CTA Performance Index

Discussions about the testing and simulation of mechanical trading systems using historical data and other methods. Trading Blox Customers should post Trading Blox specific questions in the Customer Support forum.
illuminati
Full Member
Full Member
Posts: 24
Joined: Sun Apr 15, 2012 12:52 am

CTA Performance Index

Post by illuminati » Thu Apr 26, 2012 2:46 pm

Anyone know where to get CTA index performance data? I checked with Barclay hedge and it requires subscription fee to get my hands on monthly return series in excel format.

Other index include the NewEdge, but then they only have data from 2000 onwards...not enough.

fjpenney
Contributing Member
Contributing Member
Posts: 9
Joined: Sat Mar 19, 2011 1:45 pm
Location: Canada
Contact:

Post by fjpenney » Fri Apr 27, 2012 10:33 am

One other index is the IASG Trend Following Strategy Index ( http://www.iasg.com/managed-futures/cta ... wing-index ). Be forewarned, however, that published monthly performances continue to change for at least a year. For example, the change for March is now showing -2.35%. Check the site in a week and the value will be different.

I benchmark my performance against the Newedge CTA Index.

Chris67
Roundtable Knight
Roundtable Knight
Posts: 1046
Joined: Tue Dec 16, 2003 2:12 pm
Location: London

Post by Chris67 » Fri Apr 27, 2012 2:59 pm

if you want an index of trend followers as opposed to CTA's which as we know is a broad all encompassing meaningless term, then I publish a monthly Trend Followers Index- ive been publishing it for many years - its consistent and has dead funds in in to - then PM me and I'll add you to the distribution list
Best
C

Jez Liberty
Roundtable Knight
Roundtable Knight
Posts: 123
Joined: Tue Nov 03, 2009 8:49 am
Location: London
Contact:

Post by Jez Liberty » Fri Apr 27, 2012 6:46 pm

I do publish a monthly Trend Following Index on my blog too ("Trend Following Wizards Index") but it seems that you want historical values.

A while back I compiled something like this to make up a historical composite index (results can be found here)

Here is the spreadsheet that I used to compile the data. It is reasonably clear and easy to understand. You have monthly data going back a few decades for most funds (columns B:W), the calculated equity curve (columns X:AT) and the average results and resulting index (columns AV:AY - index starts in 1990).
Each fund is charted on the second tab (as per the screenshot).

It's obviously not perfect and might contain errors (data was collated from various free/public reporting websites).
It does have survivorship bias (as highlighted in the blog post) and only contains TF Wizards (no emerging, etc.).
Also I havent updated it since I compiled it in 2010.

But that might be a start and you have the raw data so you can construct another index from it...
I'd be interested to know if you find anything else/better.
Attachments
TF Wizards.png
TF Wizards.png (44.1 KiB) Viewed 6323 times
TF Wizards history.xls
(749 KiB) Downloaded 234 times

illuminati
Full Member
Full Member
Posts: 24
Joined: Sun Apr 15, 2012 12:52 am

Post by illuminati » Fri Apr 27, 2012 10:17 pm

Thanks so much for the excel spread sheet. I've found that its important to compare system performance to benchmarks so I can get a better grip on how my ideas hold up.

For trend following in general, I've come to two possible benchmarks. Ether compare to traditional basket of CTA's, but that is hard as it requires a list of historical constituents. Another way is to simply construct a basket of core systems and construct a index based on the aggregate performance.

The first option seems to be hard or inaccessible due to costs. The second one, crude in nature, doesn't really serve as a adequate benchmark as it doesn't incorporate a lot of the finer details of system design but it may be the best we've got....?

Below are a few known indexes that track CTA performance, some have data availble for free like newedge but not all do. Barclay requires $150 subscription and the rest will require some digging on your part.

Code: Select all

Barclay CTA Index
NewEdge CTA Index
Morningstar Long/Short Commodity Index
IASG CTA Index
IASG Trend Following Index
Dow Jones Credit Suisse Managed Futures Index
Altegris 40 Index
Stark 300 CTA Index
CISDM CTA Equal Weighted Index

Source:
http://www.managedfuturestodaymag.com/t ... ta-indices

http://www.barclayhedge.com/research/cta-indices.html


Hope this will help out the people who are looking for the same thing as me.


il

Jez Liberty
Roundtable Knight
Roundtable Knight
Posts: 123
Joined: Tue Nov 03, 2009 8:49 am
Location: London
Contact:

Post by Jez Liberty » Fri Apr 27, 2012 11:39 pm

I like the transparent aspect of a simple "mechanical" index (your second option).

