Forums  > Basics  > non-quant trading  
Page 1 of 1
Display using:  


Total Posts: 88
Joined: Apr 2018
Posted: 2020-04-06 15:30
Are there any good books/websites/resources to learn non-quantitative trading?


Total Posts: 248
Joined: Dec 2010
Posted: 2020-04-06 20:31
For which asset class?

Institutional or retail?

etc, etc, ....

This is a very vague question...


Total Posts: 57
Joined: Feb 2018
Posted: 2020-04-06 21:40

Offtop: I was always interested in what is the difference between "institutional" and "retail" trading in general?

Obviously, someone running 1B book is "institutional" and someone trying to time the market with his 300K savings is "retail". But what if someone runs an "almost market neutral" strategy (means that he allows some additional exposures here and there) on his own 10M account with daily turnover?


Total Posts: 88
Joined: Apr 2018
Posted: 2020-04-07 02:23
I know it's a broad question. Just trying to understand other types of trading that people do. Any asset class with holding period hours to weeks (not HFT, not buy and hold). FX, commodities, macro bets, stocks, trend following, technical analysis, etc etc. Who are some legit non-quant traders, and what is their thought process when making trading decisions?

Put it another way, where does their alpha come from? What is their trading edge?


Total Posts: 30
Joined: May 2016
Posted: 2020-04-07 04:19
Read the Market Wizards book series by Jack Schwager. Plenty of very successful non-quant traders interviewed (fundamental equities, technical charting, global macro, etc).


Total Posts: 69
Joined: Jul 2018
Posted: 2020-04-07 09:16
Take this with a grain of salt as I'm a 100% quantitative trader, but Michael Burry (of the Big Short fame) speaks very highly of Graham and Dodd's 'Security Analysis'. That being said I believe the book teaches a very classic view on value investing, which is an inherently quantitative notion (even if it doesn't involve complicated mathematics, which most trading doesn't). Secondly even if your alpha is not quantitative, your portfolio management should be and you can't go wrong reaching for Grinold and Kahn before anything else.

There's a great quote in filthys book, paraphrasing as I can't remember it exactly but essentially saying that all trading must be quantitative in some sense, otherwise you are just guessing (i.e. even if the model you use to estimate returns and risk is your brain).

did you use VWAP or triple-reinforced GAN execution?


Total Posts: 88
Joined: Apr 2018
Posted: 2020-04-07 10:00
@tradeking: I read Market Wizards a long time ago, and from what I remember there wasn't a lot of meat. I'll take a second read, thanks.

@doomanx: I tend to agree with your statement

> all trading must be quantitative in some sense, otherwise you are just guessing

But I have seen people who make consistent money for decades with zero quant analysis. They must be really good guessers!


Total Posts: 1062
Joined: Jun 2005
Posted: 2020-04-07 10:22
> They must be really good guessers!

or insiders with good coverup.


Total Posts: 418
Joined: Jan 2015
Posted: 2020-04-07 10:38
> Any asset class with holding period hours to weeks (not HFT, not buy and hold). FX, commodities, macro bets, stocks, trend following, technical analysis, etc etc. Who are some legit non-quant traders, and what is their thought process when making trading decisions?

George Soros' two books on trading (Alchemy of Finance and The New Paradigm) might fit what you're looking for here. I think arguably he's the most successful trader within your criteria.

They're definitely not instruction manuals, any more so than spending an afternoon with Terry Tao will turn you into a Fields Medalist. But it peels back a layer on Soros' thought process. And afterwards if you squint, you can maybe start imagining how he might approach different scenarios. But again those book won't give you any alpha unless you're already starting with something pretty close to Soros' mind.

Along similar lines, if you can get ahold of it, check out Trader, the 1987 documentary on Paul Tudor Jones. PJT bought up all the copies and had them destroyed, but I believe there's a torrent floating around Pirate Bay. I'd second the Market Wizards series. Opalesque TV has some good interviews with a wide array of people who might fit your description.

One final point is be careful about comparing those who were successful in the past with the present-day. If you're talking about an hourly timeframe, markets are *way* more efficient than they were 30 years ago. I think there were a lot of successful non-quant traders from the past who were basically proxying for what's now done by stat arb desks. If you're interested in this from a historical perspective, maybe look up some of the literature on the turtle traders.

Good questions outrank easy answers. -Paul Samuelson


Total Posts: 69
Joined: Jul 2018
Posted: 2020-04-07 11:16
@EL Soros books are a pretty perfect match here. I remember as a young lad I read some books by Jim Rogers (the other guy in the Quantum Fund during the 4200% period) that I thought were pretty good at the time, seem to remember they explained why they went long oil when they did and such, was all macro analysis and that kind of thing.

My first summer job in the city I got given a copy of Reminiscences of a Stock Operator, think PTJ used to do that too. Pretty entertaining book but not so relevant now. There was another book I was given called the Great Wave that had some interesting ideas.

did you use VWAP or triple-reinforced GAN execution?


