what-if analysis in JSON facet

It's literally what subfaceting (and earlier pivot facets) does. However, it's more complicated with you know what? Right! Block-Joins. Completely maked up sample follows:

Giving those favoured SKUs with colours and sizes. I'd like to see sizes distribution (facets) depending on dominating colour per product documents. Eg I need to enumerate sku colours with term facet, then for every sku colour I want to search for products where this colour dominates - it's comprehensive combination of {!frange}s, summing {!parent score=total constant score queries ^1 and funcs like div() and sub() etc. Then I want to facet those new docset for sizes.

Proposal

• Implicitly tag incoming bucket docset to let to exclude it via domain:{excludeTags:[enclosing_bucket_docset]};
• Implicitly inject the incoming bucket docset as a query parameter to allow to include it in query like query:"{!frange l=10}{!parent .. score=total}({!query v=$enclosing_bucket_docset}^1)"

Questions

1. How you gonna do something like this alternatively?
2. What about those two non-invasive improvements?
3. How to name those tags and param?