Measuring your correlation to professional futures traders

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.
Chelonia
Roundtable Knight
Roundtable Knight
Posts: 525
Joined: Mon Apr 30, 2007 3:37 pm

Post by Chelonia »

With conditions, but not bad indeed, definately worth a look if you still believe governments are not going to default on their debt´s! I don´t see rates going up anytime soon because of the fact that the current problems come from the irreversable multification of debt by interest. This central banking monetary system has expired/terminated itself under insoluble debt. Fysical gold is probably your best bet during this transition period to hopefully a fair and sustainable monetary system. But that´s another discussion. Cheers
LeviF
Roundtable Knight
Roundtable Knight
Posts: 1436
Joined: Mon Dec 22, 2003 12:24 pm
Location: Des Moines, IA
Contact:

Post by LeviF »

Max DD or MAR is very misleading if a program with a lot of history has reduced leverage in more recent times. I guess thats what the Calmar ratio is for.
sluggo
Roundtable Knight
Roundtable Knight
Posts: 2987
Joined: Fri Jun 11, 2004 2:50 pm

Post by sluggo »

Very nice job, Klatt_attack. However I think using a single instrument may have been an oversimplification; practically all professional futures traders trade portfolios of dozens of instruments, and typically have on positions in many instruments simultaneously. Whereas your approach will have periods of time when it is (literally) flat: no position at all, and a flat equity curve. This is not what occurs in practice.

Fortunately there is a simple fix: just sample a portfolio-level returns series. There used to be a nice one stored on the Blox forum (HERE) but it seems to be gone now, probably lost when moving the Forum's web page hierarchy from one server to another.

I've excerpted Figure 8 of that file below. In brief, the first four moments of the observed returns series were:

Code: Select all

Daily Returns (Eq_today / Eq_yesterday)
----------------------------------------
Mean                1.00206
Standard Deviation  0.02123
Skew                0
Excess Kurtosis     2.63   as reported by Excel KURT()

So you could set your random number generator to have the same four moments and gen-up a returns series of your own. Or else just run a backtest of a system on a portfolio of 57 markets and capture the daily returns from the Daily Equity Log in the TradingBlox/Results folder :P Also, if you decide to repeat the work, plotting the data as a scatter plot (individual points) will eliminate the distracting diagonal lines.

Nicely done,
Sluggo
Attachments
Daily returns from a backtest which trades a 57 market portfolio
Daily returns from a backtest which trades a 57 market portfolio
histogram.png (96.68 KiB) Viewed 12074 times
td80
Roundtable Fellow
Roundtable Fellow
Posts: 60
Joined: Thu Mar 08, 2007 8:38 pm
Location: Irvine, CA

Post by td80 »

This has to be one of the hall of fame threads for this board. I'm sorry I can't add more value as I am still soaking it all in, but thank you all for your contributions and especially sluggo for lighting this candle.

Oh and by the way it is posts within this thread that make me glad I am not running institutional money (or worse trying to raise funding). There is a lot of insight into the madness of the institutional mindset in this thread.
nodoodahs
Roundtable Knight
Roundtable Knight
Posts: 218
Joined: Wed Aug 09, 2006 4:01 pm

Post by nodoodahs »

td80 wrote:There is a lot of insight into the madness of the institutional mindset in this thread.
:lol:
Chris67
Roundtable Knight
Roundtable Knight
Posts: 1052
Joined: Tue Dec 16, 2003 2:12 pm
Location: London

Post by Chris67 »

so much so that I was thinking about launching a class C shares - guaranteed no draw downs or volatility - it doesnt actually trade or do anything !!!! i think it would actually appeal to some institutions ! :D
Chelonia
Roundtable Knight
Roundtable Knight
Posts: 525
Joined: Mon Apr 30, 2007 3:37 pm

Post by Chelonia »

guaranteed no draw downs or volatility



Compared to what? :roll:
nodoodahs
Roundtable Knight
Roundtable Knight
Posts: 218
Joined: Wed Aug 09, 2006 4:01 pm

Post by nodoodahs »

td80 wrote:There is a lot of insight into the madness of the institutional mindset in this thread.
This:
http://www.cfapubs.org/doi/pdf/10.2469/faj.v65.n6.4

