There is another good piece they did that has the advanced metrics guys saying that Houston and GS should almost double the ammount of threes they attempt each game (also suggested that Milsap should be taking 3-4 threes a game for the Jazz). The Coaches and the metrics people are sort of at war, and Metrics keeps winning ground. I couldn't find it, I'll search on draft night or sometime.
It's because metrics are fundamentally rooted in hard facts while coaches generally regurgitate falsities that they were taught.
Why do you assume these are true? Everything I've read on modern coaching says they are very aware of the metrics.
Coaches are well aware of the statistics that "prove" this or that lineup is better, but they can't worship them as a holy grail. Gregg Popovich was ridiculed by one reporter after losing the series because the statistics showed playing Duncan and Splitter was a defensive nightmare. Spoelstra adjusted by going small, and after that win said we'd see "more of the same".
Making strong adjustments that might seem counter intuitive based on pure metrics is where good coaching lies.
Let me give you an analogy with Texas Hold 'em. I can be in a pot with a 33% chance of hitting the nut flush, get 5:1 on my money, adjust in my head for the possibilities of a full house or better negating my flush, and make a statistically sound call that's guaranteed to win me money in identical cash game situations in the long run.
However, I'm not a casino playing pure win:lose %'s. There are plenty other variables such as the size of the pot to my bankroll, depth into a tournament, etc. that are so situational that a purely statistical call is not guaranteed to win me money over time.
Also, to site 3 point shooting and follow it up with metrics gaining ground after you wrote that Griffin bit yesterday on how he wouldn't sniff a leading 3 point attempts title today is kinda funny.

I think the changing league outpaced the metrics on that one (and had something to do with allowing zone defense again).