@Martin FYI, my new answer here contains a query with (what I think) is the maximum possible information we'd get from SEDE. It gave a pretty accurate number (+7008) for my own account on Meta.

When searching for posts where some string is contained in the body or in the title I can use (Body LIKE '%##text##%') OR (Title LIKE '%##text##%') or (CONCAT(Title,' ',Body) '%##text##%').

Could the difference in speed between them be large enough that I should prefer one of them over the other when writing SEDE queries?

For example, if I have the two versions of this query: data.stackexchange.com/math/query/1158173/… data.stackexchange.com/math/query/1158175/…

Or this query: data.stackexchange.com/math/query/1158177/… data.stackexchange.com/math/query/1158178/…

@Martin Now I know what I missed in that query. (The value of OverCapRep was NULL on some days.) Here is a corrected one, which is much closer: Testing an estimate of reputation (with daily cup, without offered bounties).

Here is a query estimating daily breakdown: Estimate of my reputation by day (with daily cup, without offered bounties, without suggested edits). I tried this so that it can be compared with meta.stackexchange.com/reputation

Of course, I meant to write "daily cap" rather than "daily cup".