Also, google "tracking error" in regards to pension funding levels and "alm" and you'll get a snout-full of the institutional investor madness mindset.
CyTrend
Roundtable Fellow
Roundtable Fellow
Posts: 57
Joined: Mon Oct 25, 2010 5:28 pm

Post by CyTrend »

sluggo,
regarding the test_case_for_programmers.csv file you posted some time ago (thanks for that), i wrote some code to convert returns before and after fees and checked it against your file which works fine. However i am struggling a bit to understand why i can not replicate ,using that code, the results of a system run in TradingBlox with 2/20 and no-fees using the fee blox found in Blox Market Place. I am still going through it but in the meantime i was wondering if you have had similar issues and got any insights that might save me some time.
sluggo
Roundtable Knight
Roundtable Knight
Posts: 2987
Joined: Fri Jun 11, 2004 2:50 pm

Post by sluggo »

To save you some time let me suggest a layout of a spreadsheet, that you can create to check calculations such as these. Start with a nice round number for initial equity at the beginning of the first month (such as, 10 million dollars). Then just populate the returns data, fill in the formulas, copy them down to all rows, and away you go.

Column A: Date
Column B: Return_After_Fees
Column C: Return_Before_Fees
Column D: End-month Equity before deductions
Column E: Equity High Water Mark
Column F: (1/6)% monthly fee ("the 2")
Column G: Equity after mthly fee deducted
Column H: Net New Profits this month
Column I: 20% of NNP, incentive fee ("the 20")
Column J: Equity after incentive fee deducted
Column K: Calculated Return, after fees
Column L: Difference between ColK & ColB

If you want to get fancy, you can use the excel function =FLOOR(X, 0.01) . This rounds X down to the nearest penny. And the excel function =CEILING(Z, 0.01) rounds Z up to the nearest penny. But I doubt these are truly necessary.
halstian
Full Member
Full Member
Posts: 24
Joined: Wed Nov 10, 2010 4:20 am

Post by halstian »

I’m just going through some of the posts on this forum, and I came across this post. I know the post is old, but as an investment consultant in a former role, I have seen first hand from the other side what a difficult job raising money can be. I’ve listed a few observations below on the thread that I hope some of you might find useful and might make the difficult job of banging on doors looking for money that bit easier.

As students of human behaviour, I’m surprised to see that contributors to this thread made more reference to the motivations of institutional investors as “madâ€
Chris67
Roundtable Knight
Roundtable Knight
Posts: 1052
Joined: Tue Dec 16, 2003 2:12 pm
Location: London

Post by Chris67 »

Halstian

Appreciate your post - but it only really serves to confirm the irrationality and madness of institutional investors (teh majority of them)

What you r effectively saying is that institutional investors only have their own job security at heart and not the interests of their clients - and that as managed futures managers we should tell them a pile of crap in order to raise money ! This model - IMHO - is broken
You need to think a little mopre outside the box
Effectively either institutional investors change the way theyu do business or they all die - end of story
I dont change and I dont succumb to changing what I do to suit madness
Plenty of places I could sell crack or pimp out girls if I wanted to make a quick buck and devalue what I know is right
I think you'll find that by about early 2013 institutional investors will see it our way or they wont exist
From my experience of being out there raising money from people who partly manage their own money and partly have assets lodged through institutions they are all on the verge of cottoning on to what institutions who typically manage vast sums of money actually do - and they know its not in their interests to keep their money lodged in funds of funds / pension funds etc - I find that many more HNW's are actually a dam site more intelligent and will probably pull assets from these institutional managers in the next 18 months
If your outwardly stating that institutional managers have their own job security at heart over their client interests and that they are willing to give up returns to prevent themselves losing their jobs then its about time the SEC / FSA started to investigate this as I have no doubt you are right in your comments about that , currently, being the way it is
How much longer can teh U.S mutual fund industry defraud investors ?
halstian
Full Member
Full Member
Posts: 24
Joined: Wed Nov 10, 2010 4:20 am

Post by halstian »

Hi Chris

Appreciate your comments, I’d agree with a lot of what you say. I’m also very conscious that I was very lucky in the consultant job I had – I had a very understanding client base who allowed us a lot of leeway in what we could do, and weren’t obsessed with being second quartile in performance surveys every quarter. This is rare in the investment industry (due to the Tyranny of the Benchmark syndrome I referred to above). So personally I wouldn’t be critical of fund managers who toe the line – they are just trying to preserve their jobs in the face of a demanding client base whose tolerance for underperformance can often be measured in weeks. However, I’d totally agree that this is a system gone mad, and credit to you for your own persistence with your ideals and unwillingness to play this game.

