The US equities market is often held up as the exemplar of a free market system – a shining example of a public market that seamlessly integrates new information to provide meaningful, uniformly visible pricing. But behind the scenes, many aspects of it are opaque and obfuscated. Complicated pricing tiers govern what brokers pay in fees to exchanges. Undisclosed proprietary algorithms govern how brokers behave on behalf of investors, and so on. However, there are finally some headwinds toward increased transparency and accountability, and more serious debate happening about the health of the markets in the policy sphere, especially at Proof Trading. Allison Bishop, President of Proof Trading, and her team are phenomenally impacting the FinTech industry by working towards the evolution of the US stock market into having more transparency around market operations and costs via scientific rigor.
Full name: Allison Breton Bishop
Job title: President
Company: Proof Trading
Location: New York, NY
In 2013, Allison started as a tenure track assistant professor of computer science at Columbia University and ran a research group in cryptography. She loved teaching and doing research, but began questioning the priorities of her larger research community, which seemed more focused on publication rates and mathematical sophistication than real world impact.
“I had the freedom to define my own research agenda, but I questioned my ability to know what the most important problems to work on would be given I had never worked in the industry. I ventured outside the academic bubble looking for socially meaningful work, and I met Dan Aisen, a co-founder of IEX, a relatively new US stock exchange. For a few years, I worked with him at IEX to design systematic defenses against predatory trading practices that could be incorporated into the fabric of IEX’s trading platform. In early 2019, we left IEX to found our own company, Proof Trading, which is an institutional broker-dealer for US equities.”
The role of an institutional broker like Proof is to decide how to trade. “We take large orders to buy and sell stock from our clients and we slice them into pieces and feed them into the market strategically, using algorithms we’ve designed to avoid having too much impact on the prices of the securities we are trading. There are lots of individual decisions to make: how we spread orders out over time, over trading venues, how we interact with the trading venues, how we incorporate incoming information from market data feeds, and more. The ultimate outcomes of our decisions are noisy and dependent on many wider market forces that we do not control. This combination – lots of plausible choices and lots of noise in measuring outcomes – creates a lot of cover for conflicted practices and makes it very difficult to hold brokers accountable.”
Proof is working to change this by setting a radical new standard for transparency and scientific rigor. The platform publishes the research behind its algorithm designs so that the process is subject to scrutiny, and it extends its philosophy of transparency across all aspects of the company’s operations - including a policy of full transparency of compensation within the company.
Currently, Allison is focusing on designing Proof’s second trading algorithm which is not yet in production. The design rests on a statistical model of how various trading actions result in subtle shifts in price distributions. Through her professional journey, Allison has struggled to find purpose in her work, or conversely, to find work that feels purposeful to her. “It has never felt right to me to assume that just because I’m good at something, it is something worth doing, or just because someone will pay for something, that it is something worth doing. Walking away from a tenure track position at Columbia because I didn’t feel I could have enough positive impact was a difficult thing to do, and many people told me I was crazy. I’ve been very lucky that I’m still able to pursue my passion for my teaching in addition to my industry work, and I’m currently teaching as a part-time visiting faculty member at City College, CUNY.”
More on Allison
Currently lives: New York City, apartment in the financial district
Family at home: Married, I live with my husband and my thirteen year old labrador retriever.
Hometown: Don’t really have one. I moved around as a kid and lived in Indiana, Ohio, Tennessee, and upstate NY (Ithaca), before college and many more places since.
Favorite hobby: Boxing and stand-up comedy.
Favorite part of your day: Eating dinner on days my husband Richard cooks (his food is amazing!)
Favorite show to binge: Fleabag.
Daily Diary
Monday
6:45 am: I wake up and walk my dog, Paley. Even though she thinks it is too early.
7:00 am: I check emails that are generated automatically by our trading system. It’s doing stuff, and it wants me to know about it! It ends each email subject line with “Successful!”
7:40 am: One of the emails says we have 10493/10494 symbols mapped. Our perfect streak is over. I shake my fist at the sky and type some commemorating the moment in the appropriate slack channel.
8:20 am: Our CTO, Prerak, gives the thumbs up that we are up and ready for trading today. I fix coffee and an English muffin for breakfast. It’s a gluten-free English muffin. It’s not “good,” per se, but it’s food. I hate that I have celiac diagnosed this late in life after many solid decades of loving bread.
9:00 am: I work on a document describing our cybersecurity controls and policies. I get to use the phrase “threat model” a lot.
9:30 am: The market opens and I’m watching it. I think about how the stock market is kind of like a video game for people who are supposedly grown up now.
11:00 am: My knock-off brand Roomba starts and my dog has an existential crisis. This happens every morning at 11 am. I take her for a second walk for the day. I tell my husband Richard to change the automatic timer on the vacuum to some time that’s outside of market hours.
12:00 pm: I fix myself lunch at home.
1:00 pm: I’m working on some python code that will analyze our algorithm’s performance on a sample size of orders that is much too small to analyze. My goal is to get the results to correctly show that they are not meaningful.
1:30 pm: I notice something meaningful despite myself – I have messed up the arrival time benchmark.
2:45 pm: I read the day’s “Money Stuff” column from Matt Levine. I love “Money Stuff”. I think for a minute about my idea for a parody version called “Money Stuff – for her”.
