In a sense that if GM gets to decide certain things, then the player cannot be deciding them. But the same happens when you outsource the decision to the mechanics as well. The mechanics say what happens, instead of any of the participants.
I've seen this said multiple times and "+1'd" on top of that. But this needs a lot more unpacking, because if we're merely "outsourcing the decision to the mechanics" and there are no other dynamics happening within play procedures, then I would say 100 % there is either a dearth of gameable space there or a complete vacuum. It should never be something like:
* GM frames situation/obstacle without any foregrounding of consequences and in such a way that the only permissible action declarations and consequence-suites are univariate. I would go so far as to extend that to bivariate. It is certainly substantially better than the univariate dynamic of "the GM has basically made the decision for you (thereby played the game for you)," but multivariate is so much better than bivariate.
* Player makes rote, or relatively rote, decision and either (a) GM rolls behind the screen for the player or (b) the player has no handles/currency to muster or any consequential PC build or build/spend dynamic to leverage.
I mean...yeah. If the situation/obstacle-framing is utterly weak + consequences are black-boxed such that players can't orient themselves and execute a compelling & multivariate decision-tree...and maybe tack on the player having little to no build/currency decisions to employ? Yes, that is tantamount to "outsourcing the decision to the mechanics." In scheme and consequence, that looks basically like GM disclaims decision-making when handling offscreen Faction/setting (eg non-PC) resolution in Blades in the Dark (or something kindred). But those two things shouldn't look like each other at all unless a table or a game intentionally (or perhaps accidentally) composes deterministic, univariate, adynamic decision-spaces for players (which...isn't gameplay imo) or the play in question aims for or accidentally lands upon Calvinball/Ouija play (which...again, isn't gameplay imo).
Circling back to
my Torchbearer anecdote here, this is a relatively small (though of high magnitude) moment of play. This is the span of 1 singular turn in Torchbearer and 2 x tests in that singular turn. But the decision-space managed by both the Ranger Surveying and the Mage employing Aethereal Premonition was both vast in scope and consequential in impact. Though ultimately they both grabbed dice to resolve, what was baked into their decision-making scheme went way way way way way way beyond "outsourcing to the decision to the mechanics." And, what's more, all of this was downstream of the synthesis of a series of consequential decisions (around selected and employed Beliefs/Creed, around gaining the Enemy in question, around befriending the Friend in question, around actions taken in Town phase to Build Kit and Do Research, around preceding actions/resolution and elaborate currency management in this Adventure phase, around planning Camp phase which includes building up "checks" to power said phase).
I really hope we can agree on that (because man...what an absolute dead-end of the conversation if we can't agree on this); that "outsourcing the decision to the mechanics" resembles (a) GM Disclaiming Decisionmaking in something like Blades or (b) a truly feckless situation/obstacle-framing > action declaration loop of TTRPG (which, again, I would call "not rising to a gameable space")...and bears pretty much zero resemblance to something like (c) the attached Torchbearer actual play anecdote. This (a) and this (b) are not even in the same category as (c), right? Or are you actually saying they are effectively all the same?