There is an error in reasoning in this calculation:
- As more miners move to bitcoin cash, then the difficulty rate in Bitcoin cash also will increase dramatically and will make the mining in Bitcoin cash much less profitable.
- About the calculation that it will take 20 weeks, if the hash power reduce by 90%. That would be only true, if all 90% move in 10 minutes at once at the start of the 2016 blocks.
- Even you are right and 50% of miners will move in 3 weeks to Bitcoin cash. But then the Bitcoin difficulty will be reduced by 50%. And then suddenly it will be for 2016 blocks very profitable to mine Bitcoin again . And Bitcoin cash will have delayed confirmation time after. And back .. and forth
Point #1, yes, but the difficulty of Bitcoin Cash can only increase after 2016 blocks. So it is not immediate. The difficulty can downgrade in 12 hours, but the difficulty cannot upgrade in less than 2 weeks.
Point #3, Bitcoin SegWit difficulty cannot reduce until the end of 2016 blocks is reached. If substantial mining power leaves, the end of 2016 blocks is greatly delayed. The more it is delayed, the more possibility of miners leaving. It would only take the top four miners leaving in the first few days of a difficulty period to take away 50% of the hashpower, doubling the adjustment period. None of the top four miners presently have any ideological commitment or interest in SegWit.