3:00 pm: Back to writing python code.
4:00 pm: The market closes.
5:00 pm: I sneak in 20 minutes of practicing the trumpet. Richard bought me a trumpet as a gift a few months before the pandemic started because I’d always wanted to learn. I’m sure our neighbors have loved this.
5:45 pm: I diagnose and fix a bug in my python code. It came from the fact that order identifiers in our system get reused over days. Duh.
5:50 pm: I bribe Paley with dinner and another walk so she’ll sleep through my upcoming martial arts lesson.
6:15 pm: Krav Maga lesson over zoom. I’m in the living room, shadow boxing away from the rug where Paley is sleeping. Richard walks in to fix a snack while we’re talking about different ways to defend against a hypothetical knife attack. Five years into my martial arts training, he’s unfazed by this.
7:30 pm: I order ramen for dinner.
9:30 pm: I work in about 30 minutes of guitar practice, unplugged.
10:00 pm: I write a sample response for homework I assigned for my CUNY class, “Privacy for Data Scientists.”
11:00 pm: Just a little more python coding.
12:00 am: One last dog walk and then we all crash.
Wednesday
7:15 am: I wake up. Paley is ready for me this time. We take a walk.
7:50 am: I pick up my niece and escort her to school.
8:30 am: I pick up breakfast and an iced coffee. A big iced coffee. Today I am working from the office, like an adult. That means I can get a large coffee with cream and sugar.
9:00 am: I’m working on python code again and waiting around for the market to open.
9:30 am: The market opens. As usual, it is anticlimactic.
9:45 am: I consider writing in this diary that I’m writing in this diary. Seems like it could create a loop that destroys reality. Like having a meeting about meetings, or keeping a spreadsheet to keep track of spreadsheets.
10:00 am: Back to working on python code.
12:00 pm: I’m researching potential new office spaces. We are growing and need a few more desks than our current space holds.
1:00 pm: Something looks wrong with our stats!
1:45 pm: Something was wrong with our stats – it was a bug in the stats computation and I fixed it.
2:00 pm: I step out to get lunch.
3:00 pm: More writing/debugging code and research about office spaces.
4:00 pm: Chatting on zoom with Dan.
5:00 pm: I take a walk in the park and get a smoothie. The smoothie guy recognizes me from last week. I have become that predictable.
6:30 pm: I teach my CUNY course over zoom. We are talking about differential privacy and the lawsuit concerning its proposed use in the census.
9:00 pm: I answer emails and generate the volume curves and other stats for tomorrow’s trading.
9:30 pm: I look out of my office window and resolve to try to run the two miles home before a thunderstorm arrives.
9:35 pm: The rain starts. I bail and take the subway home.
Sunday
9:00 am: I wake up to Paley pacing. I had forgotten to set my alarm. We quickly go for a walk.
9:30 am: I have a donut and iced coffee for breakfast. I wonder briefly how many days in a row I have had a donut for breakfast. I decide it’s best not to keep track.
10:00 am: I get on zoom with my friend and co-author Sasha Fradkin to work on our second book. It is a draft in progress that is a follow-up to our first book, “Funville Adventures,” which is a fantasy adventure story that introduces kids to mathematical functions. Sasha is great at turning mathematical concepts in magical premises. We’ve been working on the new draft every Sunday now for several months, and it’s about three-quarters done.
11:15 am: I get off zoom and eat an early lunch.
12:00 pm: I have a trumpet lesson over zoom. I wonder if the neighbors appreciate that my triple-tonguing is coming along. Paley does not. Richard saves her by taking her for a walk in the middle of the lesson.
1:00 pm: My friend Christina texts me that she is heading to my neighborhood to join me for an afternoon run. I get my workout clothes on and head out.
1:30 pm: We run along the Hudson at a leisurely pace for about 5 miles. The day is gorgeous, if a little hot.
2:45 pm: I walk the last mile home and get a smoothie. The smoothie man in this neighborhood doesn’t know me yet, but he will. I’m listening to an audio book (“Empire of Pain” by Patrick Radden Keefe). I pause it because I have a thought about the design of our main trading algorithm. I realize something I could try to de-noise the training data a little further.
3:00 pm: I have virtual office hours over zoom. As a bouncing off-the-walls extrovert, I miss in-person teaching and in-person learning desperately. But I’d much rather have the zoom versions than nothing, so I am grateful.
5:00 pm: I catch up on some administrative work – answering emails, reviewing paperwork, etc.
5:30 pm: I write down my algo design idea in my whiteboard notebook.
6:00 pm: Richard resolves to make gluten-free bread that I will like. It appears to be a complicated process, but I’m hopeful about the outcome. In the meantime, the apartment smells like fresh baking bread.
6:30 pm: I feed and walk Paley while I wait for dinner to be ready.
8:00 pm: We eat dinner. The bread is a success.
9:00 pm: Richard needs to record some tracks for a virtual choir performance. I’m supposed to be quiet. Quiet is not my specialty, but I do my best. I work on the lyrics to a comedy song titled: “She doesn’t know she’s beautiful – what else doesn’t she know?”
11:00 pm: Paley and I go for a walk.
11:30 pm: Richard and I watch an episode of Bob’s Burgers, and then we fall asleep.
Reach out to Allison on LinkedIn.
Yorumlar