It would also allow you to quantify your "system design alpha" compared to a Trend Following index/benchmark, which provides cheap/easy alternative beta.
Taking the index as a benchmark, you could evaluate each system by calculating its information ratio (or some other similar metrics), then compare it to pre/post-fees performances from Trend Follower managers (a bit of a dual approach..)

illuminati
Full Member
Full Member
Posts: 24
Joined: Sun Apr 15, 2012 12:52 am

Post by illuminati » Sat Apr 28, 2012 2:03 am

Given what you have said, I must agree. A composite of CTA's may not be as suitable benchmark as it incorporates in it a pinch of alpha. I guess its better like you said to use beta generating benchmark..

I've read from your website about your state of trend following. Do you happen to have the composite index values in excel? I'd love the opportunity to see the calculations methods and possibly use it.

babelproofreader
Roundtable Knight
Roundtable Knight
Posts: 138
Joined: Wed Nov 10, 2004 4:36 pm

Post by babelproofreader » Sat Apr 28, 2012 1:55 pm

Coincidently I have recently been toying with the idea of creating a benchmark suite of typical, mechanical trend following systems. If I actually get around to doing this, I'll post code on my blog and link to it from this thread. The code will be almost certainly be a C++ .oct function for Octave and/or a C++ function for R using Rcpp.

babelproofreader
Roundtable Knight
Roundtable Knight
Posts: 138
Joined: Wed Nov 10, 2004 4:36 pm

Post by babelproofreader » Thu May 10, 2012 8:53 am

Further to my above post, I have started coding my benchmark suite but have a conceptual problem. At the moment the benchmark suite consists of 5 basic systems, 2 of which are stop and reverse types and the other 3 can sometimes be flat. The problem is, assuming 1 contract per signal per system only,

If I take an average of the 5 equity curves this average composite equity could only move a fractional amount of the underlying contract equity move i.e. if long 2 systems and flat 3 the average composite equity curve will move two fifths of the amount of a 1 contract move.

If I take a simple sum of the 5 equity curves, this implies de facto position sizing i.e. the above example would be long 2 contracts, but at another time this sum composite equity could be long or short different numbers of contracts, which would make it difficult to use this as a benchmark suite if the system being compared to the benchmark is only trading 1 contract.

My preferred "solution" to this is to use the combined systems in an ensemble to produce a net long/short position of 1 contract, or be flat, so that the above example would be long 1 contract. What are the forum's views on this?

mojojojo
Roundtable Knight
Roundtable Knight
Posts: 131
Joined: Mon Oct 31, 2005 2:07 pm

Post by mojojojo » Thu May 10, 2012 11:23 am

Stock indices track market cap of the underlying securities. Using a similar view to equate a manager to a company, I would think you would want your index to trace their (your systems) equity curves/AUM.

I'm pretty sure Hedge Fund indices due the same, except they are either equal weighted between managers or are weighted based on AUM.

illuminati
Full Member
Full Member
Posts: 24
Joined: Sun Apr 15, 2012 12:52 am

Post by illuminati » Sat May 12, 2012 12:48 pm

Babel,

I am not familiar with regards to the progamming language you use but I assume that this can be overcome by setting up the code to accept parameters from the user. Let the user decide on which method to use, as I believe this flexibility will make your benchmark index very convenient.

It may be nice of you use multiple contracts too, as a lot of people, (like me!) test systems on multiple contracts.

AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2040
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner » Sun May 13, 2012 3:40 pm

It may be that the IASG Indices are a little misleading. There is no detailed explanation of the methodology used in their construction - or at least none that I have found so far. I have recently been looking at the performance of "Managed Futures" in general and took a look at the BTop 50 and Altegris 40 indices. Obviously we are talking the whole gamut of managed futures here and not just trend following but it is interesting to see how disappointing performance has been over the past 5 and 12 years, at least as far as these two indices are concerned. A far cry from the bullish figures to be assumed from IASG.

I am only at the beginning of the research project I have set myself but it is interesting to see from a number of different indices, as well as from the returns from a number of CTAs which I follow, that investment returns have deteriorated over the past decade. In part this may be as a result of de-leveraging (you need to look at risk adjusted returns) but in part it may perhaps reflect the realities of ever more crowded and volatile markets as competitors gobble up each others edge. Further research may lead to more definitive conclusions...or not, as the case may be.
Attachments
Indices.JPG
Indices.JPG (22.56 KiB) Viewed 5870 times

illuminati
Full Member
Full Member
Posts: 24
Joined: Sun Apr 15, 2012 12:52 am

Post by illuminati » Sun May 13, 2012 4:07 pm

AFJ Garner wrote:... CTAs which I follow, that investment returns have deteriorated over the past decade. In part this may be as a result of de-leveraging (you need to look at risk adjusted returns) but in part it may perhaps reflect the realities of ever more crowded and volatile markets as competitors gobble up each others edge. Further research may lead to more definitive conclusions...or not, as the case may be.
Yea, Tactical just experienced their MaxDD. I assume there are just too many players in the trend following space. It may be to the point of saturation, but then thats only my view. I am sure a lot of people will counter that statement with their own beliefs.

