Forums  > General  > Is anything interesting happening in quant finance right now?  
     
Page 1 of 2Goto to page: [1], 2 Next
Display using:  

EspressoLover


Total Posts: 320
Joined: Jan 2015
 
Posted: 2018-01-30 11:50
Like the title questions says...

Is any group/desk out there doing anything cool and interesting right now? Something like convertible bond arbitrage in the 90s, stat-arb circa 2001, structured credit circa 2006, HFT circa 2008-2011 or vol-arb circa 2014. Or has low vol, long-beta mania and trading firm consolidation taken the winds out of everyone sails?

My situation's such that I'm kind of out of the mainstream gossip networks, but afaik things look pretty boring right now. Crypto's maybe the only thing I can think of. (But c'mon do blockchains have to be the hype vehicle for literally every single field on Earth?) Anything you guys are excited about right now? Random speculation, reckless gossip, and uncorroborated hearsay are definitely welcome. Even if you have to pepper it with some BS, get me pumped up about something!

Good questions outrank easy answers. -Paul Samuelson

Maggette


Total Posts: 1046
Joined: Jun 2007
 
Posted: 2018-01-30 12:45
Nope. I am out of finacne for a while in parts because of that. Would also be happy to hear interesting stories.


Ich kam hierher und sah dich und deine Leute lächeln, und sagte mir: Maggette, scheiss auf den small talk, lass lieber deine Fäuste sprechen...

tradeking


Total Posts: 16
Joined: May 2016
 
Posted: 2018-01-30 15:05
Alternative Data
Machine Learning
Cryptocurrencies

are the biggest fads going on right now

svisstack


Total Posts: 308
Joined: Feb 2014
 
Posted: 2018-01-30 15:17
Huge amount of algorithmic/hft traders from the US stocks and things like that are interested in the crypto, we helping them at CoinAPI, so there is a lot of interest in the Cryptocurrency space which is kind of different as in the past there was a disconnection between normal markets and crypto ones.

Time well wasted.

EspressoLover


Total Posts: 320
Joined: Jan 2015
 
Posted: 2018-01-30 18:02
> Alternative Data
> Machine Learning
> Cryptocurrencies

At the risk of moving my own goal posts... Do you think anyone's making real money on these fads? (Besides collecting AUM fees from hyped up investors)

Good questions outrank easy answers. -Paul Samuelson

Osiris2


Total Posts: 17
Joined: Sep 2017
 
Posted: 2018-01-30 18:03
What ever happened to mining predictive signals from social media or google search queries? Is that dead yet?

chiral3
Founding Member

Total Posts: 5059
Joined: Mar 2004
 
Posted: 2018-01-30 19:06
Not really on relative bases (relative to 10 years ago) but yes on an absolute basis. At least in the US and Europe. With regulation and a prolonged cycle this time around we’ll continue to see low vol and high(er) valuations with less edge in developed countries.

Most quants associate quant finance with the sales and trading paradigm. In that respect things are in stasis and will continue to be suppressed through regulation and punitive disincentives from on high. We are seeing fits and starts of another wave of M&A (that’s flush with Asian capital) that will fade the end of the credit cycle, beginning a cycle of integration. Banker bonuses were shit not too long ago when S&T was good and now the tables have turned in the cycle.

Wanting to be at a successful quant fund in the US today is like a 13 year-old wanting to be Lebron James. On the flip side, the definition of edge in this low-vol environment is being redefined for people that are adaptable, flexible, and thoughtful. The decade from 1998-2008 was really the only time where a pure quant wasn’t completely subjugated. The timing was off since it was a lagged effect riffing off the rate markets born of the 1980 Volker FED, initially, and then gen1, 2, 3 of equity derivs. I know plenty of quants that have been related to utter serfdom. Many make good money, few are happy, and even fewer have made a ton of money. The game is the same, though: we deal in intangibles. The least successful quants are clever and a complete nightmare to deal with. The most succussful ones are clever, thoughtful, and possess that inexplicable quality of “getting it” that allows them to harvest true value from dealing with intangibles.

Sentiment continues to drive a ton of absolute movement in asset prices. The noise floor is just too low still for anything else to rise above it. I saw index correlation pop by 10 points a couple of weeks ago and realized index vol hit a 10 handle for the first time in recent memory.

While regulation has been a pain, the pattern recognition problem of identifying patterns in regulatory filings, balance sheets, and rules has become a huge textual analysis problem on the regulatory corpus that people are solving and deriving relative value signals with.

