Would it be possible to increase performance for pagination with indexes for large collections ?

Consider a collection with 200000 documents, paginated with $skip and $limit and indexed on the sort order.

The first page takes a few milliseconds, but the last page can take a few seconds because the index is scanned from the start counting the $skip.

What if, as an optimization, the query planner would check if $skip is greater than half the amount of indexed documents and if it is: go through the index in reverse and shift the results into the array instead of pushing them.

This would result in never having to scan more than half the index.