You're correct until your last 2 sentences. You assume Turtle still gives a nice edge when spread across enough commodities. It usually did in the 1980s, but not now. There was a time when the 200-day moving average actually worked pretty well for long-term stock investing. Trader Vic talked about it in his book. Now it's whipsaw city. You'll churn yourself to death if you buy/sell every time an index goes above/below the 200-day. Edges evaporate over time--especially simple trend following.
I think it depends on what kind of asset you're trading, as well as the specific strategy employed. My experience tells me that the moving average is still a useful tool to use in forex trading, for example.
Just want to make a quick point here - a lot of people are saying that trend-following systems are no longer outperforming due to algos/technology/sophisticated market participants, and while this is true, it really only tells half the story. Part of the reason that trend following worked so well back then, is because the 70s and 80s were massive inflationary environments (10-year treasuries yielding 14% in 1981!) Forget the stock indices and physical commodities, just the huge interest rate differentials between countries back then would have created massive trend opportunities in the currency and bond markets.....and algos wouldn't have been able to diminish those opportunities.