Disintermediation via fintech is only just beginning even though it feels like it has been in full swing for a while. Sketch out any industry and work to indentify middlemen, look at what these entities do, and think about how to replicate and add value. Goldman literally did this internally and many of the partners that are out of a job are trying to do this in start up space.

If someone wants the comfort of leveraging existing skills they can look to less developed economies.

Looking farther ahead, more pie in the sky, beyond cloud, GPU, etc. I really believe that quantum computing will be leveraged in my lifetime. Most people think of crypto when they think of QC. However, it’s kind of a fun thought experiment to try and solve classical problems quantum mechanically. When I was studying physics, this would have seemed stupid and unnecessary. Kind of like using GR for asset pricing problems (this actually got quants excited a while back). But with QC the most computationally expensive classical problems become cheap.

I know nothing about cryptocurrency, particularly what most people associate with it - speculation - so I will not comment. CC has become more about identity politics and fashion in the US. In the US it’s about some asshole (without the balance sheet) ruining a dinner party arguing their cognitive biases. I am excited to see of any applications to developing countries that benefit those countries socially.

Final thought: quants can always go teach. Even if you have 2 years of experience you can start a blog, latch onto a particularly spicy identity politik, and write a book, like “Weapons of Math Destruction”. Next stop TED.

Nonius is Satoshi Nakamoto. 物の哀れ

ronin


Total Posts: 317
Joined: May 2006
 
Posted: 2018-01-31 11:36
> What ever happened to mining predictive signals from social media or google search queries? Is that dead yet?


Pretty dead. A few firms still have some VC money left so they are still ticking, but that's about it.


"There is a SIX am?" -- Arthur

frolloos


Total Posts: 38
Joined: Dec 2007
 
Posted: 2018-01-31 11:45
So basically quant finance is dead? Is it fair to say that wére just tweaking existing ideas instead of coming up with new big ideas a la Black Scholes Merton? And how has regulation contributed to stifling development?

chiral3
Founding Member

Total Posts: 5059
Joined: Mar 2004
 
Posted: 2018-01-31 12:41
I would think that all of us get to work on interesting things everyday. In that regard, life is interesting and fun. I get to work on interesting and varied things, determine strategic direction, direct large resources at things I think are meaningful, read about what I am doing or am involved with in the news from time to time, and we all get paid pretty reasonably well. Maybe the nuance in the question is whether there's anything really pivotal happening more broadly (which is how I thought of it) and what does it mean to have a some agency in determining what should be interesting. Man's search for meaning or some shit like that.

Nonius is Satoshi Nakamoto. 物の哀れ

pj


Total Posts: 3400
Joined: Jun 2004
 
Posted: 2018-01-31 12:45
What about XVAs (value adjustments)?
At least I would consider them as a job security for quants.

The older I grow, the more I distrust the familiar doctrine that age brings wisdom Henry L. Mencken

mtsm


Total Posts: 217
Joined: Dec 2010
 
Posted: 2018-01-31 17:03
Quant finance means different things to different people here in this thread.

Some here (pj, frolloos, ...) think of quant finance as being the body of knowledge applied to some extent to sell side market activities. They think pricing models, derivatives, risk and such...

To everybody else here including the OP it's about general applications of quantitative models and methods to buy-side activities and to fintechs or other startups.

Kind of different perspectives if you ask me. I guess chiral3 made this point among others.

From my perspective and since I am meeting with a lot of people on the job market nowadays, I am seeing the following (probably somewhat biased by my background):

* Alternative data (mostly for equities markets) in various funds.
* Some crypto efforts in institutional space, but not actually that much. Both in investing and fintech/services space.
* Attempts to systematize/codify (famous investor's) macro approaches and thinking in macro shops. (The anti-ML approach...)
* Large scale data platform build out and AI/ML related efforts in multi-asset shops.
* Applications of Distributed Ledger Technology in startups.

EspressoLover


Total Posts: 320
Joined: Jan 2015
 
Posted: 2018-02-01 12:14
@chiral

Thanks very much for the very well-thought out response. There's so many nuggets of wisdom there, that I kind of want to unpack everything. I'll try to restrain myself to the highlights...

> The decade from 1998-2008 was really the only time where a pure quant wasn’t completely subjugated.

If I'm grokking you right, you seem to be saying that 1998-2008 was the aberration and post-2008 is probably closer to long-run ergodic normalcy. This viewpoint (which is definitely somewhat of a bummer) never occurred to me. I always thought of quant-ishness' rise as driven by secular trends: cheaper computing power and increasing market transparency. And its post-2008 decline as a cyclical trend: QE, high regulation, deleveraging, etc. Hence the expectation that long-run normal will look a lot more like 1998-2008 - secular trends march on, and cyclical effects fade away. But your counter-narrative is pretty compelling, so I'm reconsidering...

