When it comes to investing in factor-based products, a little sin is in, said two analysts on opposing sides of the smart-beta debate.

In a much anticipated face-to-face discussion at the 2016 Morningstar Investment Conference in Chicago, Research Affiliates CEO Rob Arnott and AQR Capital co-founder Cliff Asness continued their ongoing debate over the usefulness of smart-beta products, but agreed that investing in factors like value, size and momentum work well when combined with a little market timing.

“We refer to timing the market as a sin, but we also recommend that investors sin a little,” Asness said.

Arnott reiterated the crux of a recent white paper that warned that some smart-beta strategies could go “horribly wrong” if investors buy into factors with high valuations.

“Check the price tag of what you’re buying,” Arnott said. “If you are buying a stock, ask what you’re paying for it … when you look at an asset class, you ask whether the asset class is relatively cheap compared to historic norms. When you look at smart beta, you should be asking the same question, ‘Is this strategy cheap compared to historic norms?’”

Asness agreed that some market timing when valuations were extraordinarily high or low was warranted, but said that investors would be more likely to benefit from diversifying across multiple factors.

“It’s important to distinguish cyclicality from randomness,” Asness said. “Sometimes a factor goes into favor, sometimes it goes out of favor, so investors have to be patient, too. Randomness leads to periods of underperformance because it just happens … its important to view these things in a portfolio context. I believe more in diversification than I do timing.”

Arnott pointed out that many of the most popular factors were discovered due to timing. For example, when the value effect was discovered in the late 1970s, it was following the collapse of the so-called “Nifty Fifty” large-capitalization growth stocks and a five-year period of relative outperformance by lower-value companies.