You can use VFINX and FBNDX to calculate from 1987 to 2017.
That gives me about 8.5% dual-momentum CAGR versus 10.2% for the S&P 500 index.
Note that any algorithm can be "curve fitted" by it's author to prior performance data. The test of an algorithm is how it works after publication, especially if it exists as a mutual fund.
My biggest problem with dual momentum is how it neatly dodges every downturn by going 100% into bonds. Accurately predicting 100% of future crashes seems unlikely at best. And further, the fund exits and enters the market at about the same level, gaining only the interest generated by bonds. That's fairly close to just riding out the market correction in 100% stocks. Between so little gain, and so much risk if the algorithm guesses wrong, I don't view the dodging of market crashes as worthwhile.
Momentum itself is interesting enough that I've been running an experiment on it. I just don't think "dual momentum", which can switch into 100% bonds, is a reliable way to dodge future market crashes.