When I tell people I used to trade for a living, the first thing they ask is whether I was good at math. From 2013 to 2016 I worked as a high frequency trader in United States Treasury and commodity futures, and later I covered futures and options on the institutional side at Bridgewater Associates. The math part of the job is real. It's also the part you can hire for, automate, or look up. The math wasn't where the edge came from. The edge came from three things almost nobody mentions when they talk about trading. World news. Poker. Psychology.
Quick disclaimer before I get into it. This post is about a job I did about a decade ago. It does not reflect my current role or my current employer, and nothing here is about either. I'm writing about what trading taught me as someone who isn't doing it anymore.
The Math Is Table Stakes
Anyone who runs a trading book has to do math. You have to compute basis between cash and futures. You have to roll positions. You have to know what a one basis point move does to your profit and loss. If you can't do that in your head or with a sheet, you can't sit in the seat. But every trader I worked with could do it. The math is the price of admission. It is not the prize.
Here's the part that gets missed. The market quotes are already the answer to a math problem. By the time a price hits your screen, every quant team in the world has already priced it. If you think you can find a number on the screen that the rest of the market got wrong with a calculator, you are going to lose money. The places where the math actually matters are narrow and specialized. Most of trading is not those places.
So if it isn't math, what is it?
World News
Treasury futures move on the things that move the world. A central bank press conference. An employment report. A surprise in oil supply because of something that happened in a place most Americans couldn't find on a map. The job, day to day, is to know what is going on. Who runs which country. What they want. What their political pressures are. What the speech tomorrow is likely to say, and how much of that the market has already priced in.
I read more news in those three years than I have in the rest of my life combined. The Wall Street Journal at six in the morning. The Financial Times at lunch. Reuters and Bloomberg headlines all day. The Economist on the weekend. Not because I was curious. Because if a finance minister in a country I'd never been to said something unexpected at a press conference at three in the morning my time, the price would move and I'd need to know why.
The good traders weren't the ones with the best models. They were the ones with the best context. They knew that the report on Friday wasn't the point. The point was what the report was going to mean for the speech the following Tuesday. They could hold a chain of three or four reactions in their head and bet on the second or third one before anyone else got there. That's a skill. It's also not a math skill.
Poker
If you want to learn how to trade, learn poker first. Not because the math is the same, although some of it is. Because the structure of the decision is the same. You have incomplete information. You have to act anyway. You have to bet the right amount, not too much and not too little. You have to know when the other side is bluffing. You have to know when you're the one bluffing yourself.
Poker teaches you to separate the quality of the decision from the quality of the outcome. You can play a hand perfectly and lose. You can play a hand badly and win. If you confuse the two, you'll learn the wrong lessons. Trading punishes that confusion every day. The trade that made money on a Tuesday because of a headline you didn't see coming was not a good trade. The trade that lost money because you were right about the direction but wrong about the timing might have been a fine trade. The market doesn't care which it was. You have to.
Poker also teaches you about size. The math of a position is easy. The psychology of holding it is not. The trader who is right about the direction but sized too big will get stopped out before the move happens. The one who is right and sized correctly will sit through the noise and collect. Position sizing is more poker than math. You're managing your own ability to stay in the hand, which is mostly a question of how you handle losing chips you weren't expecting to lose.
Psychology
Every trader I knew got better when they learned to read themselves. The market is a machine for finding out what you don't know about yourself. If you tilt after a loss, you'll find out. If you press too hard when you're up, you'll find out. If you anchor on a price you bought at, you'll find out. The good traders had a kind of internal weather report running constantly. How am I feeling. Is that feeling making me want to do something. Is the thing I want to do a good idea or just a way to make the feeling go away.
The same applies to reading other people. The futures pit is gone but the social structure of the market isn't. There are still big players, and there are still moments when you can tell from the tape or the chat that somebody got run over and is trying to get out. There are still salespeople with tells. There are still days when the desk is collectively in a bad mood and the market does what the desk's mood is doing. Reading that is a skill. It looks nothing like math.
Bonds Trade Through Chat
Here's the part that surprises people most. A huge share of bond trading still happens in chat windows. Not exchanges. Not order books. Bloomberg messages and similar tools. A trader on the buy side wants to sell a corporate bond. They send a message to four or five dealers asking for a price. The dealers come back with quotes. The trader picks one. The trade prints. That's the workflow. It's called request for quote, and even on the electronic platforms, the underlying logic is the same as picking up a phone. Most corporate bonds in the United States are still traded over the counter rather than on a centralized exchange.
Compare that to equities. If I want to buy a share of Apple, there's one ticker and one consolidated tape and a dozen exchanges all quoting the same instrument and the price is the price. The whole market collapses to a single number. For bonds, none of that is true.
Why hasn't it standardized? Three reasons.
First, there are way too many bonds. Apple has one common stock. Apple also has dozens of bonds outstanding, each with a different maturity, coupon, and call schedule. There are about forty thousand outstanding corporate bond securities in the United States. Each one has its own nine character identifier called a Committee on Uniform Securities Identification Procedures number, or CUSIP. Most of them barely trade. A typical investment grade corporate bond might change hands a few times a month. You can't run a continuous order book on something that trades a few times a month. There aren't enough buyers and sellers in the same minute.
Second, the existing system works for the people who profit from it. Dealers make money on the spread between what they pay for a bond and what they sell it for. In an opaque, fragmented market, that spread is wider. Bid-ask spreads in corporate bonds are meaningfully wider than in equities, which is another way of saying the dealers make more per trade. A move to a fully transparent exchange model would compress those spreads. That's good for end investors and not good for the desks. The desks are the ones with the seat at the table when the rules get written.
Third, bonds aren't really fungible the way stocks are. A share of Apple is a share of Apple. Two corporate bonds from the same company with the same coupon and the same maturity might trade at different prices because one is held by buyers who can sell and one is held by buyers who can't. The instrument is the same on paper. The supply and demand for it is not. Centralized markets work best when one unit is exactly like every other unit. Bonds don't quite clear that bar.
Treasuries are a partial exception. The on-the-run benchmarks trade on electronic platforms and look much more like an exchange. The off-the-run issues, the strips, and the older bonds revert to the chat model. Even there, you have fragmentation across CUSIPs that economists have argued for decades should be consolidated.
Why Any of This Matters
It matters because the popular picture of trading is wrong, and the wrong picture leads people to the wrong careers and the wrong investments. People think trading is a quiet room full of mathematicians. Some of it is. Most of it is people in a chat window trying to figure out whether the person on the other side knows something they don't, while reading a newswire and a transcript at the same time. The skills that matter are reading other people, reading the news, and reading yourself. Math is the floor.
If you're considering the field, learn poker. Read the news every day for a year. Watch how you react when you're wrong about something small. The math you can pick up. The other parts take longer, and they're the parts that actually decide who keeps the seat.