Fully automated futures trading

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

  1. #4341     Jan 11, 2025
    newbunch and danw like this.
  2. VIX generally goes down, because of the futures roll down. So it's great for trend following but mostly you are short

    GAT
     
    #4342     Jan 11, 2025
    danw likes this.
  3. A more interesting question is - would you consider it a trend? In my opinion, forward vol (and generally, term structure roll up/down) is more of a mean reversion phenomena.
     
    #4343     Jan 11, 2025
    danw likes this.
  4. The term structure might mean revert (which you'd profit from with a calendar spread trade), but there is a systematic bias towards an upward sloping futures price curve which means going outright short will generally be profitable - basically the term premium for taking implied volatility risk is positive.

    GAT
     
    #4344     Jan 12, 2025
  5. danw

    danw

    Thanks guys, I just added both indexes and put them through the backtest. It seems trend DOES work for both (though only the VIX has long enough history to really look at).

    Not a huge amount of trades, but makes money for what is there - but only on the short side.
    Zero longs were profitable.

    Only issue is you need to use the front month for the best results, and that has a 60% higher margin at IB. Fine if size is small enough and not running other strategies, but might be a problem some time in the future if IB increases margin on other positions during some big event.
     
    #4345     Jan 12, 2025
  6. Yes this makes sense as the front has more roll down but also nastier skew which the roll down pays for and requires higher margin. Personally I trade the 2nd contract.

    This is discussed at some length in AFTS

    Rob
     
    #4346     Jan 12, 2025
    danw likes this.
  7. newbunch

    newbunch

    Volatility is mean reverting, but that doesn't necessarily mean that vol futures are.

    I have VIX and V2X futures in my system just like all other futures. The Sharpe ratio for them is about the same as the average of all other instruments. The back months test better than the front months, but that could be due to noise--the difference is probably not statistically significant.

    That said, my system rarely trades the VIX or V2X futures because they have high (inverse) correlations to their underlying indexes (-0.81 to -0.71 for VIX, -0.66 to -0.62 for V2X) and because their costs are higher than those underlyings, so it usually just trades the indexes.

    However, my system did buy one Mini VIX contract during the last vol spike (on December 18) when the VIX was closed as 27.61. I paid 19.27 for the Jan25 contract. While VIX has fallen sharply since then to 19.53, my VIX contract is down only slightly to 19.01. The point being that VIX and VIX futures are two different things.
    upload_2025-1-12_12-43-27.png
     
    #4347     Jan 12, 2025
  8. That is true, but if you trade also carry and/or skew, vol futures usually have higher (absolute) scores than their underlying indices, at least in my experience.

    I'm stating the obvious, but one thing one needs to be careful about is position sizing on the short side. Given their skew profile, sizing based on historical volatility can be dangerous.
     
    #4348     Jan 12, 2025
    jtrader33, Spacious and newbunch like this.
  9. newbunch

    newbunch

    Oh yes, that can be an issue if you only use a short-term vol measure.
     
    #4349     Jan 12, 2025
    younggotti likes this.
  10. I use a mix of short term and long term, but still find it dangerous to use it to scale vol futures.

    Based on my calculations, in the 6 trading sessions from 2018-01-26 to 2018-02-05 front VIX future spike was 4.1x its historical blended volatility (and 8.3x its short term one). For SPX future the ratio was less than 1x. I know I am cherry picking, but such events are not uncommon for vol futures.
     
    #4350     Jan 12, 2025