> Sentiment continues to drive a ton of absolute movement in asset prices. The noise floor is just too low still for anything else to rise above it.

You mentioned something like this in another thread, and I'll just remark that it's a great insight on the current market regime. It never occurred to me before, but after hearing it seems so obvious and descriptive.

> While regulation has been a pain, the pattern recognition problem of identifying patterns in regulatory filings, balance sheets, and rules has become a huge textual analysis problem on the regulatory corpus

Another great insight. Lemons to lemonade

> Disintermediation via fintech is only just beginning...

Don't disagree, but at least as an outsider, it doesn't really seem like quants have any comparative advantage in this space. Is a company like Stripe or SoFi really doing anything more complex than what MBAs do in Excel everyday? Maybe they're doing it with more scale than the regional consumer bank. But that's more an engineering problem ("do this simple calculations, but at 100,000 transactions per second") than a quant problem ("how the hell should we even go about pricing this thing"). If anything, adtech seems like a more natural home for an emigre from quant world.

> However, it’s kind of a fun thought experiment to try and solve classical problems quantum mechanically.

I'm getting off on a tangent here, but I'd recommend Quantum Computing since Democritus. Scott Aaronson's approach to quantum mechanics reminds me a lot of what you're saying. (Well more with a CS perspective than a physics perspective). But it's a good take on QM qua QM. What can we say about QM as a pure mathematical system? What are the most fundamental differences between quantum and classical systems? Why probability amplitudes instead of probabilities? How big are quantum states relative to classical states? Compared to the standard textbook treatment, which is more about how to use QM to solve problems, this book's trying to posit why might QM have these properties in the first place. Aaronson makes a pretty convincing case that QM isn't just some weird set of arbitrary rules, but there might be pretty deep philosophical reasons to expect reality to be quantum even a priori.


@mtsm

Yeah, I'm definitely coming from a buy-side/trading biased background. It probably seeps through, but definitely still interested to hear about stuff in sell-side/pricing world.

Thanks for posting that list. It's pretty interesting, and some of the points hadn't occurred to me...

* It will be interesting if systematic macro really gets off the ground. Especially given how poor most regular macro managers have done the past few years. What's relatively unique about that space is how small the data sets are. If you're forecasting monthly moves in major FX pairs, you're not getting much more than 5,000 data points even if you go back 40 years. So it's definitely understandable to see why starting with good priors is really important. At least compared to some large data set where you have enough degrees of freedom to mine for signal in a black box.

* Alt-data could definitely be the next big thing. But I wouldn't bet the farm on those signals having high R-squared. It's kind of tough, because even if you have something like credit card transaction data, you don't necessarily know what's a good vs. bad number. Earnings come with analyst forecasts, so it's easy to divide bearish from bullish results. But most alt-data doesn't have any equivalent baseline for comparison. Like yeah, maybe credit card transactions indicate a 4% uptick in sales this quarter. But how do we know that wasn't already baked into market expectations? It seems like regular old meatspace analysts will be able to contextualize that data better than algorithms.

* Distributed ledgers are the only thing on that list I'm unequivocally bearish on. (Which sorta sucks, because back office deserves to get a cool project once in a while.) Distributed ledgers certainly have specific uses and unique advantages. But I think 95% of the demand for the tech is simply coming from senior execs who've been sold on the idea that it eliminates the cost and headaches of running a persistent service. This seems plausible because bitcoin operates pretty seamlessly without a Bitcoin LLC paying huge AWS hosting fees and waking up their CTO at 3 AM. But that's basically because the miners get paid assloads of money to make that happen, like orders of magnitude more assloads than a DBA team.

Good questions outrank easy answers. -Paul Samuelson

Maggette


Total Posts: 1046
Joined: Jun 2007
 
Posted: 2018-02-01 16:12
"Alt-data could definitely be the next big thing. But I wouldn't bet the farm on those signals having high R-squared. It's kind of tough, because even if you have something like credit card transaction data, you don't necessarily know what's a good vs. bad number. Earnings come with analyst forecasts, so it's easy to divide bearish from bullish results. But most alt-data doesn't have any equivalent baseline for comparison. Like yeah, maybe credit card transactions indicate a 4% uptick in sales this quarter. But how do we know that wasn't already baked into market expectations? It seems like regular old meatspace analysts will be able to contextualize that data better than algorithms."