With regards to indices, I am still waiting for IASG to send me their indexes...it seems like forever.

mirec79
Full Member
Full Member
Posts: 16
Joined: Fri Sep 02, 2011 6:41 am

CTA and Fees

Post by mirec79 » Tue May 15, 2012 2:54 am

In last few weeks, i am watching a lot of traders searching for a benchmark or a performance of different trend followers of TF funds.

Some traders are comparing their performance with different TF funds.

What should be noticed here, is that the results are after fees! So it means, that the real performance and drawdown are much better than these numbers are showing.

Just for illustration i prepared two curves. One is showing the performance of DUNN Capital managing 100 000 USD of customers money from 2000 to 2010. This is the blue line.

Second curve is the same performance of this fund without commision! (or said another way, with reinvested commisions.) This is the red line.

Quite a difference, or?

(The data is direct from Dunn Capital Management, fee is 25% of profitable month, if a month is negative no fee is paid, there are no other expences).

To show my calculations i am attaching a excel file.

I presented my graph to few traders and their first reaction was alway that i have some kind of mistake in my computations. :-). I dont think so.

So my point is, that when someone is comparing his performance with "big funds" he should also consider the fees and expences.

A lot of investors are complaining about the TF performance in last decade, but the real/pure performance of such fund is still very good. See my example graph with Dunn Capital.

Mirec.
Attachments
Dunn Capital Management.jpg
Dunn Capital Management.jpg (88.66 KiB) Viewed 5787 times
Dunn Capital Management.xls
(55.5 KiB) Downloaded 217 times

AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2040
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Post by AFJ Garner » Tue May 15, 2012 3:45 am

You have taken no account of the High Watermark provisions. No performance fee is charged unless the fund is at new highs. You have added back in the 25% performance fee in cases where none would have been charged. I fear you must revise your spreadsheet to take account of this. You will find it makes a considerable difference.

AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2040
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Dunn

Post by AFJ Garner » Tue May 15, 2012 4:18 am

See below.
Attachments
Dunn.JPG
Dunn.JPG (27.26 KiB) Viewed 5757 times

mirec79
Full Member
Full Member
Posts: 16
Joined: Fri Sep 02, 2011 6:41 am

Dunn

Post by mirec79 » Tue May 15, 2012 4:57 am

I selected Dunn because the fee structure is simple, just 25% percent from profitable month, i was contacting the office if Dunn and they told me they are not charging any other fee.

So one example, there is a month with 10% profit. (after fee). So it is actualy 75% of the real profit. Dunn took already his 25%. So if Dunn is claming that the profit for this month is 10% in reality it means that the profit was 10/3*4=13.33%.

Or another way to calculate is, that if the real profit for month is 13.33%, so commission for Dunn is 13.33*0.25=3.33% and for the client it is 13.33*0.75=10%.

This is the calculation in my excel file.

Profitable month/3*4=real profit.

Mirec.

AFJ Garner
Roundtable Knight
Roundtable Knight
Posts: 2040
Joined: Fri Apr 25, 2003 3:33 pm
Location: London
Contact:

Re: Dunn

Post by AFJ Garner » Tue May 15, 2012 5:06 am

miroslav_krajcir wrote:I selected Dunn because the fee structure is simple, just 25% percent from profitable month, i was contacting the office if Dunn and they told me they are not charging any other fee.
Do they operate High Watermark provisions?

mirec79
Full Member
Full Member
Posts: 16
Joined: Fri Sep 02, 2011 6:41 am

Dunn

Post by mirec79 » Tue May 15, 2012 5:10 am

Ok, i ll drop them an email and i ll be back with the answer. AFJ you are right if they are not charging fees on new highs this matters a lot.

Mirec.

stopsareforwimps
Roundtable Knight
Roundtable Knight
Posts: 199
Joined: Sun Oct 10, 2010 1:47 am
Location: Melbourne Australia

Re: Dunn

Post by stopsareforwimps » Tue May 15, 2012 6:46 am

miroslav_krajcir wrote:Ok, i ll drop them an email and i ll be back with the answer. AFJ you are right if they are not charging fees on new highs this matters a lot.

Mirec.
Either way the 2+20 provisions make it hard for the punters to win.

Someone calculated that if Warren Buffet had charged the investors in BRK-A 2+20 they would have underperformed someone who indexed the S&P500.
Attachments
corzine-obama.png
corzine-obama.png (193.29 KiB) Viewed 5721 times

Post Reply