Not really wanting to get into yet another "they do work / they don't work" dabate, but I believe the statistics showing their effectiveness largely come from reported head injuries. Not everyone reports those injuries though. I've written off two helmets in my time and neither incidents resulted in a trip to A&E. The damage done to the helmets wasn't done to my bare head.
Saying a helmet offers no protection is like saying shoes offer no protection. But I wonder how many people would turn up to their doc with cuts and blisters if not for the minimal cushioning from shoes.
If helmets mitigated serious injuries in to minor ones to a significant degree then the serious injury rate should go down. It doesn't appear to.
Saying helmets offer no protection is indeed lazy. It should be qualified in to something like they appear to offer no consistently useful protection against serious injury. There isn't convincing evidence that wearing a helmet on a journey makes you more likely to finish it without a serious injury. (There's not really any evidence one way or the other that your chances of finishing the journey without a minor injury change either way. You can reasonably expect (from the materials used, not injury studies) that if
you hit a protected part of your head you'll have less of a headache and more complete skin cover.)
The shoes analogy would make a lot more sense if we banged our heads as often as we press our feet against hard/sharp/abrasive surfaces. Can't speak for you, but that's not really the case for me. As it happens I routinely go barefoot in benign foot-friendly places, and I regularly wear sandals to go mountaineering in summer, and my feet are not regularly mutilated despite many assumptions that they must be battle-scarred puddings because I don't clump about in protective boots everywhere. I find that the old "look where you're going" system reaps dividends.
I work next to a walking/footwear research place, and have the occasional natter with the folk there. I learned that in places with walking populations and a lot of bare feet there are actually more doctor-worthy injuries to people wearing shoes...
Back to helmets, the basic known effect against serious injury for a notional cyclist starting a notional journey (not "in an accident", you have to account for the possible increase in chance of being in an accident) is zero plus or minus error bars. Difficult to say how big those error bars are. The basic effect against minor injuries is unknown (as you point out, data is not reliably collected), but they're minor. If you bang your head on a bike significantly often that you'd collect minor injuries I would suggest you wear one. If you don't then it's something of a moot point.
Often seen riding a bike around Dundee...