It's clear why some might want to go back to a previous version, based on preferences, nostalgia (not a bad thing!), or really, a bunch of reasons.
Nostalgia is the 'n word' of the OSR community. The very mention of it drives them up the wall. The average member of the OSR community hearing the word thinks that you are saying that they have no real reasons for liking OSR games. The average person using the word merely means that they have a lot of fun playing games 'back in the day' and want to recapture that magic.
What's not clear to me is how people think games designed in the 70s are designed better than modern ones. I can understand preferring them, but, as [MENTION=4937]Celebrim[/MENTION] points out above, one would have to acknowledge the issues.
The very touchiness they have about criticism of the system is I think pretty telling. I can't imagine a player of 3.X whether 3.5 or pathfinder actually objecting to the idea that 3.X has serious issues out of the box. Indeed, I've never had a discussion with someone who is deeply familiar with a system where I'm also deeply familiar with the system about the systems problems that regularly goes in the direction talking about OSR does.
And again, for me the most telling thing is that if you do get one of these OSR people to defend the rules, like 8 out of the 10 things that they'll talk about aren't rules but play processes like rolling attributes, random magic item placement, more challenging encounters, proposition filters to use more fiction specific actions rather than 'moves', and so forth. All of those things can be and frequently are used as either rules variants or simply processes of play using more modern rule sets. Let's get real: 3e had random magic item tables and the default treasure placement methodology in its guidelines is random. Granted, it won't randomly generate a +5 vorpal sword in the hands of a low level goblin the way the 1e tables could, but then I seriously doubt that if the dice did generate that result in 1e AD&D (or OSRIC) that the results would stand, because 1e AD&D and it's emulators had the metarule "don't let the dice ruin the game" that applied to random treasure and random encounters.
And then invariably one of the things that they'll cite is that it doesn't have rules, which either undermines the claim or more telling means that they probably either don't know the rules or don't use them. Certainly, if you made the claim '1e AD&D doesn't have rules for X' you'd almost certainly be wrong. What you almost certainly meant was, "While AD&D does have rules for X, they are bad rules and therefore no one uses them."
Anyway, my whole point of all this grognardish grumbling is that what OSR is really about is not substance but style. And by that I don't mean that it is about something unimportant any more than by nostalgia I mean 'bad'. By nostalgia I mean good, as in most people aren't nostalgic about things that weren't good. Style is very important and is at least as important as rules, but it doesn't require an OSR ruleset to have an old school game.
Like anything else involving design, things get better as time passes, whether it's tech, social issues, education techniques, sports, whatever...and that makes sense, because designers today have access to everything that's come before.
Well, as long as you are willing to concede that this progress is not a smooth and even thing, but is filled with setbacks, mistakes, digressions, and other sorts of failed experiments, then I agree. The mistakes become part of the lessons learned by future designers, which is how we very slowly get better at things.
So, if someone says "As a wizard player, I loved 3.5" I can understand.
I can too, but if someone admitted to that it would likely imply very big things about what they really wanted in a system - most of them not healthy. For example, it would imply that they liked to have as an individual player an answer for every problem, and quite possibly that they liked to have spotlight in every challenge. Depending on the sort of wizards that they played and the sort of game that they ran, it might mean that the preferred to play a game without real challenge where they could just go from success to success. Of course, a really self-aware player might actually admit that, though that might not be the traditional metrics of good design, as a practical matter it is actually what they like.
By way of contrast, I never had a player play a successful Wizard in any of my 3.X games. Indeed, even when I was running RAW games Wednesday night in open dungeon crawls for all comers at the local game shop, the most successful returning player decided that the only way he was going to survive as a wizard was start a fighter, and then accept that a less powerful but less squishy wizard was the only thing that would work. This is so very much the opposite of CharOp decision making that I can only assume that if the CharOp people are playing anything other than a theoretical character building minigame, that they must be playing a very different game that I was running - even when I was running the same ruleset.
They might dislike 5e for addressing that imbalance...but they'd have to acknowledge that the game is better balanced. The design is better, but it no longer fits the preference.
But do they have to acknowledge that? To acknowledge that the 5e spellcaster is better balanced than the 3e full casters is to say that in some sense their victories in 3e weren't earned, but they simply the result of exploiting badly thought out rules. And that word 'exploit' could very well trigger them just as hard as 'nostalgia' triggers members of the OSR community.
You have to think about why people play RPGs. There are ton of different reasons for doing it, most of them boiling down to some variation on the 'illusion of success'. And for some people, the illusion of success requires them to not see through the illusion so that it feels like real success. So for our hypothetical player that loved 3.X for its full spellcasters and what you can do with them, to tell that player that that is bad design is to attack them emotionally - you are imperiling the illusion that makes the game fun for them. You are likely to end up in a strange argument with said player about how balance isn't important to an RPG (back to the John Wick school of gaming) and really RPGs are supposed to be unbalanced (or something of the sort) where what's really going on is a proxy argument for "stop attacking my illusion of success".
Now obviously, there is something very different going on here, as I've only met a few AD&D players who hard core exploit the spell rules as hard as a 3e CharOp player - though it certainly can be done if you exploit the illusion rules, for example. And the OSR style seems to involve a lot more dying all the time (at least in theory) than the sort of mostly on rails 1-20 AP campaign that become associated with 3e.
Again, I don't fully understand it and don't claim to understand it. My best theory is still the one I put forward - for most people their habits of play get attached to particular rules set and if they want to change their habits of play they have to change the rules. From the sort of complaints that I see, the sorts of styles of play and the sorts of habits of play that became stereotypical in say the 3.X era don't appeal to them, and to get away from those ingrained habits or to get their associates away from those ingrained habits they had to change the rules and the atmosphere. And if that worked for them, then more power to them. But for me, the cost of going back to those jankier rules is less than the cost of creating new ways of thinking about 3.X rules, or just tweaking the 3.X rules to nerf spellcasters a bit. And you might have notice from my discussion of 1e AD&D, I don't mind system mastery and optimization either as a player or a DM. To me, well created and conceived characters mechanically just means that there will be greater campaign continuity because the PCs will survive long enough to not just have a great scene but a story. I mean one thing that is definitely telling to me about these discussions is how often he OSR players all assume the best stories about D&D are tales of how some character died.