Total Posts: 1225
Joined: Jun 2007
Posted: 2020-04-07 12:03
IMHO the problem with all these big names is, that all these legends might be survivors, and many of them in a spectacular way. But given their profile (lot's of severe losses...) I am afraid I to this day think they might be a huge bit of randomness and luck involved.

I think we are looking at a survivor-ship bias here.

Also many of these examples are quite old.

I would like to know how the discretionary traders at DE Shaw are working. What are they looking at? What are their holding periods/their trading frequencies?

I do believe that a cyborg, a combination of a human or a team of humans and lots of data and software tools, might add some value. I have seen this outside of finance a lot. But never in finance.

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


Total Posts: 88
Joined: Apr 2018
Posted: 2020-04-07 14:59
Thanks EL and doomanx for the suggestions. Will check them out.

I'm also wary of survivorship bias. I read somewhere that Buffett only does one thing (value investing) and it just happens to work in the long run. On the other hand, Soros makes many many independent bets and still makes money, so he couldn't just be lucky.

I'm also curious to know about discretionary traders in the present day at DE Shaw, Citadel, SIG, and potentially a lot more under the radar.

RE cyborg, I believe some relative value traders fall under this category. They look at quantitative signals but make discretionary bets, and possibly even leverage HFT infra to execute.


Total Posts: 418
Joined: Jan 2015
Posted: 2020-04-08 13:16

I really enjoyed Jim Rogers' travelogues as a kid too. Even if you completely overlook the investing discussions, it was fun to read about a super-rich dude driving around the world on a motorcycle with his model girlfriend.


I can't speak to DE Shaw specifically, but a lot of the funds with quant DNA take a unique approach to fundamental/discretionary equities. Basically you start with human analysts who stack rank the stocks in their universe. That's then treated as an alpha, which is dumped into a stat-arb portfolio optimizer pooled across the entire group.

I think this is actually a pretty good approach. One problem with human portfolio managers, even otherwise good stock pickers, is the don't do enough to diversify across orthogonal bets. The human tendency is to take your strongest conviction idea, and just plow everything into it. Even if risk management limits your single-name concentration, often times a PM's positions will just be different riffs on the same theme.

If Michael Burry had been running 25 orthogonal bets instead of being obsessed with a single thesis, maybe he'd only have half the final returns, but with virtually none of the drawdown. So, I think overlaying stat-arb portfolio construction can also drastically increase the Sharpe ratios even without any improvement to stock picking skill. The secondary benefit is that it becomes a much more tractable and less noisy challenge to assess the skill of your individual analysts.

Good questions outrank easy answers. -Paul Samuelson

Founding Member

Total Posts: 1363
Joined: Mar 2004
Posted: 2020-04-08 13:42
In terms of specifics - what non-quant traders follow, the data, the contact networks etc - it'll vary a fair chunk between different markets. In term of generic tendencies I'm sure there's a lot of common threads.

I'm not aware of any solid resources. Speaking from my own experience in the commodities world I can say there's a huge range of different types of non-quant traders doing pretty different things. Some people are focused on a small niche, nurture a big network of contacts they use to get a good feel for what's going on (e.g. people at all levels in the value chain), trade directionally and rely on being right a lot. Others look at a lot more markets, are more data driven, trade more RV but with a discretionary process factoring in also "soft" data, or data that is unreliable and hard to systematically collect.

I guess I'm saying it's a broad church.

Capital Structure Demolition LLC Radiation


Total Posts: 348
Joined: Mar 2018
Posted: 2020-04-08 13:45
@EspressoLover its not quite clear to me how this stat arb optimiser would work. Theres no residuals from which to measure the amount of dislocation. I would have thought that the stack ranks would be relatively changing


Total Posts: 418
Joined: Jan 2015
Posted: 2020-04-08 15:00
Yeah, you're right. Technically it's not kosher, because you're treating ordinal rankings as cardinal scores. But in practice, you can usually pick an arbitrary zero-centered mapping from rank to signal, and the exact details rarely make that much of a difference. Scaling magnitudes isn't that important, because turnover on a fundamental long/short portfolio doesn't incur that much in transaction costs.

For example, pretty much every academic paper on cross-sectional anomalies does this by assigning -1/0/+1 to the bottom decile, middle brackets, and top decile of the universe respectively. Even this simple approach is pretty much good enough, that it's been the standard approach for 50 years.

Although, I've never been directly involved in the process, I'm sure most of the real-world systems have some sort of accommodation for conviction. Both within an analysts' universe as well as between analysts. By this point, some of the funds have been doing something like this for nearly two decades. So, they should have pretty good empirical data for making evidence-driven tweaks.

Good questions outrank easy answers. -Paul Samuelson


Total Posts: 88
Joined: Apr 2018
Posted: 2020-04-10 05:07
I'm most curious about traders like this guy. He has been profiled in Bloomberg several times. He trades serious size in Nikkei from his bedroom, and judging from his Twitter it doesn't look like he does any sophisticated analysis.
Previous Thread :: Next Thread 
Page 1 of 1