Corey
|
|
| Total Posts: 234 |
| Joined: Feb 2008 |
| |
|
Someone forwarded me along the following article, "Adaptive Asset Allocation: A True Revolution in Portfolio Management", written by the guys who run the Darwin Funds.
Now, my bullshit alarm is screaming on the backtests, but I can't put my exact finger on where I think the idea is flawed. Momentum is a pretty well established phenomenon and there has been a lot of recent research around min/low vol portfolios historically out-performing -- so I don't feel I can necessarily call bullshit there.
I feel pretty comfortable saying that the length of the backtests for the assets in the portfolio are unrealistic; not only are the assets being included unrealistic for many portfolios, historically, but many of them didn't have easily tradable and liquid vehicles until a few years ago.
Discounting that hindsight bias by using an equally unrealistic benchmark (their "equal weight"), they still get some pretty freaking phenomenal returns.
So why is my bullshit alarm internally screaming about historical over-optimization? I can't put my finger on it... |
"Then there was the man who drowned crossing a stream with an average depth of six inches."
W. I. E. Gates |
|
|
 |
|
In my opinion one should watch their back in anything to do with Macquarie.
They have Jeremy Grantham quotes in their marketing material and CFAs all over the show as if they were preaching a deep-value Benjamin Graham style. The actual product, however, appears to be technical-ish using trend following with vol & correl weighting. If the strategy is supposed to be the latter, one might expect computer and mathematics talent.
With the backtesting, one can never really tell. That is a handicap of all startups, and is certainly not unique to them. In the past, I have found it useful to try to couch the backtest results in the context of my discussions with the investment team - is the package consistent from all angles, does the strategy make sense at the simplest of levels, and how serious are the obstacles to good performance under different scenarios? You would need to talk to them to make a decent judgement here.
Most people judge backtests by the out-of-sample behavior. I have done this as well, but have been reconsidering where the data set is time-ordered and probably serially correlated. Out-of sample testing would be more valid for unordered data sets with zero serial correlation. But if all you have is a hammer, then everything looks like a nail... |
|
|
 |
NIP247
|
|
| Total Posts: 450 |
| Joined: Feb 2005 |
| |
|
| Just had a quick glance and it seems to be a strategy from the "voltarget" family. In my opinion, the underlying assumption of such strategy is that (1) vol clusters (which we can agree on) and (2) you experience higher volatility with negative returns. Assumption (2) has been historically kind of correct for equities (except end of dotcom bubble '99 in recent history) but I believe that's a very strong assumption- The Artemis article on volatility had a pretty good example of a macro regime that would flip that assumption. |
On your straddle, done on the puts, working the calls... |
|
|
|
Lapin
|
|
| Total Posts: 244 |
| Joined: Feb 2006 |
| |
|
NIP247,
Are you talking about the following article:
Volatility at World's End? |
|
|
|
NIP247
|
|
| Total Posts: 450 |
| Joined: Feb 2005 |
| |
|
@Lapin,
yes |
On your straddle, done on the puts, working the calls... |
|
|
|
Lapin
|
|
| Total Posts: 244 |
| Joined: Feb 2006 |
| |
| |
|
Corey
|
|
| Total Posts: 234 |
| Joined: Feb 2008 |
| |
|
| Thanks for the responses and the Artemis paper. Seems like assumption #2 that you mentioned, NIP247, could be a real downer for min-vol strats in the future. |
"Then there was the man who drowned crossing a stream with an average depth of six inches."
W. I. E. Gates |
|
|
|
Lapin
|
|
| Total Posts: 244 |
| Joined: Feb 2006 |
| |
|
Corey,
I may be wrong but this strategy seems to complete the TAA from Mebane Faber.
You have a momentun strategy and you add a layer of risk budgeting.
I had the feeling that TAA a la Faber made sense for a long term investor ... without real risk management constrains |
|
|
|
|
is that graph on the lower left of page 6 a real plot? I can read it, but the plots look fake.
|
in the words of one such quant ‘were on the whole either less quanted or
not quanted at all’. |
|
|
|
|
| This fund would make a good candidate for a 'dead pool' thread. |
|
|
|
nadtim
|
|
| Total Posts: 8 |
| Joined: Jun 2012 |
| |
|
Could it be that the system was developed by these folks: Automated Trading System
There is a testimonial from: Adam Butler @ Darwin Funds |
|
|
|
|
|
posted by: nadtim - "Could it be that the system was developed by these folks: Automated Trading System. There is a testimonial from: Adam Butler @ Darwin Funds "
I spoke to them a while ago and can confirm the AUM quoted in the testimonial is what they told me was currently being traded by the strategy. That said, in the testimonial it also looks like they've modified their approach/code from what was done (if anything) by ATS.
Given they focus on managing risk and diversification - imo that's a cut above most retail offerings. They seem like good people with a reasonable product, hope it works out for them. m2c. |
|
|
|