To me that kind of boils down to the question of "values vs price". Assuming I do get data that somehow has an impact on the "fundamental outlook" of a company. Who cares if that is not really represented in the prices...becuase I am the only one who gathered and analyzed the data? Even if you believe in markets beeing that efficient and that a (probable) drop in earnings or rise in costs for the respective company forecasted by my alternative data crunching will effect prices...you have to get the timing right and still have to know when it happens and to what extend.

Ich kam hierher und sah dich und deine Leute lächeln, und sagte mir: Maggette, scheiss auf den small talk, lass lieber deine Fäuste sprechen...

ronin


Total Posts: 317
Joined: May 2006
 
Posted: 2018-02-02 12:55
The problem with alt-data is a bit more conceptual.

Alt-data can't reconstruct material non-public information. And if it did, it would be illegal to trade on it.

So it mainly means taking non-material public information, and turning it material. But that won't go very far. Once the market catches up with the fact that (say) credit card clicks constitute material information, credit card clicks will become subject to standard disclosure rules. If you are lucky, you may get a payout or two before that happens, but then you are back to square 1.

Not to mention that there is a lot more to any price than sales upticks and downticks. At best, you can get a bit of a partial picture - which can be worse than no picture at all.

Say your uptick in sales is due to a big increase in cost of sales, and margins are actually negative. Is that good or bad?

"There is a SIX am?" -- Arthur

heyman


Total Posts: 1
Joined: Feb 2018
 
Posted: 2018-02-02 13:28
Buying home furnishings online has so many advantages like – plethora of choices, lowest prices, prompt delivery, and best services. In short, online shopping gives you enormous options and one can have a choice to search until they get their favourite colour, design, and style etc. Nothing can be more happening than the limitless options, freedom of selection, and finding things from any part of the country.

We offer Fine Home Furniture and Accessories for your home, office, condominium or rental property, Visit our store & buy Home Decor & Modern Furniture Online. modern furniture online

deeds


Total Posts: 393
Joined: Dec 2008
 
Posted: 2018-02-02 13:56
Ronin -

may be showing my age, things may have changed

'Alt-data can't reconstruct material non-public information. And if it did, it would be illegal to trade on it.'

Wouldn't this kind of activity generally come under "Mosaic Theory" of materiality...
https://en.wikipedia.org/wiki/Mosaic_theory_(investments)

seems certain that AI/ML implementations will be a test of how robust that 'theory' is

Maggette


Total Posts: 1046
Joined: Jun 2007
 
Posted: 2018-02-02 14:18
@deeds and ronin

Very interesting.

I guess there is also a kind inheritance here? I mean, beeing able to identify insider traders or market manipulators and just run with them?

For example if I use machine learning to identify successful "pusher" accounts. "Influencer" of the dumb money in the pink sheet world.

And if I buy the stocks they are pushing, I guess I am doing something illegal?


Ich kam hierher und sah dich und deine Leute lächeln, und sagte mir: Maggette, scheiss auf den small talk, lass lieber deine Fäuste sprechen...

deeds


Total Posts: 393
Joined: Dec 2008
 
Posted: 2018-02-02 16:59
Maggette -

I think the case you are describing would be excluded in practice...the demands for being able to explain how you assembled the mosaic are pretty high.

If one element was 'identification of insider trading' I think there would be an issue.

ronin


Total Posts: 317
Joined: May 2006
 
Posted: 2018-02-02 16:59
@deeds,

Of course. Ultimately, that's what analysts are paid for - take public information, and construct a price target.

Didn't Raj Rajaratnam famously ask female analysts to try out some tights made by the target company to work out if it's a good product that people would buy?

But there is still a conceptual problem, which I think people are largely just glossing over.

Here is my classification of information, a-la Donald Rumsfeld / Goedel:
- "known knowns" - that would be all past numbers, content of all announcements and press releases, plus all published analysis. This isn't really a target for alt-data or AI. Maybe you can construct a better valuation model, but not through alt-data or AI.
- "known unknowns" - this would be current quarterly numbers, up until the day they are disclosed. Unless you are an insider, you don't know what they are. You may steal them, but basically we are in criminal territory here.
- "unknown unknowns" - all sorts of sh.t that may be somewhere behind the horizon, from geopolitics to new competitors to changing customer preferences to who knows what. You don't know, insiders don't know - basically, a wild guess.

None of these have much to do with alt-data or AI. The only place left is number 4, which would be
- "unknown knowns" - stuff that people know about, like credit card clicks, but it doesn't really mean anything because it's noisy partial information.

So before I spend a ton of money on chasing some clever alt-data, the first thing I would quantify is how much price uncertainty is due to the "unknown knowns", as opposed to "known unknowns" and "unknown unknowns". And as a wild guess, I wouldn't say too much.

