Hi Rob, have you considered adding spreads into the system? My trend following portfolio has 30% made up of spreads, mostly calendar with a few inter-market ones (KE/ZW, ZN/ZB, GF/LE). The result is decent. Transaction costs double, and risk of breaking correlation is there. The reduced margin requirement really helps. Some spreads trend better than outrights. I would appreciate if you can comment on this. Thank you.
Fully automated futures trading
- Thread starter globalarbtrader
- Start date
Hi Rob, have you considered adding spreads into the system? My trend following portfolio has 30% made up of spreads, mostly calendar with a few inter-market ones (KE/ZW, ZN/ZB, GF/LE). The result is decent. Transaction costs double, and risk of breaking correlation is there. The reduced margin requirement really helps. Some spreads trend better than outrights. I would appreciate if you can comment on this. Thank you.
You say that transaction costs double, but don't you have to trade more contracts to get the same amount of vol with a spread as you do with an outright? If you have to trade 5x the number of contracts with a spread, isn't your transaction cost ten times greater?
I should have been more clear on transaction cost. By double I only mean for a single transaction (two-leg spread) comparing to outright. For total transaction cost per year, I think it depends. I do fixed position sizing instead of vol targeting, because I don't think the simple model of taking recent vol has enough predictive power for future vol.
For spreads, they partially remove the common mode noise, so the differential mode can trend better (stay longer in trend/fewer reversals). So they may result in higher SR in the end (together with the reduced margin requirements).
Like this one, CL, 3rd month outright vs. spread between 3rd and 6th month. I find the spread is better to trade in terms of SR.
I supoose spreads exhibit high negative skew, so even if it has higher SR, you probably don't want to leverage it up too much, unlike with positive skew strategies.
It works now. Thanks a lot, that's very interesting data. I will see how much my returns differ from this.
In that data in 2018 i am seeing a period where the things adds up to -48% (drawdown), is that correct or did i add up wrong?
Last edited:
I see drawdown trough at -40.50% on 2019 January 9th.
Max drawdown seems to be -41.52% on 1971 May 13th, but it's too far back.
This is what I get when compounding the daily returns.
P.S. I also get CAGR of 39.3% over the whole 52 year period.
That would be about right then, the real drawdown (40%) would be less the quick and dirty 'drawdown' figure I got by simply adding up the losing period days one by one (48%).