I wouldn’t necessarily see institutional managers as being all bad – their funds do generally serve client needs, though not as well as if they had a free hand. I remember a conversation I had with a CTA we interviewed some years ago, good track record 8-10% per annum, low drawdowns. I’d done some research on the returns that may be possible and asked them to look at providing something with less emphasis on drawdowns. They came back with results of a fund that had higher drawdowns (max DD of about 25% from memory) and CAGR of 40%. They said they had tried to launch it some years prior to that and couldn’t get anyone to invest in it – the drawdowns had scared the investors away. So there is definitely a trade off between serving what clients think they need, and what is actually best for them. At the end of the day they are (generally) mature adults who make their own decisions, so giving them what they think they need is usually what is required in order to build and keep an investor base.

But on your views of the changes to the institutional investor / institutional manager way of operating in the near term – I’m not so sure. In my experience there are 2 types of institutional investor / manager. The first and by far the smaller group know the deficiencies of the system but while everyone else plays the game this way, they can do nothing about it. The second, far larger group are so concerned with trying to get ahead in the mad system i.e. trying to beat benchmarks, pick the best managers (consultants’ goal), or select the safest investments (pension trustees). They cant stand back and see what they are actually doing is not all that productive outside of the strange microcosm that they now inhabit. Though I think we could say that for a lot of aspects of our current economic and social model!

There is a cohort of people who have stood back and assessed the larger picture, some of whom you are evidently talking to with some success at the moment. In fact, a lot of this cohort would frequently write on this forum. The conversation above with the CTA showed me that it is in fact possible to achieve CAGR of 40% per annum with a respectable drawdown, and so I set out on the long and winding road to putting together my own managed futures fund to manage my own money. My own view has always been that if most people are locked in a mad system where they seek marginal outperformance of benchmarks but mainly get marginal underperformance, there must be plenty of opportunities for those who are willing to take a more unconventional route.

Good luck in your education process of investors – its a difficult task but those who understand what you are doing and buy in are most likely to be long-term investors for you, so it could work out very well for you. But having to do the education is a difficult process, and may not be for everyone trying to build an investor base.
SimJimons
Senior Member
Senior Member
Posts: 31
Joined: Mon Jun 20, 2011 9:59 am

Post by SimJimons »

Obviously low correlation is good, in general. However, I've seen a lot of examples where people try to sell their system/fund on that premise, but once you disect what they bring it becomes evident that they are snake-oil salesmen. Recently I was approached by a "CTA" claiming to have very low correlation to most other CTAs, but what they actually offered was a very HIGH correlation to the trendfollowing equity index leg of most CTAs. Stripping out asset classes will obviously reduce correlation overall, even if you are at 1 within equities. Hence, while I agree that low correlation is important I wouldn't bend backwards to try and achieve it versus other players. Having said that I do think that low correlation, if it is for real, may get institutional attention. I guess QIM is a good example of that.
Last edited by SimJimons on Thu Aug 18, 2011 5:15 am, edited 1 time in total.
Chris67
Roundtable Knight
Roundtable Knight
Posts: 1052
Joined: Tue Dec 16, 2003 2:12 pm
Location: London

Post by Chris67 »

Halstian

Thanks for your post - i personally thought you put this stuff together really well - and I totally agree that quite a few institutions have certainly got their act togther - but it seems the vast majority still dont

I think with regards to the stuff also about correlation its an impossible situation that can be best looked at in simplitic language
If someone said to me today would you like to be correlated to Eckhardt at 0.99 I would definately take it ? If you belive , as I do (rightly or wrongly) that in teh long term there is only really one way to trade financial markets , then but definition all teh successful players have to be highly correlated ? a sort of backward logic
finding 5 CTA's with a 20 year track record of great returns that are un-correlated is not possible ? Finding 5 cta's who, over the last 3-5 years seem to all performa nadbe un-correlated I would suggest is an accident waiting to happen and exposing one to more risks than they think ?
Listen nothing is that black and white / cut and dry but at the end of teh day our job is to make money - then whatever will be will be

