Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

What the interns have wrought, 2020 edition

blog.janestreet.com

Post reply on HN
Print view View original

171–180 of 219 posts

Re: What the interns have wrought, 2020 edition

#171

Post by kevstev »

nickbauman wrote: post #164
kevstev wrote: post #161

Earlier quoted context omitted.

Technically HFT is a subset of algorithmic trading. Algorithmic trading can be done on both the buy (IE hedge fund) and sell (Investment bank) side, but does not need to be high frequency. High Frequency also means low latency- sub millisecond, and I have never seen an algo trading system of any sort that deals in latencies that are not in the millisecond range, though this isn't really a requirement. Does this help?

Yes. So seller A wants to sell for $1.00 and buyer B is willing to buy at $1.06, an algorithm will calculate the presence of this potential (or actually have knowledge from other systems that this a certainty) and purchase seller A's stock with a margin of overhead of $1.03 and sell to buyer B for $1.06 netting $0.03. The algorithm is essentially a parasitic entity.

What you are describing is market making, and has existed pretty much as long as there are markets. Before computers rendered guys in the pits extinct, the spreads were much wider and those guys were known to literally front run you on occasion. You had legions of brokers and floor traders all taking home fat six figure paychecks each year and your trading costs were in the tens of dollars, not pennies. Those legions of traders have been replaced by a much smaller number of programmers who take home comfortable six figure paychecks.

They take on obligations as market makers to always buy and sell no matter what the market is doing.

How that worked was that he got a phone call saying "buy me 50k IBM" from a broker. Guy on the floor thinking smart money is behind this, or even not thinking there is smart money behind this, he just knows an order like that is going to move the market- will place his own order ahead of it and likely exit the trade immediately after. The advent of electronic trading, even when people were still on the floor, made this much harder to do since it became easier to slice and dice that order up to a bunch of different people, and there were now much better paper and electronic trails of the activity. Now its all computers talking to one another and this is not possible.

Re: What the interns have wrought, 2020 edition

#172

Post by atq2119 »

yellowstuff wrote: post #150
nwsm wrote: post #146

Earlier quoted context omitted.

> moving money around isn't that societally valuable I would say hedge funds have massive negative social value (especially quant/HFT ones), and actively contribute to wealth inequity. Robert Mercer is a prime example of this. Not saying I think your friend is a bad person or should quit, but it's kind of a naive justification.

I'm curious why you think HFT is harmful. They do the same job that human market makers used to do, buy vastly more efficiently, leading to lower prices for "customers" (IE lower spreads), and lower profits for the industry: > TABB Group estimates that US equity HFT revenues have declined from approximately $7.2 billion in 2009 to about $1.3 billion in 2014. https://web.archive.org/web/20140404072855/http://tabbforum…

HFT specifically is a form of largely zero-sums arms races. Is it really useful to society if there are entities that issue orders with microsecond (or less) response times rather than on the order of seconds or minutes? Yet a significant amount of brainpower and resources is essentially wasted on this problem.

The story about lower spreads is also rather dubious. I can believe that going to sub-second HFT reduces spreads slightly, but what's the point? By how much, exactly, is the spread reduced by going to the extremes that HFT goes to, and how does it compare to just your regular intra-day swings?

What's useful to society at large is long-term capital allocation. Regular people don't do day-trading, they place orders "at market" maybe a few times per year at most (or perhaps monthly as part of an automatic plan). The loss from intra-day variations will dwarf the measly reduction in spread that is achieved using sub-second HFT.

So in that light, it's good to see that apparently HFT is becoming less profitable.

Re: What the interns have wrought, 2020 edition

#173

Post by cltby »

nickbauman wrote: post #167
cltby wrote: post #165

Earlier quoted context omitted.

Can you explain to me very carefully why the orders of seller A and buyer B weren't able to match before Mr. Evil HFT came along? Moreover, given that buyer and seller weren't able to match without Mr. E. HFT's help, hasn't he in fact rendered a valuable service (for which he should, of course, be paid)?

So that's the missing part you say? The Internet has been the biggest catalyst of disintermediation in history. But this is not disintermediatable without HFT?

My question is pure nuts and bolts.

A and B wake up one morning and want to trade MSFT. Both submit orders to "the market." You claim HFT somehow inserted itself into the middle of the transaction.

I'm asking if you can fill in the blanks on how this is possible.

Re: What the interns have wrought, 2020 edition

#174

Post by nickbauman »

erik_seaberg wrote: post #170
cltby wrote: post #165

Earlier quoted context omitted.

Can you explain to me very carefully why the orders of seller A and buyer B weren't able to match before Mr. Evil HFT came along? Moreover, given that buyer and seller weren't able to match without Mr. E. HFT's help, hasn't he in fact rendered a valuable service (for which he should, of course, be paid)?

The HFT needs latency low enough to win a race against B.

Thanks, that's what I thought.

Re: What the interns have wrought, 2020 edition

#175

Post by nickbauman »

kevstev wrote: post #171
nickbauman wrote: post #164

Earlier quoted context omitted.

Yes. So seller A wants to sell for $1.00 and buyer B is willing to buy at $1.06, an algorithm will calculate the presence of this potential (or actually have knowledge from other systems that this a certainty) and purchase seller A's stock with a margin of overhead of $1.03 and sell to buyer B for $1.06 netting $0.03. The algorithm is essentially a parasitic entity.

