I work for a midsize HFT firm in Chicago (headcount ~60), trading equity index futures. I broke into the industry after completing my Bachelors degree but decided to leave and take on further studies. I hold Masters & PhD

We arent solely MM, we do liquidity taking too.

I also played a lot of PLO in Macau when living in Asia.

What's your average profit/contract? Is it purely a game of having the lowest latency or is everyone creating unique strategies?

Profit per contract is thin, but we compensate by doing alot of volume. Having lowest possible latency is obvs very important, even if youre model is good, you still have to get the trades you want so the infra is very important. We dont want to play the latency game as technological overheads can be very expensive and because speed and latency is relative, there is someone who will always try and ink out an extra nano, and for deterministric exchanges like CME, this extra nano matters. Rumour has it if your 10nanos slower, youll just never get the 'obvious' trades.

We are aiming to be smarter, as we think we have kind of reached the stage where you can only squeeze out a few nanos (if that) from the tech stack. There are a few in this space, who get the obvious trades, and they do it very well, Others have designed really smart systems and algorithms.

* How does market-making work. Do you accumulate a position if the market continues to go in one direction?
* How do you manage risk? What level of stress can you endure?
* Average profit, average loss, average duration?

Market making just means we quote both sides of the spread. We arent obligated to quote, but MM is just one strategy. We have a fair value and we trade around this. How you decipher this fair value is non-trivial. Market making is about inventory management, but you also need to have some predictive capability to prevent getting run over. As long as the market moves, we are happy either way.

Can hedge with a similar product or just bear inventory risk for some period of time. We hold positions for about 10 -15 seconds. I dont get stressed nowadays, as its all very systematic. I run a headless system and spend 80% of my day researching and all i see on my 1x 27 inch screen is the overall risk metrics and pnl. Wont disclose profits. Would be easy to see who i work for.

Can you speak more about inventory management?

How do you manage inventory risk? Let's say you accumulate an inventory of long S&P e-mini contracts, and it starts to selloff/crash. Do you use stops, buy puts etc?

I didn't mean your personal level of stress, I meant market stress/shock lol.

Can you share profit to loss ratio and win-rate?

Thanks for making this thread btw.

There are many ways you can hedge, granted you can trade and hedge with the back months etc. We are ok with risk and dont look to hedge immediately. so a rudimentary view of what we do is just hold until we can offload it passively, this partially comes back to being smarter. It used to be that market making was a computer science problem, buts gradually less and less so. If you have some forecasting ability and forecast a big sell off, worst thing that can happen is you cross the spread and submit market orders.

With market making, we are making money every minute, with prop stuff there is alot more variance in pnl . perwin is good, but not the best from what i have heard of others having.

Do you have to be the utmost expert in an instrument to be a market maker for it? Otherwise you'd get run over by buyside traders?

How many instruments are you a market maker for?

How high are the trading fees you pay for your market making?

Does the exchange compensate you in any way for being a market maker?

Thanks!

do you think there's one firm/setup that's *the* fastest? or is the speed advantage more widely distributed/fluctuating? what percent of the pie, so to speak, could the very fastest setup capture?

even if you aren't trying to be the fastest , do you guys still find the need to pay the co-location fees?

do you do options market making?

in your space, who are considered the best?

You dont have to be an expert in any given instrument, it obvs helps, but there is a structured relationship between alot of products and things tend to move in tandem. if you are an expert, youll capture things that others wont, this is pure alpha.

approx 600

we have membership fees etc but no other perks

no rebate for us unfortunetly, some markets do, but less liquid.
if i had to take a guess, id say jump or tower are the fastest due to the number of infra guys they have relative to researchers and traders. could be very wrong here. Speed always varies among the top shops with code now larely written to hardware where trading logic is more 'dumb'

mid-freq is slowly becoming high frequency, so yes, colo is needed. as you decrease the frequency you trade, the better your models need to be.
not right now, have done so in the past with a good firm
wouldnt know. probably an unknown firm. though there are firms im sure alot in hft wouldnt have heard about who ink money.

so there are a bunch of firms essentially racing to capture the inefficiency between the value of the future vs. the sum of the individual components of the future's underlying in cash/equity mkts?

thx for this thread, good to see things from the other side of the fence.

I guess you can phrase it like that, but even things like the etf vs futures, but there are lots of ways you can trade a product. in the end i think it benefits the retails guys alot, that is, reduce volatility, reduce bid ask spread and increase liquidity.

There are firms who dont compete in the race with tech, they are just smarter, who its not all about speed, though it doesnt hurt.

EDIT: I must preface by saying that if you can do this efficiently, you end up with all the money, ie futures vs cash

What do you enjoy the most about your job?

What do you dislike?

The work is extremely fun and challenging. the hours can be long(ish) ~ upto 12 hours a day.

What's the pay like? Have you ever thought about going off on your own or would higher latency systems not suit your skills or be too much of a handicap?

somewhere between 175-200, slightly below industry standards, but better perks.

it has crossed my mind then i see the work people put in and then realise it would take 5 years or so to be competitive and lots of millions of dollars.

How wide would you quote the S&P e-mini under "normal" market conditions?

Why did you leave? What are you pursuing now?

impossible to answer in a vacuum. the market is extremely dynamic. we want to trade and actually provide liquidity so its not wide by any stretch of the imagination
didnt leave.

175-200 is interesting to me, i would have assumed with such an insanely horizontal strategy that more could be pulled out of the environment. must be incredibly competitive. what's keeping you from scaling up in the size of your quotes/lifts?

175-200k usd is the salary. There is lots that go in MM strategy and indeed competitive. Scaling up upto some point works and after you’re trading with informed flow/toxic flow

How involved are you in commodities, especially spreads. Things like the crush (beans) or the crack (oil) or yield curve spreads? Things like these and the cross exchange NG and CL spreads were some of the things I cut my teeth in prior to HFT. Back in the early days of screen trading these things were the wild west. But once HFT came in it pretty much killed the individual market maker.

NG and CL very active, the others not at all. The NG and CL trade on various exhcnages so opportunity for lead/lag arbs. ie CME & MCX. Outrights are more fun

What did you mean by "but decided to leave and take on further studies"?

What skills are HFT firms looking for? Obviously programming and statistical modelling, but anything specifically?

Would a HFT firm seriously consider someone with just a bachelors?