The tactical asset allocation model that’s outlined in Meb Faber’s widely read paper–the most downloaded paper on SSRN.com, in fact—went to cash at the end of last month, Bloomberg reported earlier this week. The risk-off move has attracted a fair amount of attention, which isn’t surprising, given the fame and influence of the 2006 paper (and its revisions)– “A Quantitative Approach to Tactical Asset Allocation.”

The model’s highly regarded reputation can be traced to a combination of simplicity and success, at least from the vantage of the paper’s back-tested results, which span four-decades starting in the early 1970s. In many respects, the work is a tour de force and has inspired countless studies that consider (and reconsider) momentum-based strategies for managing risk and otherwise avoiding deep losses a la the 2008 crash. But for all the well-deserved accolades, there’s a caveat that tends to be overlooked: the model’s not flawless.

That’s hardly a radical statement. Indeed, no model is perfect, which is to say that every model is flawed in some way. At its core, there’s much to admire about Faber’s model, if only as a foundation for deeper analysis. There are also plenty of questions in terms of how (or if) to apply the model’s signals in real time. According to one pundit, the prudent thing to do is to water down the all-cash message. “I wouldn’t go 100% anything,” writes Brett Arends at MarketWatch.com.

That’s a prudent bit of advice for an obvious reason: nothing’s guaranteed in money management. If you have 100% confidence in a model’s ability to provide accurate real-time insight for navigating the future, extreme decisions could be rationalized. But ours is a world with more than a trivial amount of uncertainty, a challenge that no amount of backtesting can overcome. Accordingly, no model—even one as influential and acclaimed as Faber’s—should be put on a pedestal.