What you are describing is market making, and has existed pretty much as long as there are markets. Before computers rendered guys in the pits extinct, the spreads were much wider and those guys were known to literally front run you on occasion . You had legions of brokers and floor traders all taking home fat six figure paychecks each year and your trading costs were in the tens of dollars, not pennies. Those legion…

Right, gotcha: the market needs intermediaries to function properly.

Re: What the interns have wrought, 2020 edition

#176

Post by cltby »

erik_seaberg wrote: post #170
cltby wrote: post #165

Earlier quoted context omitted.

Can you explain to me very carefully why the orders of seller A and buyer B weren't able to match before Mr. Evil HFT came along? Moreover, given that buyer and seller weren't able to match without Mr. E. HFT's help, hasn't he in fact rendered a valuable service (for which he should, of course, be paid)?

The HFT needs latency low enough to win a race against B.

If HFT and B are racing for the same order, they evidently have a very similar trading objective. In which case it's bizarre to ascribe some uniquely negative social value to HFT winning that race.

Re: What the interns have wrought, 2020 edition

#177

Post by claydavisss »

Scarbutt wrote: post #131
hvs wrote: post #58

Earlier quoted context omitted.

I think Jane Street is single-handedly keeping OCaml a production-level language (which is a good thing, because it's a great mix of functional programming with a practical/pragmatic approach). If you haven't tried OCaml (or F# which is based on it), I highly recommend it.

Is there any reason learn OCaml over F# for a new project?

To turn this around a bit, I don't think dispensing with F# is that difficult. Practically everything interesting comes from the .Net SDKs and as a result, any nontrivial F# code will be sprinkled with C#-style API calls etc. I expected a friendlier Haskell but then realized it was more like a Typescript for C#.

Re: What the interns have wrought, 2020 edition

#178

Post by dom96 »

wk_end wrote: post #145
dom96 wrote: post #66

Earlier quoted context omitted.

I don't know about anyone else but I have a negative perception of tech companies in the financial sector. Perhaps my perception is wrong, so I'm curious, how does the culture compare to one at Google or Facebook?

You can't really compare JS to Google or Facebook. It's still a company of * Very flat hierarchy and very limited bureaucracy. At JS there's like at most two people between you and the folks who run the place. * JS is technically isolated. They're not really part of the same conversation as the web-focused and much more public tech giants are. Lots of NIH, custom hand-rolled solutions. Maybe part of that is a consequ…

> More of a 9-6, butt-in-chair mentality than at other tech jobs I've worked. There's some flexibility about the 9-6 thing (and 9-6 can be pretty optimistic at times), but they're limited by the nature of the markets. At the time and for me, anyway, there was zero possibility of ever doing any work at home. COVID may have changed this.

This sort of thing precisely. My perception of these companies is that they are not as understanding about work-life balance, PTO, working from home, and ultimately trusting you to be independent.

There is also a more corporate image that these companies have in my eyes, one thing that would likely exemplify this is the dress code. Does Jane Street empower you to be your true self or do they prefer that everyone is the same kind of suit-wearing professional that banking institutions often want?

In general, all of this together gives me the perception that working at such a company would be far more intense. Working at a FAANG is sometimes stressful enough, knowing that there is a flexible and (what I at least consider) a progressive culture backing the company makes the stress easier to manage in some ways.

Re: What the interns have wrought, 2020 edition

#179

Post by nickbauman »

cltby wrote: post #173
nickbauman wrote: post #167

Earlier quoted context omitted.

So that's the missing part you say? The Internet has been the biggest catalyst of disintermediation in history. But this is not disintermediatable without HFT?

My question is pure nuts and bolts. A and B wake up one morning and want to trade MSFT. Both submit orders to "the market." You claim HFT somehow inserted itself into the middle of the transaction. I'm asking if you can fill in the blanks on how this is possible.

I thought that was clear: prior to HFT, the seller would have made the deal with the buyer without the intermediary HFT because if the latency. HFT exists because of speed. Very hard to say "High Frequency Trading" without saying "High Frequency".

Re: What the interns have wrought, 2020 edition

#180

Post by fractionalhare »

dom96 wrote: post #178
wk_end wrote: post #145

Earlier quoted context omitted.

You can't really compare JS to Google or Facebook. It's still a company of * Very flat hierarchy and very limited bureaucracy. At JS there's like at most two people between you and the folks who run the place. * JS is technically isolated. They're not really part of the same conversation as the web-focused and much more public tech giants are. Lots of NIH, custom hand-rolled solutions. Maybe part of that is a consequ…

> More of a 9-6, butt-in-chair mentality than at other tech jobs I've worked. There's some flexibility about the 9-6 thing (and 9-6 can be pretty optimistic at times), but they're limited by the nature of the markets. At the time and for me, anyway, there was zero possibility of ever doing any work at home. COVID may have changed this. This sort of thing precisely. My perception of these companies is that they are no…

Almost no one at Jane Street wears suits. The tech-driven buy-side of finance is very different from the sell-side you seem to be familiar with. It's much closer to small tech companies.
Post reply on HN