Does that make sense?


"There is a SIX am?" -- Arthur

involution


Total Posts: 6
Joined: Nov 2015
 
Posted: 2018-02-02 21:43
@ronin

Thanks for your very interesting perspective. Are you saying that using alt-data to estimate "known unknowns" (sales, margins, growth, etc...) is likely to constitute insider trading? Is that because such data sold by vendors may be prohibitively expensive (and hence not considered publicly available?). What do you think about the widespread use of expert networks (also quite expensive)?

chiral3
Founding Member

Total Posts: 5059
Joined: Mar 2004
 
Posted: 2018-02-02 21:46
Today is a great day. It feels... normal again.

Nonius is Satoshi Nakamoto. 物の哀れ

ronin


Total Posts: 317
Joined: May 2006
 
Posted: 2018-02-05 11:57
@involution,

Public companies have to release their earnings to all investors at the same time. You can look them up on their websites, on the relevant exchange websites, on yahoo finance etc. But generally speaking, the fact that there may be a subscription fee attached doesn't change the status of the data. If it is published to paying subscribers, it is considered to be in the public domain and it can be traded on.

To answer your first question, if you can reliably reconstruct confidential numbers, then there is a breach somewhere. Either you are effectively an insider even though you may think you are not, or the company has published some information that it shouldn't have. Either way, the legal risk is high.

"There is a SIX am?" -- Arthur

jslade


Total Posts: 1132
Joined: Feb 2007
 
Posted: 2018-02-06 04:09
Fintech is an astounding shit show. People in this space are insanely, preposterously clueless as to very basic concepts like risk models ... hell linear regression and discounted cash flow models. They're still eating up business of classical finserv companies simply by putting it all on a website and burning VC money. If you want to make a big difference as a quant, go fix one of these companies.

Crypto: most of the shit you see are, as you expect, speculative bubbles on bullshit. I still have made ridiculous returns on it; I'm kind of in shock this is how I've managed to make my number. The trading stuff; there is plenty to do here. Old fashioned algos which doesn't work in equities or futures any more work just fine in crypto for trading. There are also ordinary services to provide; making options markets or finding liquidity for example. For trading, the biggest problem is exchange risk. It's actually a very big problem; the Bitfinex/tether/goofytoken thing is only scratching the surface of what's going on here. It may come out eventually; I should encrypt something and stick it here so I can say "I told you so" in two years without giving it away. There are also legitimate ways to use this technology to bring banking services to underserved communities (I'm advising a company doing this). And of course, it's also useful having a sort of distributed SMTP for money.

Old school quanting: there are at least 2 newish things in stock market "analysis" -one of which chiral alluded to (using machine learning on regulatory/filing/sentiment stuff), the other of which is basically modeling cash flows in different kind of corporations using Bayesian techniques, aka knowing what your ops cash flow and accrual components are and taking positions on these. Probably 1-2 others I haven't heard of; side info and so on -I'm mostly an outsider here, but people talk to me. I think most of the juice in stock market stuff now requires fairly large organizations to take advantage of; aka no more 5-10 man crews at SAC. I still think there are lots of interesting but illiquid opportunities in futures and options for teams of this size.

FWIIW I started writing papers in quantum computing in 2003. I don't think the field has matured at all since then, and I'm coming around to the belief that it's all a socially gamed cargo cult the way noodle theory was. If this was real, people would invest in hardware which works, but none of the hardware seems to work any better than it did at the Dartmouth Gordon conference I went to back in 2001 or 2002 or whenever it was. Meanwhile hundreds of half frauds bloviate theory articles on this subject; achieving tenure in a subject which is about an imaginary object that may never exist. Imagine if there were hundreds or a thousand computer programmers or "computer scientists" before anyone actually built a usefully functioning computer. That seems the present situation in this field. Would be happy to be proved wrong. Would also love to see a historical example of any actually existing technology which has developed in this way.

"Learning, n. The kind of ignorance distinguishing the studious."

deeds


Total Posts: 393
Joined: Dec 2008
 
Posted: 2018-02-06 13:46
jslade - thanks for the excellent insights

"basically modeling cash flows in different kind of corporations using Bayesian techniques, aka knowing what your ops cash flow and accrual components are and taking positions on these."

Possible to give any references or pointers for this type of theory/practice in context you mention?

Are these no arbitrage (risk neutral etc) approaches or forecasting approaches?

thanks

EDIT: added question
Previous Thread :: Next Thread 
Page 1 of 2Goto to page: [1], 2 Next