Man Investments Employee Sounds Off

Someone from Man Investments posted this today:

“As an employee of one of the largest CTAs I can catagorically [sic] state that we do change our models. This is, amongst other things, what us researchers are employed to do. I think that the assumption that all major CTAs run the same or similar models is also fairly amusing. Not all CTAs using EWMAs for example. At our shop we use four types of models to determine the level of momentum from markets, EWMAs being just one.”

Later in the day the same person posted:

“I agree that they are not changed every day but it is not as black and white as you appear to be making it.”

So you have a system that changes the system at Man?

Isn’t that a system?

13 thoughts on “Man Investments Employee Sounds Off

  1. I think that the system that changes the system are called the research team.

    This all sounds familiar to what I posted a few weeks ago…

  2. Isn’t trying to predict which of the models are going to work best on the next trading period a bad principle to incorporate in a system?

  3. it has been my experience that regardless of the system quants are using, whether the are trying to predict the momentum from markets, the power of earnings revisions etc, the guys that run the models are very often attempting to “optimize” their models to make the right fir for current conditions. That, in my opinion is where so much of the problem starts. Instead of saying this works and we are going to stick with it they adjust things, most often placing too much reliance on recent data and tilting too far in that direction right before the correction. The VAST majority of people that employ models do NOT run money the way the firms in Michaels’s books do. That is the point. The Kings he describes have found things that work and go with it. Many continually try to match models with what will work right now.

  4. @ Andre

    It’s in fact imperative that new systems be forward tested rather than exclusively back tested to counteract our tendency to optimize results by curve fitting. We HAVE to predict and test. This is why you should NEVER trust a system advertised through a pile of back testing (i.e. 99% of systems advertised…I could throw darts and get results that beat the S and P on back testing).

    It’s also important that a HUGE amount of forward testing be done as entropy ensures that all peaks and troughs in your data will smooth out and the biggest opportunities come from non-peak parameters regressing to the entropy curve. Only massive amounts of forward testing can identify those parameters.

    As an aside…Ed Seykota sings “The Whipsaw Song,” which states that if you’re going to trend follow, be prepared for the whipsaws. This isn’t just a folksy tune; it can be shown through mathematical proof that the more you optimize your money allocation parameters for maximum geometric growth, the larger the drawdowns.

    In other words, it is impossible in a systematic trading method of any kind to produce exceptional results without exceptional drawdowns.

    Any attempt to reduce drawdowns leads to an ARITHMETIC decrease in drawdown, but a GEOMETRIC decrease in profit. I have forward tested systems that anticipated 85% drawdowns…but, you guessed it, 85% returns as well. Good luck being able to stomach THAT system.

    It’s unfortunate that most fund managers are MBA’s and not mathematicians…if they knew a little more math, they’d embrace the drawdowns in good systems, and not treat them as pseudo-risk parameters to be avoided at any cost.

    By avoiding guys like Bill Dunn, for example, because of his drawdowns, they are seriously limiting their long-term performance for the illusion of stability. And when Jerry Parker limits his drawdowns to appeal to big money, he does serious damage to his results…unfortunate but unavoidable in an investment world that equates drawdowns to risk.

  5. Hi Michael

    What do YOU think of this, as the largest trend follower in the world changing their systems ? I understand tweaking, or small refinements, but changing constantly? Also they are not going to define what they change or the quantum of change, so difficult for us to see if that improves things or not? What do the smaller CTAs in your books do ? eg, Beach, Drury , Clarke etc..

  6. Atul, are there trend followers who use marketing hype? You betcha. Don’t pay attention to the words that make little sense, look at the performance.

  7. This is a ambiguous statement. It really does not say much of anything. Larry Hite in Michael’s documentary said it best . Most information like most opinion is suspect. Someone like David Druz with a couple people working for him putting up better numbers for longer periods of time. No army of Phd’s.

  8. Robert,

    Absolutely true…but what pension fund will invest with a guy working out of a small office and a single lap top?

    The trader is forced to throw a few Phd’s on staff, get a solid oak boardroom table, and refer to their “extensive research” into all manner of investment parameters to appeal to the morons who want more than good results.

    That’s why banks and investment houses spend money on large buildings with lots of glass and make all their emplyees wear suits…it gives the illusion of security.

    Shakespeare was right: “All the world’s a stage…”

  9. Michael

    The largest trend followers have more of an allocation to fx, in their portfolio due to size, and it is said that more of an allocation to commodities, gives better returns? The percentage of allocation of a 120 market portfolio, surely must count in performance numbers ?

  10. At any given time I’m sure the individual systems being used at MAN are mechanical – the question is whether the decision to tweak or discontinue systems is based on some level of subjectivity. Some smart traders I know allow systems that stop working to “extinguish” themselves on their own. As Michael mentions, if the approach to system selection is systematic you can end up with a “system to change your system”.

  11. The problem becomes, when do you pull the plug?

    On P. 37 of “Trend Following” (2009 updated edition) Michael has a chart of the British Pound labeled: “Unfavorable Market For [Bill] Dunn.” If you were a TFer who started out trading the Pound on that chart, you would’ve lost your shirt.

    Question is, then: Would you have continued to play Dunn’s TF system after a year of significant drawdowns in the Pound?

    In retrospect it’s easy for us to say yes because we know TF works….But just visualize yourself in that market for a year. Would ANY of us have the fortitude to carry on? Or would we throw out the system in disgust?

    Salem Abraham caught a good tending market when he first started trading…How would his story have turned out if he’d started off with the 1995 Pound in this chart? He had six months (if I remember right) to convince his grandfather he knew what he was doing as a trader. Sometimes timing can be everything.

    Just a thought.

Comments are closed.