains when it happens?

A lot of that search is automated now. You can pick up smaller trades that a person watching a screen would never connect. Some fundamental events also give you a place to start looking because you’ve seen the same kind of flow around them before. If there’s repeat flow that’s constantly re-pricing the option, the desk might widen its spread or quote less size against it (reduce risk).

Pete: What has all the retail interest in short-dated options (like 0/1/2 DTE) done to the vol surface⁶in popular stocks?

QUANT: It can create very profitable dislocations across the surface. That’s about as specific as I can get.

Pete: Is there an edge in looking at realized vol⁷?

QUANT: Realized vol is useful as an input for more complex models but on its own, it’s pretty basic. You might give an intern a project on how it behaves in a particular set of conditions. If there’s something there, it can become one feature in a more developed model. There’s a big jump from showing a relationship in the data to having something you can trade on its own.

The Development Curve

Pete: What was the training like once you got in?

QUANT: They had a lot of different modules, almost like college courses, with different subjects: poker theory, trading sims, options pricing, a primer on learning the firm’s tools. You’d find a trader to shadow and ask whatever you could. Later on, we’d have to study for our FINRA exams.

Getting to the desk was a different experience. You get a few projects and a lot of room to figure out where you can actually help. That’s where a mentor matters. I’d ask someone, “Why did you delta-hedge right there?” If they say, “Because I know how to trade,” okay, man. What am I supposed to do with that? The best mentors will tell you what they were watching, what made them act, and what would’ve made them wait. They can break down every single input into the decision.

Pete: How do you teach yourself to make those decisions?

QUANT: I break it down into inputs and outputs, similar to how I thought about games and puzzles when I was younger. Inputs can be market trades, our fills, news, what’s on Bloomberg, what related products are doing. Outputs are the things I can actually change. I can move the vol curve, adjust rates or system parameters, or hedge delta. Once you understand those mechanics, you can start asking which action makes sense under which conditions.

P&L is part of the feedback. I also like simulation because you can run the same decision through a lot of different markets. That’s where the poker analogy comes back in where it’s almost like you look for the game-theory optimal decision. You can test whether a response holds up across a range of situations or whether you just happened to get paid the first time you tried it.

Pete: You keep a big trading journal. What actually goes in it?

QUANT: Ideas from papers, things I see online, observations from other traders, and trades that made me stop and ask what happened. I categorize them and backtest what I can. If there’s a big jump in options prices, I want to know what changed, when the information showed up, and which expirations moved.

One example was when Roaring Kitty announced a GameStop livestream. Suddenly there was an event the options had to account for.

Early earnings releases are another example (something not on the calendar). First you adjust by adapting a web-scraper for your earnings stocks. Then you need to know what to do when it happens. In a situation like that, I’d think about when to pull the market-making quotes and when it made sense to switch to a market-taker model.

I try to do the same thing after a loss. Figure out what I saw, what I did, and what I’d change if I saw it again.

Pete: Give me an opinion that would go against the grain in your industry.

QUANT: People should learn to be a market taker before they’re a market maker. It’s easier to be right once a day than all day. As a taker, you get to pick your spot. As a maker, you’re putting prices out across names, strikes, and expirations, and somebody else gets to pick when to trade with you.

The thing that drives me nuts is watching somebody really smart get personally offended by a stock price. They’ll be like, “This company sucks. The vol is insane.” Okay. It can suck and still squeeze in your face. These tend to be the junior traders who don’t make it.

I’ve watched people get so invested in being the smartest person in the room that they stop asking why anyone would pay that price. That question is the whole job. What happened? What did you miss? What would you do if you saw it again?

Pete: What does career progression at a quant firm look like?

QUANT: You can spend a whole career trading on a team, just making base and bonus. As you get more senior, you might own a bigger piece of the strategy or take responsibility for more of the book. If you run a desk, you’re also deciding who to hire, what infrastructure to pay for, and which ideas are worth putting capital behind. Depending on the firm, you can get a bigger share of the desk’s P&L. Further up, as a partner or founder, you might have multiple desks under you and decide where the firm puts its capital. Think Jeff Yass at the founder end of that spectrum.

