18
$\begingroup$

Over the years I met a couple of dozen of (human) traders who I'd consider good (being in the business for several years, generating a reasonably steady income). I had chances to chat with some of them regarding their trading strategies. All of them admitted they are not doing anything "out of the ordinary", whether it's some sort of pairs trading, trend following, curve fitting, and such.

Since these techniques are well known, I'm wondering why are they (still) able to generate profits? Is trading some sort of a CAPTCHA, one that machine-learning or other algorithms unable to crack?

Improve this question
$\endgroup$
4
  • $\begingroup$ A very warm welcome to Quant.SE and thank you for this very interesting and good question! Please see my answer below. $\endgroup$
    – vonjd
    Commented Feb 21, 2016 at 11:08
  • $\begingroup$ There's nothing special about trading versus other competitive areas like sports gambling. There the techniques are not secret either but some are still far more successful than others. $\endgroup$
    – Chan-Ho Suh
    Commented Feb 21, 2016 at 19:24
  • $\begingroup$ If any of the answers were helpful it is considered good practice here to accept it - Thank you. $\endgroup$
    – vonjd
    Commented Mar 2, 2016 at 11:57
  • 1
    $\begingroup$ I recommend the book Way of the Turtle by Curtis Faith. It explains in a lot of detail the very famous Turtle experiment that Richard Dennis and William Eckhardt made in the 80's. The turtle system doesn't work anymore, but it is an extremely interesting read. A group of diverse traders were all given the same (winning) trading rules, but in the end, only a few made money because most were simply unable to follow the simple rules. So to answer your question, the people you met can win consistently because of their personalities and particularly how well they handle the fear of losing money. $\endgroup$
    – Valtinho
    Commented Mar 16, 2016 at 15:55

4 Answers 4

Reset to default
9
$\begingroup$

First we have to clarify what we mean by profits: I think your question can only address the fact that some human traders beat the market (because you also make profit by just buying the market, e.g. through an ETF).

I think there are two, perhaps even three main sources:

  1. Randomness, luck (as @PerAlexandersson) correctly pointed out - financial markets are highly stochastic environments!
  2. From a systemic point of view the diversity of trading strategies: The more people follow a certain strategy - no matter how clever - the less profitable it will be (and vice versa)
  3. Level of fundamental research (all the way up to insider knowledge) (I know that this is still a matter of intense debate!)

For 2. see e.g.

N.B.: In this point is also factor investing included (e.g. value investing) - which when you look at it is nothing else but some sort of trading strategy!

For 3. see e.g.

(Just for your information: If you know some German the last two papers are very well explained in Schredelseker: Den Finanzmarkt verstehen (2015))

Improve this answer
$\endgroup$
1
  • 5
    $\begingroup$ @Downvoter: It is good practice here to state your reason! Is there anything I could do to improve the answer? Thank you $\endgroup$
    – vonjd
    Commented Feb 21, 2016 at 20:01
4
$\begingroup$

This might not be a suitable answer, but I refer to the book "The Black Swan" by N.N. Taleb.

If you start with 1000 equity traders, that base their investments on dice throwing, most will make profit when the market goes up. When it goes down, surely, some traders will make a loss and leave the game, but even after a long time, people with no strategy at all will have made consistent profits just by sheer luck.

It might be that you only know the lucky traders, and that traders who failed usually do not mention this.

On another note: Humans can probably also assess how CEO's will change companies, and see higher-level patterns. Remember, computers cannot even deal well with sarcasm yet. Computers are good at analyzing data, but humans are still needed to select which data to analyze.

Improve this answer
$\endgroup$
0
$\begingroup$

When comparing computers to humans they are both very different but don't think computers are better than humans. Firstly, they are programmed by humans, and secondly the human brain is undoubtedly far more powerful, it just differs how it operates.

Some of the most intelligent traders I have met think through all possibilities and have the capacity to weigh up exogenous market influences and design optimum strategies to maximise profit from those scenarios. This takes deep knowledge across multiple disciplines (ML usually encompasses a single pipeline of algorithms on one dataset or problem) and experience to be consistently good. Often the most profitable periods are via uncharted events which is why a programmed computer wouldn't necessarily work, because it gives rise to previously unseen data. If you are someone can utilise ML alongside this heuristic analysis you have the potential to be very good. I am sure many of these people operate under the radar and are keen to remain secret whilst generating profits.

You may have heard of trading computers being switched off in flash crashes etc. since they can't really control or function properly in those environments. Anecdotally, I recognise my own strengths and weaknesses and certainly my main strength is recognising when I might lose serious money and prevent it before the fact. I was far better at not losing money than having big wins, but consistently grinding out profits works in the long run.

Improve this answer
$\endgroup$
0
$\begingroup$

Well, it's pretty unfair to think all humans make money randomly. After all - quant strategies are designed by humans like you and I, meaning they have identified an edge in the markets and found a way to automate it.

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.