A Different Spin on the “Perfect Alignment” Theory
This is a different spin on CSS Analytics’ Perfect Alignment theory.
As we saw in our previous test, criterion #1 was the more predictive of the two (252-day relative performance of the S&P 500 vs Russell 2000, AKA the Generals Lead the Troops). But criterion #2 has had an additional benefit we haven’t shown yet: predicting when the Nasdaq would outperform the broader market.
In this post, I’ll combine those two observations.
[logarithmically-scaled, growth of $10,000]
The graph above shows the S&P 500 in grey, versus CSS’s original criterion #1 (AKA the Generals Lead the Troops) in blue, and my own spin on the strategy in red, from 1989 to present.
The original (blue) strategy rules are: go long the S&P 500 at today’s close if the S&P 500 will have outperformed the Russell 2000 over the previous 252-days (1-year), otherwise move to cash.
The new modified (red) rules are the same, but trade the Nasdaq Composite instead of the S&P 500 when the Nasdaq will have outperformed the S&P 500 over the last 42-days (2-months). Put another way, I’m using CSS’s first criterion to determine when to trade, and the second criterion to determine what to trade.
Geek notes: these results are frictionless (i.e. do not account for transaction costs or slippage) but could’ve been easily reproduced using actively-traded mutual funds (our weapon of choice) net part of an annual expense ratio. As I do with all tests of longer-term strategies like these, I’ve included a return on cash of half the nearest interest rate composite.
Numbers for the number-lovers…
In terms of absolute and risk-adjusted return, the new strategy outperformed, but I’m always wary of just looking at averages – perhaps all of the improvement came during a single (curve-fit) period of the test.
So in the next graph I’ve shown a rolling 5-year Sharpe Ratio of the original (blue) vs modified (red) strategy from 1989.
Clearly the modified strategy has consistently outperformed the original (one of my main criteria for determining if something is as likely to continue working in the future as it has in the past).
Closing Thoughts
Readers know I’m an active swing trader chasing short-term moves in the market, but I think longer-term strategies like this one have a place in swaying the general bias of the trade.
And for determining that long-term bias, I like the idea of coupling a traditional trend-following strategy (like 50/200-day MA crossovers) with inter-market relationships like these because they approach the problem from a different angle and often generate signals that are completely contrary to a traditional trend-follower.
To see this multi-timeframe concept at work in a very simple way, check out the free daily State of the Market report.
Happy Trading,
ms
. . . . .
To stay up to date with what’s happening at the MarketSci Blog, we recommend subscribing to our RSS Feed or Email Feed.
Filed under: Trading Strategies | 14 Comments
I like this Michael, a great example of blogging collaboration!
One idea – given that the SPY and Nasdaq outperform during this period, presumably rotating amongst a small basket of the constituent issues (within the relevant index) with the highest rate of change would give even better performance. Just food for thought – i know your married to the rydex funds :)
cheers
Ramon
RE to Ramon: great idea and I agree on both points: I think looking at constituent stocks would bear a lot of fruit, but yes, I am married to the active mutual funds. [great opportunity for another blogger to step in here though] michael
Interestingly the original strategy is not nearly effective during the great bull run of
1988 – 1999 when buy the dips was the rule.
RE to Aristotle: two thoughts: (1) I’m not sure how this strategy would be related to “buying the dips”. That’s an short/intermediate-term concept. This is a long-term inter-market relationship.
But more importantly, (2) longer term strategies like this that are holding positions for so long (and not using leverage) are at best going to keep pace with strong bull markets (but not outperform) – we’ve talked about this quite a bit on the blog: long-term indicators like this one really shine during sideways and bear markets, but never strong bull markets. Thanks for the comment, michael
Re- “buying the dips”, I simply meant every time the market dipped in ’88 to 99, it roared back strong. So the double negative condition (perfect mis-alignment) showed strong returns. I was hoping it would pick up some of the minor dips but it’s probably too slow to react.
Thanks, really enjoy your blog!
Michael,
As an investment adviser, I love your testing and tweaking of these strategies! Thanks so much. Also, like Meb Faber’s system, they’re very actionable.
The issue with a strategy like this is even though the returns/drawdowns are EXACTLY what most retail investors ask for, they would also jump ship in the 05-06 period at EXACTLY the wrong time. It kills me when an investor says they are in it for the long run, but then they get impatient and bail.
Human nature will never change. I guess I shouldn’t complain, cause that means there will always be opportunity for me…
RE to Tommy: agreed, in this business you’re only as good as your last trade. I did a post on that subject a while back:
http://marketsci.wordpress.com/2009/08/05/random-thought-chasing-the-hot-hand/
michael
Michael,
Good stuff, in general, but I would advise caution here.
Looking at the equity curve, you can plainly see that almost all if not all of the out-performance of the updated strategy came from 2 years (1998-2000) when we know the NASDAQ100 was in a one in a hundred or more years bubble. In fact, the top 10 components of the NASDAQ100: MSFT, INTC, CSCO, WCOM (remember them?), and a few more accounted to well over 50% of the market cap of this extremely rare bubble.
RE to Ariel: I agree with you and that’s why I produced the 5-yr rolling the sharpe ratio of the two versions to show that over the entire life of the test, the modification improved risk-adjusted performance. Returns are an illusion of leverage, exposure, etc…R/A returns are not. michael
How does this fit in to the setup? Ron
http://stockcharts.com/h-sc/ui?s=SPY:QQQQ&p=M&yr=4&mn=0&dy=0&id=p15291165689
RE to Ron: not sure I understand your question. Can you elaborate? michael
MICHAEL, When brick is clear go with SPY and when brick is red go with qqqq. I use this ratio on 3xETFs to confirm trend, long or short. Sub $rut for your setup. Ron
http://stockcharts.com/h-sc/ui?s=BGU:BGZ&p=D&yr=0&mn=6&dy=0&id=p63905099793
http://stockcharts.com/h-sc/ui?s=BGU&p=D&yr=0&mn=6&dy=0&id=p95172011310
Is this strategy still active? It has shown a status of Cash for quite some time on The State Addendum.
Thanks
Hello Ernest – yes, still active…in cash for a long spell. michael