The problem with being good at sports betting is that most bookmakers learn who you are and refuse to take your bets. Why? Good sports bettors will cut into a bookmaker’s profit, especially when placing larger bets. However, this all changed in 1998 with the founding of Pinnacle Sports.
Pinnacle proclaimed that it was happy to take larger bets (up to a maximum) from anyone who played. And if Pinnacle found a player who consistently made money, they did not stop them. This was considered heresy in the sports betting market in the early 2000s, but Pinnacle had made an intriguing discovery. The Perfect Bet: How Science and Math Are Taking the Luck Out of Gambling reveals Pinnacle’s secret (emphasis mine):
Whereas all bookmakers look at overall betting activity, Pinnacle also puts a lot of effort into understanding who is placing those bets…Pinnacle generally posts an initial set of odds on Sunday night. It knows these numbers might not be perfect, so Pinnacle only takes a small amount of bets at first. It has found that the first bets almost always come from the talented small-stakes bettors: because the early odds are often incorrect, sharp gamblers pile in and exploit them.
But Pinnacle is happy to hand an advantage to these so-called hundred-dollar geniuses if it means ending up with better predictions about the games. In essence, Pinnacle pays smart gamblers for information.
Pinnacle was able to revolutionize the sports betting market because they paid a premium for data and insights on the outcome of sporting events. All they had to do was follow the money.
Pinnacle’s discovery reveals an important lesson for investors that I have recently come to accept: you are likely to outperform a traditional buy and hold investment approach…if you follow the money. When I say “follow the money”, I mean using a simple trend following rule. One such rule for investing in the S&P 500 could be:
- When the current price > average price over the last 12 months, stay invested (or buy in if you aren’t invested).
- When the current price < average price over the last 12 months, sell everything and move to cash (or bonds).
This rule translated into simpler terms would be:
- When everyone is buying and the market is moving up, stay invested.
- When the price starts dropping below where it was over the last year, get out.
That’s it. That simple rule has outperformed buy and hold for almost every 40 year period I tested starting in 1900–1939 and going through 1978–2017 (assuming no taxes or transaction costs). It has also done so with far less downside risk than buy and hold.
To visualize this, I put together a chart that shows both a buy and hold approach and the trend following rule (“12-month simple moving average (SMA)”) outlined above. I also shaded in green in the background every time the trend model “moved to cash” (i.e. this is inflation-adjusted cash, so it technically moves to TIPS, which I refer to as “cash”). As you can see, the trend following model outperforms over the 1900–2017 period:
[Author’s Note: I was informed after posting this that the prices I used from the Shiller data are average monthly prices, instead of closing prices. I have been told this can overstate the outperformance of trend following. My understanding is that trend following still works, but the size of the outperformance is not as large as the charts below display. If I have time to change the data in the future, I will update these charts and remove this note.]
If we zoom in on the period of 1990–2017 we can see more detail on how this model outperforms:
As you can see, during bull markets this model tends to underperform because it moves to cash too quickly. It acts a little skittish and doesn’t hold on. However, it makes up for this underperformance during drawdowns by moving to cash and waiting it out until the storm clears. You can clearly see this when it moves to cash in the DotCom bubble (2001) and the GFC (2008).
However, underperformance can last a long time. If we look at the 1978–2017 period, we see that this trend model didn’t start to outperform until 2003 or so:
My point with all of these charts is that this trend following model has worked well historically, but it has done so with periods of underperformance.
As a side note, I also tested trend following rules using shorter (3-month moving average) and longer (18-month moving average) time frames. However, I found that the only trend models that consistently outperformed buy and hold were in the 10–12 month range. Why? The 3-month moving average seems to be too skittish and moves to cash too often, while the 18-month moving average doesn’t react quickly enough to some drawdowns and can experience large losses. As a result, the 10–12 month moving averages seem to be the Goldilocks of trend following.
Why does this approach work? Simple answer: it follows where money is flowing. Trend following anticipates the mass of investor behavior to some degree and acts on that information. Just like Pinnacle Sports, trend following pays a premium of underperformance a lot of the time in order to gain information. Sometimes that information provides a huge edge when it comes to avoiding losses. By avoiding these losses, trend following can outperform.
Abandoning Buy and Hold?
Coming to accept that trend following can outperform buy and hold was difficult for me. I first heard about it through Michael Batnick, but was quite skeptical. And why shouldn’t I be? Buy and hold is the alter of the Gods in investing. Who would dare challenge it?
It wasn’t until I had dinner with Jake from EconomPic a month ago that I started to dig deeper. Jake sent me Meb Faber’s revolutionary white paper that popularized this idea using a 10-month simple moving average rule and then I decided to test it myself with the Shiller U.S. stock market data. I am convinced it works, but I wouldn’t recommend it for a large portion of someone’s tax-advantaged portfolio because (a) what if it stops working and (b) the bouts underperformance may make it difficult to stick with.
Instead, it might be better to use trend following for altering your equity exposure in your tax-advantaged portfolio. For example, instead of moving from 100% stocks to 100% bonds, maybe you move from an 80% stock/20% bond portfolio to a 60% stock/40% bond portfolio.
Either way, if you are interested in learning more on this, I highly recommend Meb’s white paper and Newfound Research’s detailed discussion on two centuries of momentum (note: trend following is a form of absolute momentum). Lastly, big thank you to Jake for inspiring me to write this. Follow him on Twitter if you don’t already.
Happy trend following and thank you for reading!
If you liked this post, consider signing up for my newsletter.
This is post 55. Any code I have related to this post can be found here with the same numbering: https://github.com/nmaggiulli/of-dollars-and-data