dav1312

Looks like Discord doesn't like `yellow` as a color... *sigh*

Is this still an issue? I rescheduled this test https://dfts-0.pigazzini.it/tests/view/62ef25a56f0a08af9f766ad5 and the bounds were kept https://dfts-0.pigazzini.it/tests/view/62f4ccb663129400ae5affa6

I have a way to make it easier to realize if you have the wrong bounds but there are some PRs that I want merged first

> The SF version on lichess is changing very often, right? Depends on what you consider "very often". They are currently using a +1yo version. https://github.com/official-stockfish/Stockfish/commit/1946a675 

In dev it does feel slow af to change pages, ifk if its the change or the server ☹ Edit: 😳 

> As expected it is very slow. The indexes (created in server/utils/create_indexes.py) should have a partial filter expression `deleted=False` I think. Something like this? ```diff db["runs"].create_index( [("finished", ASCENDING), ("last_updated", DESCENDING)],...

@vdbergh https://github.com/glinscott/fishtest/pull/1425 will enable more complex searches to be performed easily. Like STC, STC + Green, LTC + Yellow, etc. Should indexes be created for all of those cases too?

Yeah, prod doesn't like it too much  But adding all of these combinations seems too tedious and definitely not scalable...

> To me 300ms seems fine for a complex query (if I understand correctly your numbers). I'm taking the values from [freemonitoring](https://cloud.mongodb.com/freemonitoring/cluster/V726YST63VGLDMUUVB4GT34WZHIY6L4D) Going to this page [Prod LTC + Greens](https://tests.stockfishchess.org/tests/finished?page=80&success_only=1&ltc_only=1)...

> The bench of the master query with the indexes is 200x slower! You mean the new indexes with master?