Pete: Did you ever want to trade your own money?

QUANT: I’ve thought about going out on my own. Then I picture opening my personal account and seeing a six-figure swing in a week and it’s not palatable to me. How badly do I need to be my own boss, really? At the firm I get the infrastructure, I trade firm capital, and every year they pay me a little more to have this exact conversation with myself. That’s the golden handcuffs. Starting a firm sounds exciting right up until you start pricing out what it takes to get one running. I’ve actually heard of firms that run -EV models solely for moat driven reasons, possibly to price out new competitors.

The Market Maker Myth

Pete: What do retail traders get wrong about market makers? Is it true they take out your stops, or that they manipulate markets for their own greedy agenda?

QUANT: People talk about “the market maker” like there’s one guy or one central trading desk deciding where a stock goes, when it’s not like that at all.

In practice, you have firms competing with different models, different inventories, and different reasons to trade. When a stock goes through your stop, think about how many people might have been watching the same level and what happens when they all react. It’s the whole apparatus, every participant, moving a stock through a price level, it’s not one guy who knows your customer tag and he happens to hate your guts.

The market maker’s job is to provide liquidity or improve the price someone can trade at. A lot of what you see in our quotes comes down to how much risk we’re willing to take at that price, given what we already own and what the market is doing. It doesn’t start with trying to move a stock to a certain price.

The market maker boogeyman has been around for as long as I’ve been trading and lurking on message boards. The GME squeeze and the whole debacle over Robinhood restricting buys gave that corner of the peanut gallery a much bigger audience. Years before GME, I remember the DNDN and OSTK boards and more OTC stock threads than I can count turning into full-blown market maker conspiracy hubs. I suspect some people just want an excuse for why their stock sucks or why their trade didn’t work out. Hopefully this post sheds light on what market makers can really control.

Pete: What do you think of some of the retail strategies you’ve observed, or just technical analysis in general?

I’m open to the idea that technical analysis can have an edge in certain markets, especially when it reflects retail psychology. But the profitable retail traders I’ve seen use it alongside another edge, and that other edge usually does the heavy lifting.

Take a simple example–shorting a junk stock when it breaks below the 200-day moving average. The break gives you an entry, but the edge probably comes from knowing which stock is worth shorting. Same with the capitulation trades and blow-off trades I see from pro discretionary traders. They know when to take it and which names to take it in. A beginner sees the first green candle, turns it into a TradingView signal, and starts firing at every ticker. They fixate on the charts but not on the real edge.

Pete: How does AI fit into your workflow? Do you see AI agents eventually replacing junior roles on the desk?

QUANT: AI is already a huge part of my workflow. We’ve automated a lot of the tedious stuff, and it’s way easier to go from an idea to a working prototype. You still need someone who understands the desk well enough to check the output, polish it, and actually put it into use. I expect LLMs to take on more of that work over time.

The idea that AI is going to wipe out junior desk roles feels like mostly marketing hype to me. I think the jobs will change. Interns and new hires will be able to take on more complex projects earlier. Even this summer, some of the stronger interns across quant firms were using AI to do most of the work on their projects. They spent their time verifying the output and learning from full-timers. I think we’ll see a lot more of that.

Pete: What advice would you give to other traders and quants?

QUANT: If you want to go pro, go pro. Give it a real shot. Find people who’ll tell you exactly where your thinking is sloppy. Read the papers. Test the idea. Write down the losing trade while you still remember what you were thinking when you put it on. A week later, you’ll have a much better story about why you did it.

Your neighbor is not a genius because he made money buying Micron last week. Can he tell you what he saw before the trade, what would’ve made him get out, and what he’d look for next time? Ask yourself the same questions after a win. There are insanely talented people in this business who still do that work every day. I love that about it. You can spend years getting good and still sit next to somebody who makes you go, “Wait, how the hell did you see that?”

Ultimately, you want to provide a service to the market. Either price discovery or price improvement. A market maker gives you a price you can trade at. A market taker helps show where that price should be. Find systems (inputs) that take you to that outcome (outputs).

Me: Thank you QUANT for being an interview guest on the fifth iteration of Churning and Chatting!