This question is probably to those who work in prop trading, specifically to those who employ algos. I'm working on an equity stat arb strategy myself as part of my master thesis and I came across lot's of papers about various strategies that got me thinking do any of them matter at all? Just following common sense, if the strategy shows superior results there is no point in making it public, researcher should utilize it himself. If he makes it public, he gives out the edge which will quickly disappear. So my question is to those who work in the industry do you guys actually follow the academia? or your strategies are being developed in-house and are mostly different from what is publicly available in research papers?
Is there any value in research papers about trading strategies?
- Thread starter Arti
- Start date
Yes much value. There are several parts to strategy building. 1) Merging your trade ideas with fundamental economic laws 2) Merging your strategy building and testing process with a quality scientific process, paper can be very useful here for cross-industry treatments 3) Bounding creativity and dreaming within previously discovered limitations and sensibilities - and so while reading a bunch of dry academic papers with complex equations (not necessary to fully understand by the way) may seem like a dead branch effort, do it for a month and it'll save you from years of bad ideas.
I try to read academic papers with some regularity. Market neutral, Asset Allocation, Commodity forecasting, etc etc. I read this stuff mainly to spark unrelated ideas in my own mind and I go on to develop things that are proprietary and not published anywhere. I rarely have found anything of value explicitly defined in academic publications. If you are going write your thesis and you stumble upon a new discovery, don't publish it or even tell your professor. It will be worth more to you as a trading edge than any peer accolades. You should do what most academics do - produce a derivative re-hash of what's already been written with a slight twist or two.
Most value from those research papers is for those who "sell" them to promote their books, advisory service, custom software, etc.. or whose job is to produce those papers (professors / professional analysts) , but rarely there is any real value in practice, at least for small-scale/individual traders.
Last edited:
First, if you can't submit your "edge" to the rigor required to published than you probably don't have one and you're just data mining or on a lucky streak. That doesn't mean you need to or even should actually publish, it means that you need to subject yourself to that type of rigor or else just admit that you're merely gambling.
I would also submit that there are a large number of people in the world who are motivated by more than money. I know its hard for people on ET to believe, but its true. I know brilliant finance PhDs who regularly get lucrative offers from hedge funds but work in academia by choice. For them it is much more rewarding to publish a novel paper on a strategy that generates abnormal returns than to trade that strategy. Also, you have to remember that true abnormal returns, not the kind an ET punter claims to have gotten but those that have passed the rigor of publication, are on the order of a couple percent above market returns. Unless you're running a hedge fund, those aren't enough to get rich on and the PhD may very well get much more out of the prestige of publishing it than the $2,000 they'd get from trading their $100,000 portfolio with that strategy.
I would also push back on the concept that everything an academic does is derivative and that makes it worthless. That is kind of a fundamental misunderstanding of the scientific method which is basically to build on the accomplishments of those before you. By those standards Newton, Einstein, and Feynman, and the rest are a bunch of derivative rehashers because the built on what their predecessors did!
There is a lot of very novel stuff out there in financial papers and the edge from it persists for some time. For example, there was a much researched and published edge from time arbitrage of mutual funds based on overseas indexes. The opportunity from this persisted for at least a decade, the edge clearly published in plain sight for anyone willing to wade through the papers to see. It was published for years before I saw anything about it on internet message boards like this one, and even more years before it was finally shut down by the funds.
I would also submit that there are a large number of people in the world who are motivated by more than money. I know its hard for people on ET to believe, but its true. I know brilliant finance PhDs who regularly get lucrative offers from hedge funds but work in academia by choice. For them it is much more rewarding to publish a novel paper on a strategy that generates abnormal returns than to trade that strategy. Also, you have to remember that true abnormal returns, not the kind an ET punter claims to have gotten but those that have passed the rigor of publication, are on the order of a couple percent above market returns. Unless you're running a hedge fund, those aren't enough to get rich on and the PhD may very well get much more out of the prestige of publishing it than the $2,000 they'd get from trading their $100,000 portfolio with that strategy.
I would also push back on the concept that everything an academic does is derivative and that makes it worthless. That is kind of a fundamental misunderstanding of the scientific method which is basically to build on the accomplishments of those before you. By those standards Newton, Einstein, and Feynman, and the rest are a bunch of derivative rehashers because the built on what their predecessors did!
There is a lot of very novel stuff out there in financial papers and the edge from it persists for some time. For example, there was a much researched and published edge from time arbitrage of mutual funds based on overseas indexes. The opportunity from this persisted for at least a decade, the edge clearly published in plain sight for anyone willing to wade through the papers to see. It was published for years before I saw anything about it on internet message boards like this one, and even more years before it was finally shut down by the funds.
I'm not really looking to produce a derivative of other research paper rewriting what is already known, I'm more interested in exploring the strategy and applying different methods to find the one that returns a better performance and perhaps suggesting some of my own. Thus I wanted to know if any of the methods that are usually described in research papers are actually the ones that are used by the market participants since if they are not then it's worth to take a different approach and maybe try to talk to some industry professionals. I think that academia and the industry should be close to each other, you would usually get the core of the idea from public sources and then add some "secret sauce" of your own, but perhaps there are some professionals here who can shed some light on this topic.