Unfortunately, that's what the BIP101 schedule confers, and so you are advocating (indirectly) for outgrowing network resources. Only industrial scale miners can cope with above trend changes, hence the centralising effect.
One correction here is that this is true for the pools, not necessary the miners, since the miners are not effected by an increased block size, Miners and pool operators are today mostly separate entities. However you are correct that BIP101 might increase the blocksize limit beyond projected trends in terms of network resources, from the perspective of the pools and other full node operators. Which would indeed decrease node count and therefore would be bad for decentralization.
More users does lead to more full nodes, however in order to support more users we need to increase the block size which does make it more difficult to run full nodes and would therefore decrease node count. This does need to be a balancing act, and your criticisms of BIP101 are valid that it might not achieve the best balance between these opposing pressures.
A choice between two bad options isn't much of a choice at all. I am confident we will have a far wider range of options on the table to choose from.
I agree, and I am also confident that we will have a wider range of options available to us in the near future as well.