Incidentally we are launching a new share class in 2 weeks time where we are effectively putting our own money where our mouth is - we are looking at 5-10 year CAGR of 50-100% with monthly (even daily) volatility probably higher than what most institutions would accept annually - we are perfectly willing to see 50/58% draw downs (as most L/T TF's have done)
I wouldnt market this to institutions and teh intention wasnt really to market it anyone else but, surprisingly enough, there seems to be a little bit of demand out there - its a balls to the wall strategy and I think theres a market for it
jas-105
Roundtable Knight
Roundtable Knight
Posts: 130
Joined: Sat Aug 02, 2008 2:32 am
Location: London, England.

Post by jas-105 »

Does anyone know what the threshold of correlation would be between two (or more) TF funds that would make them attractive in the institutional sense to be held together in a fund, assuming all other things equal ?
sluggo
Roundtable Knight
Roundtable Knight
Posts: 2987
Joined: Fri Jun 11, 2004 2:50 pm

Post by sluggo »

Here's how you could estimate it yourself.

Choose a portfolio of markets that duplicates the portfolio of one or more large CTAs (found on IASG.com). Run each of the eleven presupplied trading systems that ship with Trading Blox, on this portfolio. Save the equity curves of each (Blox stores it in the file Results\Daily Equity Log.csv).

Choose the best performing eight systems / discard the three worst, according to some measure-of-goodness such as Robust Sharpe Ratio or R-Cubed or something.

Pull the equity curves into a spreadsheet. The spreadsheet will have nine columns: Date, ECurve#1, Ecurve#2, ..., Ecurve#8. Then use the Excel correlation function =CORREL() to calculate the 8x8 correlation matrix among these eight systems.

Find the lowest correlation in the entire matrix. This represents the correlation between the two lowest-correlated systems of the bunch. Call it "No Brainer, Public Domain, Freebie Trading System Correlation" (NBPDFTSC). Asset allocators can get correlation values of NBPDFTSC or lower, very easily: they don't need YOU for that.

So to catch their eye, you need correlation lower than NBPDFTSC. A lot lower. (Your correlation to the systematic trader's index and your correlation to Winton, Blue Trend, etc, needs to be a lot lower than NBPDFTSC).

How much lower? Your correlation needs to be below (0.25 * NBPDFTSC) and it also needs to be below (NBPDFTSC - 0.50). Less than a quarter of no-brainer correlation, AND less than (no brainer correlation minus zero point five).

Run the simulations, assemble the spreadsheet, and you'll have your answer. If you post your research and your results here, I feel confident they will quickly be copied into the blogosphere, which might make you famous!
jas-105
Roundtable Knight
Roundtable Knight
Posts: 130
Joined: Sat Aug 02, 2008 2:32 am
Location: London, England.

Post by jas-105 »

Thanks Sluggo.
I already manage a fund and I know how it correlates to others (I should have told you this first, I know) but looking at it from the point of view of those folks who allocate to CTAs is that number more attractive if it's below a certain level ? I've heard 0.75 mentioned but that doesn't seem very uncorrelated to me, would that get you excited ?
sluggo
Roundtable Knight
Roundtable Knight
Posts: 2987
Joined: Fri Jun 11, 2004 2:50 pm

Post by sluggo »

Galen Burghardt, a pretty sharp dude, gets excited at a very different correlation than +0.75. Have a look at Exhibit 10 of his STTI document found HERE
Chris67
Roundtable Knight
Roundtable Knight
Posts: 1052
Joined: Tue Dec 16, 2003 2:12 pm
Location: London

Post by Chris67 »

SURELY 2 TF's with correlation below 50% are not TF 's ?
Seems carazy to expect 2 firms who both follow trends in global futures markets to have a correlation of less than 0.5 ? if tehy do something maybe wrong
Once again Institutional investors looking for what probably doesnt exist ? looking for the holy grail perhaps ?
Of course many TF's have lowly correlated performance over the short term but as we all know that counts for diddly

IncidentallyJAZ in the World of TF - Correlation of 0.75 would get me very excited ? 0.88 gets me going over the long run hehe
Best
C
Post Reply