Fully automated futures trading

AvantGarde said:
Hi GAT,
Great post. Your posts to date have focused on your futures portfolio, yet your equity portfolio seems to be significantly contributing to your returns. Can you offer more insight into your equity portfolio? Does it have a value/momentum/etc bias? What sort of strategy is it and how does it work?
Thanks!
More...

It's a historical accident really. Rather than fund my account with a big heap of cash, I transferred in some shares I already owned (plus some cash). But I wanted the majority of my performance to come from my futures trading, so I set up hedges against the stocks. Getting the hedge right is difficult; partly because to make my life easier I do all my hedging in the eurostoxx, but my portfolios is also spread over several non euro countries including the UK (and yes if anything there is a value bias).

Also the correct hedge ratio is hard to get right. I was probably overhedged for much of last year which means as the market fell I was gaining too much; I've now reduced that hedge.

I'm not "trading" this thing exactly except adjusting the hedge ratio, and opportunistically selling things for tax purposes.

GAT
 
Thank you for all the info you're sharing!
Do you do also swing trading?
How would you translate simpler trading rules into forecasts?I mean smth. like if A>B & B>40
Thank you. Hopefuly see you in London next week.
 
sidzkosic said:
Thank you for all the info you're sharing!
Do you do also swing trading?
How would you translate simpler trading rules into forecasts?I mean smth. like if A>B & B>40
Thank you. Hopefuly see you in London next week.
More...

I don't really know what "swing trading" is. Nobody has ever given me a satisfactory consistent answer. So if you tell me what you mean by it I'll tell you :-)

You could translate if "B>40 then go long" into something like (assuming B is say between 0 and 80) forecast = (B - 30)

Then if B=40 forecast = 10
If B=50 or above (forecast cap) forecast=20
If B=30 forecast=0
If B=20 forecast =-10
If B=10 or lower (forecast cap) forecast=-20

GAT
 
globalarbtrader said:
It's probably easier to show you my code.

Code: 
def breakout(x, ws):
  max_x=pd.rolling_max(x, ws, min_periods=min(len(x),int(ws/2)))
  min_x=pd.rolling_min(x, ws, min_periods=min(len(x), int(ws/2)))
  sig=[breakout_one_row(idx, x, max_x, min_x) for idx in range(len(x.index))]
  sig=pd.TimeSeries(sig, index=x.index)
  sig=pd.ewma(sig, span=int(ws/4.0), min_periods=int(ws/8.0))
  return sig

def breakout_one_row(idx, x, max_x, min_x):
  r_px=x[idx]
  r_min=min_x[idx]
  r_max=max_x[idx]
  return 4*(r_px - mean([r_min, r_max]))/(r_max - r_min)
More...

Hi GAT, is there a reason why you use a value of ws/4 for the EWMA of the breakout? Why 4? I see you have also multiplied the denominator in min_periods function.

What sort of values do you look at for breakouts? 10, 20, 30, 40, 50 etc?
 
AvantGarde said:
Hi GAT, is there a reason why you use a value of ws/4 for the EWMA of the breakout? Why 4? I see you have also multiplied the denominator in min_periods function.

What sort of values do you look at for breakouts? 10, 20, 30, 40, 50 etc?
More...

Four is nothing special. I've never fitted or done any sensitivity analysis on it. It just feels right. Probably any sensible number would do. Too high, and you'll incur a lot of costs. Too low and you won't adjust fast enough to the new trend that the breakout has picked up.

Breakouts over ws= 20,40,80,160,320 days

GAT
 
globalarbtrader said:
Four is nothing special. I've never fitted or done any sensitivity analysis on it. It just feels right. Probably any sensible number would do. Too high, and you'll incur a lot of costs. Too low and you won't adjust fast enough to the new trend that the breakout has picked up.

Breakouts over ws= 20,40,80,160,320 days

GAT
More...

When I simulate the breakout formula, I see the way you have defined it means it is bound by +2/-2. If I multiply this by 10 it fits nicely with the rest of the scaling.

As the method is bound, it is normalised. Can I correctly infer from this that you do not calculate a forecast multiplier for breakouts (ie so the the absolute average value of the forecast is 10)?
 
AvantGarde said:
When I simulate the breakout formula, I see the way you have defined it means it is bound by +2/-2. If I multiply this by 10 it fits nicely with the rest of the scaling.

As the method is bound, it is normalised. Can I correctly infer from this that you do not calculate a forecast multiplier for breakouts (ie so the the absolute average value of the forecast is 10)?
More...

Yes, change the 4 to a 40 and you have something that ought to be nicely calibrated.

Actually I do calculate a forecast multiplier. The scaling will only be correct if the distribution of the forecast has the right standard deviation. This isn't guaranteed.

However you might not want to bother. This is pretty anal stuff, it won't make much difference.

GAT
 
Hi Rob,

Just been reading your blog posts & book. I have a couple of simple questions.
  • Can you elaborate on why you roll manually? What are the problems you encounter in automating this task?
  • Why do you favour commodity futures versus other instruments, and why are commodity futures the principal instruments traded by large systematic hedge funds?
Thanks!
 
isotope1 said:
Hi Rob,

Just been reading your blog posts & book. I have a couple of simple questions.
  • Can you elaborate on why you roll manually? What are the problems you encounter in automating this task?
  • Why do you favour commodity futures versus other instruments, and why are commodity futures the principal instruments traded by large systematic hedge funds?
Thanks!
More...

To be clear the rolling is automatic (my execution algo issues eithier a spread, or two individual leg orders), but the decision to roll is manual. This is because there are multiple things to consider as http://qoppac.blogspot.co.uk/2015/05/systems-building-futures-rolling.html discusses. It would be very complicated to code up all these considerations into an automated decision making process, which would save very little time (a few minutes a month).

Not sure what you mean by "commodity futures" since I also trade equity index futures, bond futures, STIR and fx. If you mean "why do you favour futures", the main reasons are:

- huge liquidity (more important for multi billion dollar CTA's than me)
- they trade on exchange. The advantages of this are too numerous to list and would require a separate post.
- they are very cheap to trade (slippage and commissions for a given level of risk)
- easy leverage: margin requirements are low
- cheap available long history of data
- path dependence; In my professional career I spent nearly all my time working on futures models (with some dabbling in equities, and fairly serious work on OTC interest rate products), so makes sense for me to it now

GAT
 
You must log in or register to reply here.
Back
Top