People have been saying that because it's such an outrageous dichotomy, but it's also not really what happened in those cases.

To begin with Epic picked a disadvantageous test case because mobile is only ~6% of Fortnite with the large majority on PCs and consoles. So when Apple banned it on iOS, most of the iOS users just bought their Fortnite stuff on their PCs and consoles instead and Apple could say "see? not a monopoly" which got them a market definition that included Google Play. The market definition is about the single most important thing in antitrust cases.

But it wasn't really Google Play that people were switching to after Apple banned them and that could turn out a lot different for apps primarily used on mobile rather than trying to go after a mobile company over an app primarily used on consoles. That was the main reason Epic lost against Apple -- Epic had an app where people would actually switch to something other than iOS and Apple had enough evidence of that to convince the judge.

In principle that could have been the case for Google too, but they got a different judge, a jury trial instead of having the judge decide the facts, a correspondingly different market definition, and then it went the other way.

What confuses people is that Google partially got in trouble for things like forcing third party OEMs to install Google Play on the home screen of their Android devices, which is not a good look, whereas Apple isn't forcing third party OEMs to do anything because they don't have any third party OEMs. But the thing they got in trouble for wasn't having third party OEMs, it was strongarming them, which is obviously not the same thing.

And -- this is probably the most important part -- locking down the device isn't what gets you out of a market definition of "aftermarket for customers of that OS". It was more that Apple presented evidence that customers would actually switch to alternatives specifically in the case of Fortnite and their judge bought that but a jury in a different case didn't.

If anything the lesson for Google here should be to not strongarm third parties, because that's plausibly what pissed off the jury. And I'd be interested to see a case against Apple where the plaintiff is doing >50% of their business with iOS users instead of a single digit percentage, i.e. the ones where they actually have market power, though of course you then have the irony that those are the ones most afraid to bring the case.