The main idea behind BCH is that they are going to allow as many transactions as possible to allow micro-transactions to take place for a low fee. They use ideas like Moores Law that says hard disk space can increase to cover all transactions forever (although there is no proof). The coin will become centralized over time possibly running on maybe 1000 servers or less.
The reason why it is a poor idea is it will become too costly to do spinoffs. Bitcoin has already had 2 code changing ideas, the lightning network and segwit. Imagine over time, leaders of BCH are going to come up with more ideas. But, unlike spinoffs, you are going to be locked into the BCH system because it will be too costly and complicated to move.
This may turn out to be true, but from an economic perspective, surely it's still better for them to try it out and then know for sure, rather than just making an assumption? If it transpires that on-chain scaling isn't viable due to centralisation, at least we'll then have the evidence to demonstrate it as fact. In this regard, I tend to view BCH as a "testnet-after-the-fact". More of a "what if" than an actual contender for market dominance.
I'm not entirely sure what you mean by "spinoffs", but it doesn't look as though they're having trouble making changes to the BCH codebase, as they're about to increase their maximum blocksize again, currently scheduled for 15th May 2018. The lack of SegWit and Lightning in BCH isn't due to cost or complication, it's more of an ideological difference. BCH supporters simply don't approve of those particular ideas. They forked away because they have a different vision of how this should work. If it does emerge their economic vision proves to be a mistake, it's their loss. Either way, it's all useful (and interesting) data.