As regular readers of Stockopedia will know, we are big believers in learning from data and statistical techniques what investing strategies have worked when. This kind of "quant" thinking is embedded in our stock reports - which feature risk indicators that have been found to be predictive like the Piotroski F-Score - and in our screening centre where we are tracking the performance of a wide range of different strategies - from Quality to Value to Growth to Income over time.
Leveraging an analysis of what's worked well in the past to develop investment and risk reduction strategies for the future makes sense to us. As George Santayana once wrote, those who cannot remember the past are condemned to repeat it.
The recent explosion of data availability and the falling cost of computing power also makes this kind of screening & modeling based on historically significant factors an important tool in the investing tool-box.
That said, it's also important to be careful with historical analysis too lest it lead you astray. It should serve as a useful device, but needs to be handled with caution too.
3 Pitfalls to be Wary Of
As wise regulators the world over like to remind us, past performance is not necessarily a guide to future performance! This counsel is of course true at many levels and there are lots of potential pitfalls in naively extrapolating from the past, including:
- Mean Reversion - Far too often investors suffer from the phenomenon of “performance chasing”. As soon as they see a hot asset class or sector, they pull their money out of their other investments and pour it into the new object of their affection. That's clearly a bad idea! Markets often experience mean-reversion (this means that, when asset prices deviate too much from their long-term trend, they will often come right back - for good or for ill).
- Data Mining - As we'll discuss in another article, another danger of focusing too much on the past is excessive "data mining" or confusing correlation with causation. If you look hard enough, you will likely find some seemingly effective rule that looks great, in the past. The issue is that the market has only one past but many possible futures. Even if stocks have rallied the first Friday in June for the past 30 years, it doesn’t mean they will rally again…
