To state what should be obvious to most of you: go-arounds are pretty important, because nearly half of all accidents are landing accidents. Most of those accidents probably followed unstable approaches. So why don't we go around? I tend to think this is partly due to the difficulty of adapting stability criteria to airplanes with different handling characteristics.
If we're landing an SR22 at Palo Alto (2400'), the winds are calm, and we're on glide slope but 7 knots fast, we gotta go around. We really might not get stopped in time. If we're in a Cessna 172, 15 knots fast, 150 feet high, but at Half Moon Bay (4000') with a 15 knot head wind, go for it. On the slow side, the Cessna 172 will tolerate a crazy high sink rate so long as you land on the main gear. If you're a little too slow in a Cirrus, you will either run out of elevator authority to arrest your sink, or you will tail strike.
The go-around criteria are maybe just another example of a very well-intentioned effort to turn judgment and humility into a checklist to be followed. It's a nice idea if you can get people to comply, but the evidence says we haven't.
This is helpful in clearing up some of the misconceptions about Va, though I wish they wrote more about turbulence. Slowing to Va will reduce loads encountered during turbulence, but it can't guarantee protection. If you a flying in a thunderstorm at Va in a parcel of air moving down at 10,000 fpm, then cross into a updraft moving at 10,000 fpm, being below Va isn't going to help you. The lift on the wings is going to take them right off. (Technically, it's not lift. It's drag, but because the drag is caused by air moving up, it's easier to think of it as lift.)