I see that in the release notes for BC 31 they have added spender indexing. This is integral to BS, and I assume they are trying to close these gaps where possible. This will give us a more comparable store size metric, since this is a large index - especially if you also store whole blocks independently as BC does. With this and tx indexing enabled and with our address index disabled, the indexation is fairly comparable. Yet we are about the same size without it, which means this should highlight the space inefficiency of storing whole blocks. IIRC client filters are optional for BC (as for BS).
(edited)