I don't think the efficient market hypothesis is universally accepted among economists.

The scope for disagreement is on to what degree a particular market is efficient (and what form of efficiency there is). Many markets (particularly smaller ones) are far from efficient and strong-form efficiency (where it isn't even possible to make money based off insider information) more or less have to be contrived scenarios. The markets are anti-inductive post that is being discussed roughly speaking only claims that markets have pressures pushing in the direction of efficiency. John's claim is a little stronger, claiming that the particular market under discussion is likely to be more efficient than you are.

The heavily traded global markets that we have now are not perfectly efficient. It is clear however that they are sufficiently close to efficient that extracting money from them from nothing more than historical trends in the prices is not easy.

What's your support for the assertion about "countless traders" using practices that concentrate risk? It does seem intuitively plausible, but is there serious scholarly support for this idea?

I don't believe I am referring to something controversial. Someone else more familiar with the nomenclature may probably more suited than I for the task of bludgeoning the principles home with authoritative references than I. All I have is an example, the famous popularisation and the name of the risk that has not been hedged against or accounted for when making this particular mistake.