I think poor design and/or poor play. If I'm doing nothing but making uninformed decisions in play, I'm likely to be dissatisfied. I expect map and key play to allow for ways to learn things so that I can make informed decisions. For me to be able to make moves that allow me to gain greater understanding of the situation.
As you may have gathered, I'm more sympathetic to
@Hussar on this particular point.
Up until a few years ago, I would have told you that I had very little interest in map-and-key play. But I did run a handful of quasi-dungeon crawls in 4e, mostly when the PCs were in Heroic tier, and I did enjoy them, but they were not really about the sort of informed play you mention in your post, and nothing like the sort of dungeon adventuring Gygax advocates for in his PHB. I tried to come up with an interesting series of rooms that would provide fun encounters in a 4e vein. The closest I remember to the sort of informed decision-making you refer to was this: the players entered one room from a downward-sloping corridor, and they conjectured that there might be some sort of slope-or-pit related trap. And so they roped themselves together. And indeed there was a pit in the room, and a Deathlock Wight who made the PCs recoil in horror; and when one of them duly fell down the pit as a result the roping together paid off.
But mostly the maps were just devices for creating spaces for the encounters to occur in, and for creating a type of geographic and resulting "story" logic. The last of these sorts of scenarios I recall running was at lower Paragon tier, which would be around 2011 or 2012. And
here's the last one I recall that really used the map to establish the logic of the encounters in any interesting way; in subsequent ones really were spaces for encounters to occur in, at least to the best of my recollection. (And when I later revisited D2 Shrine of the Kua Toa and G2 Glacial Rift of the Frost Giant Jarl, they were shorn of their dungeon-crawl elements altogether.)
Between then and 2020 I played a few sessions of AD&D (which is map-and-key down to its bones!) and continued to use some maps in 4e play as framing devices (as well as their fundamental role in combat resolution). But mostly I thought I had left map-and-key behind, as I had adopted other techniques (influenced especially by Burning Wheel) to manage the relationship between encounters, even inside buildings, without relying on a map-and-key.
My first real return to map-and-key as a serious thing was in Classic Traveller. Not at first - for instance, when the PCs assaulted an enemy installation (which was the first Traveller combat I'd GMed in about 30 years), I used the encounter distance rules to generate a sense of geography, and then sketched the installation map around that. But a bit later on I used maps (and adapted the keys) from the classic adventures Annic Nova and Shadows. I did these in a not-very-secret way: the maps were generally in the middle of the table, and the players could see where they were going. And because these are actually rather boring scenarios as written, revealing the information in the key didn't really involve much more than the players having their PCs wander through the places and look around at the obvious stuff.
It's only with Torchbearer, over the past 12 months, that I've been getting into map-and-key play that is remotely like what Gygax had in mind, with a hidden gameboard and the players trying to obtain information about it so as to make informed decisions.
This post ended up a bit more autobiographical than I had in mind starting out. But I think what I'm trying to say is that I can understand that there might be map-and-key play taking place, where the idea of informed player decision-making is not at the forefront.
EDIT: Upthread I talked about a type of play I'd mostly experienced in 2nd ed AD&D, but in some ways the 4e play I've described above resembles it. For me, the difference is that the 4e - influenced by player build decisions, player authored quests, etc - involves a different approach to prep from AD&D 2nd ed, and also the resolution framework is different enough that the way the map and key factors into
resolution, as opposed to framing, is different. But if those things are put to one side, it does look pretty similar to what I described and what I think (from post 763)
@Hussar has in mind.