It is an interesting question. We can only guess, but my guess I think I'm fairly confident on. Response online to the playtest Sorcerer was more positive than to several other less-changed playtest classes, and whilst it could be misleading, I really doubt it's that misleading. So I think we have our culprits.
The "special thanks" group of playtesters.
So not any of the people in the open playtest, but the somewhat strange group of mostly-grogs, including a lot of avowed OSR players who indeed kept playing OSR games after 5E came out. These were people essentially hand-picked by Mike Mearls (to the point where he engaged in some very bad decision-making about one of them later), and a lot of them were people who essentially didn't even want 5E. That doesn't make their advice worthless, but it colours the hell out of it.
There were a bunch of other strange changes after the final open playtest, which seemed to be either completely unconnected to or fly in the face of what people had been (openly) saying about the playtest. Not least the much-discussed change from 3-4 to 6-8 encounters/day (and easier ones) as the default. That was a terrible match for both 3E and 4E, so it's unlikely to be veterans of those. Even in 1E/2E, days with 6-8 encounters would strongly be the exception not the norm. Only in dungeon-crawling OD&D and OSR games is that more normal. But I'm still unconvinced it was those.
So I'm guessing the "special thanks" group influenced a whole flurry of last minute changes. I definitely don't think WotC had already shaped 5E and was rubber-stamping it, because 5E, whilst a cool game, is in many ways a mess, and really seemed very unfinished - the 5E DMG is the most unfinished-seeming core D&D book I've seen from any edition, and not by a small margin.