Fully automated futures trading

Discussion in 'Journals' started by globalarbtrader, Feb 11, 2015.

  1. Yes. This is actually a really interesting result. Basically the skew for a trading strategy will only become apparent at frequencies in line with it's holding period. So if you're intra day trading, and trend following, you'd expect to see positive skew at daily horizons. But with a holding period of weeks the positive skew doesn't kick in until you're looking at monthly results. At a daily frequency you end up with skew that is in line with the skew of the typical position you're holding.

    Now this is the interesting bit; why is the daily skew negative? This I haven't investigated, but I have some ideas.

    In principle imagine a long/short system on one asset. Suppose that asset is the S&P 500. Since stocks have gone up more often than not, we're going to be long on average. But stock indicies have negative skew; so on a daily frequency we'll have negative skew (though less than the index long only).

    The question then begs itself, once we have a system that is trading multiple asset classes why this effect doesn't even out. For example you'd expect bonds to have positive skew, and we're on average long bonds.

    A couple of weird things here. Firstly many assets you'd expect to have positive skew (like US bonds) don't on daily returns; although they do at slower periods. Is this a weird microstructure effect?

    Also, and more intriguingly, is it possible that trend following systems tend to end up long assets when they are likely to have negative daily skew and vice versa? There has been some work showing that skew can be treated like a risk / return premium (I'm talking directional not option trading here); now if it turns out this is correlated with trend following then that is an interesting result (at least academically - not much use for trading).

    Last edited: Sep 7, 2016
    #501     Sep 7, 2016
    KCOJ likes this.
  2. KCOJ


    Bang on … longer-term trend following does indeed suffer from some big down days despite a positive skew when viewing monthly returns and as GAT says ... it's all about the holding period. Well spotted tradrjoe.

    Recently I carried out some work on this very topic and I’d say no … “profitable” trend following is most definitely not correlated with going long markets with a negative skew or vv. At least not when applied solely to commodities but I guess it’s always possible that the case may be different with financials … I might take a look at that.

    Meanwhile if you’re interested, you can see results of the skew strategy as applied to commodities here …
    #502     Sep 7, 2016
  3. I can't believe I missed that other thread. Really interesting and fantastic work. I have some comments I'll make over on there.

    To be clear is it your view that because going long skew doesn't work, and trend following does, that they aren't correlated; or have you actually checked the correlation specifically? (I'm guessing the latter as you seem like a careful guy). Did you do something like partition a price series into times when a trend following system would be long, and times when it would be short, and see if the conditional skew was significantly different between those two periods?

    #503     Sep 7, 2016
  4. KCOJ


    Actually sorry to say it’s simply the former … not so careful :sneaky:

    Based on the disparity in results between the two strategies it seemed largely academic to investigate any further.
    #504     Sep 7, 2016
  5. <slaps forehead in realisation> but they're not the same thing. One test looks at historic skew to decide it's position. The other test looks at trends to decide it's position, and then looks at future skew conditional on that. So unless skew is a perfect predictor of future skew you can't make the connection between your research on the other thread, and the hypothesis on this one.

    Might still be something in this.

    #505     Sep 9, 2016
  6. KCOJ


    Ahhhh of course. Now re-reading your original question …
    …you are absolutely correct, my research proves nothing of the kind … my bad.
    #506     Sep 9, 2016
  7. Don't worry about it, it took me a while to realise that as well

    #507     Sep 9, 2016
  8. Ouch. Down 5.5% yesterday. Hope everyone else is not too bad

    #508     Sep 10, 2016
  9. tradrjoe


    Which instruments/strategy did the losses come from? Did the short vol portion explode or were the trend and carry parts also carried out?

    Are you adjusting your sizing downwards due to the loss?

    Hope the markets pick up for you the rest of the month.
    #509     Sep 10, 2016
  10. BobbieJ


    Why do you think Winton has very low skewness compared to other CTAs? Too much carry?
    #510     Sep 10, 2016