ppy
Unexpectedly high star rating change from slider tail leniency
Type
Other
Bug description
Beatmap: https://osu.ppy.sh/beatmapsets/660630#osu/1398809
Web: 7.64* Master: 8.01*
It looks like the change that caused this was https://github.com/ppy/osu/pull/24966, but the difference is much higher than I remember was an acceptable difference from that PR. It should probably be investigated, for sanity's sake.
Screenshots or videos
No response
Version
ef2e2302d4cf3f049c5637527f99b311d999c43b
Logs
On a quick look, yes this beatmap did indeed gain that much star rating due to the slider tail leniency change, however it looks like an edge case to me. The biggest difficulty peak on this map is the following "buzz"slider:
Note the length of it and the 5x slider velocity. osu-tools shows this:
| overall | single object | |
|---|---|---|
| pre-#24966 | ||
| current master |
This just looks to me like the actual diffcalc/ppcalc code may need adjusting to account for the new slider leniency mechanics correctly. I'll ping @ppy/osu-pp-committee to confirm the above assessment, but I don't really believe this needs to be p0 at this time.
Notably the beatmap here has even been mentioned in review of the pull in question and the difference is explained by fp stupidity which I can believe on a slider this jacked as this...
As I said before, I think its fine-ish - the map is getting higher star rating and more pp, but it's not really upsetting the balance in any way, not really abusable (the pattern is still hard and the map overall is underweighted) and can be addressed later by diffcalc devs.
