FWIW I got it working with some hacker-pokery:https://data.stackexchange.com/stackoverflow/query/1537146/link-only-answers-that-suggest-duplicate-questions-split-by-reputation?opt.withExecutionPlan=true
From your last message it seems that you solved your problem in the meantime. But at least it is a reminder for the visitors of that room, that also this SEDE-related room exists.
@Pureferret I think this one has the same functional meaning but without the top 50,000. It selects the answers from users under 100 rep where the answers contain http, https or stackoverflow, groups those into 3 reputation buckets (<15, [15,50], > 50) and count the number for answers for each bucket.
But you would like to do it in a way which avoids having to put stackoverflow twice in the ULR. (Once as a parameter and once as a site where you're running the query.)
If nothing else, at least I think I understood what you're trying to do.
When people do various network-wide queries, they use things such as sitename or hostname. But I am not familiar with this well enough to be able to help.
You can see an example of a query run over several sites here:
Nope. Don't worry, it is not much work. most of the query comes from my default template I have for that: https://data.stackexchange.com/stackoverflow/query/1004129/multi-db-template-with-hostname