Too many things change during development of medium to large titles for a multitude of reasons to make this viable.
Not only that, but there's great risk in showing/talking about work that is still being iterated on, as people might (read: will) latch onto an idea of your product and then it'll be much harder to change their mind vs. having just shown it when it was ready. Show off too many different iterations of your idea and it seriously starts to muddle the message of what your game is.
There are basically no (or very few) advantages and many, many disadvantages to ever doing this. Maybe more viable with smaller projects that are more akin to a cinematic experience vs. a game with actual systems, gameplay, and scope that changes drastically from pre-production to launch.
Exactly. A movie pretty much has to be set in stone before it really begins production. A game is completely different as things aren't often locked down until it's a year or two from release, which is when games are typically announced.
I mean, imagine Splatoon. The initial proof of concept had no squids and would of had people scratching their heads. Then we'd hear it might become a Mario game and be even more confused.
I could see it working for series that have pretty regular development, like Assassin's Creed, Call of Duty, Battlefield, and a few others. But for the cast majority of games it just wouldn't work and it's an awful idea.
Plus it can really deflate excitement for a game. For movies we may get a cast list and general synopsis when it enters production. Maybe a few unintentional set photos, but we don't really get trailers or promotional material until a year or so before release, just like games. Mount and Blade Bannerlord is a game that has been in development forever and they try to do regular and open updates, but at this point my excitement has dropped down substantially, even though I am sure the end product will be good.