There's another interesting anti-cricket argument in this type of debate which was first made by Goughy a long time ago where he argued iirc that a cricketer (Botham in that case) should be rated more highly for performing well against teams which were roughly on their level and they had a realistic chance of beating (like India and Australia) as compared to performances by a player against a far superior side which their side did not have a chance at beating (like Kapil's great performances v WI). I think it's harsh to not rate lower for performing their best against the best sides but it was a pretty interesting argument nevertheless.
I made this argument about a decade ago, albeit not fully believing it. IIRC it wasn't one of my more popular ideas. Most people hated it.
As an Ireland fan I would 100% take a home specialist over a batsman with evenly distributed runs. Away runs aren't very useful for us. We'll struggle to get results in the subcontinent for the foreseeable future, and one player who can handle those conditions won't change that. If we want a test victory, we need players who are good at home.
That's not exactly the same thing as home players being "better". But there's an extent to which batsmen choose which skills to develop, and I'd like it if they made that choice with the intention of winning matches. Suppose one of our players works hard on mastering Irish conditions, and another tries to master all conditions evenly, and they end of with similar records overall. I would strongly object to the idea that the latter player is automatically better because his runs were more evenly distributed.
All that said, I agree with your post on the whole. It's impressive for an Indian or Sri Lankan player to bat well in England or Australia, and vice-versa. It's not that the logic is wrong, it's just given more weighting than I think it should have, and Sehwag/Jayawardene types are underrated as a result.
I think the Goughy logic is stronger when it comes to valuing fifties, centuries, and double centuries. If you modelled the marginal effect of each run on win probability you'd probably get something n-shaped. Sometimes people use double centuries as a kind of tie-breaker, but that just seems wrong to me. The runs between 180 and 200 aren't very likely to make the difference between winning and losing.