Whatever point you are trying to make here, it would apply to any version of dnd.
<snip>
So the rules, instead of the gm, decide that you "can't just get help by praying" unless you are at least 10th level.
Like you, I assume that the answer in 5e D&D is
no, you can't successfully pray for divine intervention unless you're a cleric - because otherwise the cleric class ability would be pointless.
Conversely, in 4e D&D it might be the inclusion of Religion in a skill challenge. We saw that quite a bit in our game.
So I don't think my point applies to any version of D&D. That's why I expressly called out an OSR game. I also had in mind that Gygax's DMG has a rule that formalises the chance for divine intervention, and I can guess exactly where that rule came from! It's Gygax writing down some version of the procedure that he, or someone else he gamed with, came up with in order to resolve that sort of situation, which is highly forseeable in a fantasy game with active gods, especially among players who have read Elric stories.
Why is that better than the gm just making a reasonable decision and possibly allowing a roll, even at 1st level?
Well, as I posted upthread - I think in reply to you - I've got no view on whether it's better or worse. (As far as preferences go among versions of D&D, I prefer 4e.)
My point is what I posted:
I'm not saying that there's anything inherently wrong with GM decides as a resolution method. But I think there are obvious reasons why people adopt other methods too. It's not that "GM decides" has never occurred to them!
<snip discussions of various situations like divine intervention and charging through a line of soldiers>
These are the sorts of practical questions that have been at the forefront of D&D adjudication since 1974. Gygax made up answers, wrote some of them down, and published some of that in his AD&D rulebooks. His suite of answers perhaps wasn't the best - in particular the needless proliferation of baroque subsystems - but the reasons behind them are easy enough to see.
In other words, what explains the subsystem proliferation in Gygax's DMG is not just a fetish for rules! Is that the same situations come up, and he's worked out ideas for resolving them, and then he is sharing that with future GMs: it's like an extended, one-person OSR blog.
Consistently with my own preference for 4e among D&D versions, I think there are better ways to approach these issues than Gygax's. Of classic games my favourite is Classic Traveller, in part because I think it has a better suite of solutions to these issues.
If we are playing b/x, there is no specific rule related to divine intervention (at least that I can find in a quick search of my OSE books). As a gm, I might decide that a player has a 1% chance, and as gm, would still decide what that intervention looks like should it be successful. One can imagine a scenario in which a player says, "wait, that's not fair! It should be at least 5%." But if that's the hill your player is going to die on, then you do not have a high trust game.
I think trust here is an absolute red herring. Why is a player not entitled to die on that hill? - eg if they are playing a paladin who has been exemplary in honour and alignment and is on a holy mission that will fail unless the divinity intervenes here and now!
I mean, if you're playing B/X what hill do you think a player is more entitled to die on? That their MU PC is allowed to cast spells? That their fighter gets to roll a d8 for hit points? That the GM not have red dragons on the 1st level wandering monster charts?
I don't see why the divine intervention chance is a special case for any general principle of trust in the GM to make things up.
To me, this seems like the exact situation that calls for "ruling not rules." The rules cannot cover every edge case.
I don't see why divine intervention is any different from shooting an arrow at an Orc, or trying to get a good price on a bardiche from the local polearm shop.