Dual Momentum Technology ETF

Hi Vladimir,

Thanks for sharing.  Clean and simple with good performance too!

I'm trying to learn how to code on this platform and your code is serving as a helpful example to understand and follow.

One question - where does the TimeSpan.FromDays() method come from?  I couldn't find it in the docs anywhere but an online search suggests that this is a C# method.  Can we use C# methods inside python code? 

Steve Jost,

It came from T Smith post.

Using several ETF and using exponential regression for momentum

Carsten,

I haven't reached the framework yet.
For comparison, I used the same symbols and dates as in my Dual Momentum Technology ETF.

Vladimir,

I did the tutorial and liked the framework, thats why I started directly with that....I'm around 2 weeks in this and I'm a lousy programmer. Should be quite easy form you. The biggest issue is the warmup stuff, totaly nerve ripping. I'm choose the framework, beacuse I belive its later more easy to run several strateys and control them depending on the actual regime. I might be easyer to achive a higher return from several average strategie than go for one high return strategie.

This program was a translation from zipline. With Quandel data and a slightly different set of ETF I got around 15 CARG and 10 DD, and yes, the set of ETF's is importand. With this one abouve I get at least CAGR and DD on paar, both 16%. could be imroved, less DD would be nice.

My interpretation from the Dual Momentum was 1) relative: get the Top X relative ETF and 2) absolut: only invest if the  slope > 0 for the ETF. Than i was looking for ETFs which have a more inverse corrolation. Unfortunatly there are some years when everything underperformed. One could experiment with more ETFs, worldwide. I tried as well to switch the stategy on of but got worse results. As the draw down is on par with its gains one could leverage up the strategie to 24% cagr and 24 DD with a leverage of 1.5.

@Vladimir, Do you have one for FOREX? 

Carsten,

Crefuly analizing you code I find out that your code with some exeptions is very similar to
Andreas Clenow Momentum.
So I changed definition of self.Value, trading logic and published it in this thread.

BukavuTrader,

Do you have one for FOREX?

Not yet, but you can try yourself.

'DataFrame' object has no attribute 'close' at __getattr__ return object.__getattribute__(self in generic.py: line 5574 at Momentum prices = self.History(symbol in /QuantConnect/backtesting/./cache/algorithm/project/main.py: line 30 at rebalance if self.Momentum(self.ASSETS[0] in /QuantConnect/backtesting/./cache/algorithm/project/main.py: line 36

Hi Papa Bear,

This error occurs in the first algorithm in this thread because the history request doesn't return any data yet the algorithm tries to access the close column of the data frame that returns.

The other two algorithms in this thread don't have this issue.

Best,
Derek Melchin