Fully automated futures trading

[truetype]

AHL in FT today --

Several big names have struggled. AHL’s Alpha and Dimension funds have
gained only 1-2 per cent this year; BlueTrend, a fund run by Leda
Braga’s Systematica, has lost 6.4 per cent, while David Harding’s
$9.9bn Winton Futures Fund has trod water. But some have done well.
For example, AHL’s Evolution fund is up 9.6 per cent in the year to
the end of July, and Systematica’s Alternative Markets Fund has
returned 11.4 per cent. Crucially, however, these two funds take
advantage of momentum in less liquid, less efficient markets, not the
major ones that tend to be popular with trend-followers.

 

[traider]

AHL in FT today --

Several big names have struggled. AHL’s Alpha and Dimension funds have
gained only 1-2 per cent this year; BlueTrend, a fund run by Leda
Braga’s Systematica, has lost 6.4 per cent, while David Harding’s
$9.9bn Winton Futures Fund has trod water. But some have done well.
For example, AHL’s Evolution fund is up 9.6 per cent in the year to
the end of July, and Systematica’s Alternative Markets Fund has
returned 11.4 per cent. Crucially, however, these two funds take
advantage of momentum in less liquid, less efficient markets, not the
major ones that tend to be popular with trend-followers.

What are the less liquid markets?

 

[algo_fool]

I believe some of these less liquid markets to be German and Scandi OTC power contracts, quarters and cals

 

[truetype]

What could possibly go wrong?

 

[algo_fool]

Does anyone know where I can download Korean 3 year bond futures data? Quandl only goes back to Sept 2016

 

[HobbyTrading]

Does anyone know where I can download Korean 3 year bond futures data? Quandl only goes back to Sept 2016

Interactive Brokers has this data. However, they only keep data for expired contracts available until two years after expiry.

 

[globalarbtrader]

Does anyone know where I can download Korean 3 year bond futures data? Quandl only goes back to Sept 2016

Another few years from here:

https://github.com/robcarver17/pysystemtrade/blob/master/sysdata/legacycsv/KR3_carrydata.csv

GAT

 

[globalarbtrader]

What are the less liquid markets?

Big components of AHL Evo fund include:

Interest rate swaps (admittedly some of these are extremely liquid but they're OTC so not suitable for a traditional managed account)
OTC FX (also very liquid, although EM crosses less so) forwards and options
Single equity and ETFs (again quite liquid)
US agency debt
CDS indices
Power markets (electricity, coal, iron ore)

http://www.aifm-federation.at/index.php?id=631&no_cache=1&cid=3514&did=2354&sechash=cdf7863d
page 16

GAT

 

[globalarbtrader]

So looks to me like 2014 is going from 100 to 130. So thats 30% roughly, how do you get to 100% ? is it just scaled by volatility 2,5 to get annual volatility target of 25%, and them perhaps performance fee added?

Yes I multiply by around 2.5 but ignoring performance fees (which obviously handicap AHL in the comparison). I'm using the precise monthly figures from a spreadsheet hence a slight difference in the numbers you get reading off the graph.

My figures already account for my change in vol target (i.e. they assume I've always been running at 25)

GAT

 

[Elder]

Hi GAT, I am trying to stitch together an initial futures price database for backtesting the system in your book. My default roll policy was going to be to ‘roll to the 2nd nearby’ with the exception of seasonal contracts. Where this is not possible I was going to use the 1st. The rationale for staying at the front is based on my prior that the steepness and vol of futures curves for non-seasonal contracts tends to be greatest at the front (with the exception of STIR). After reading your blog on rolling, I understand you have a more complicated roll policy which is to roll to optimal carry contracts for live trading, but its not clear whether you apply this more complicated policy for building your core pre-live trading database and whether I would be missing something by not attempting to do the same? Partly I am daunted by the coding effort to get this flexible roll policy into a system and partly I am thinking this could be over-complication for perhaps marginal benefit. Any insights about your approach and views on mine would be very welcome.