RaneFire wrote:
If you browse the "beatmap difficulty" page here, there are a lot of maps that should not be there, but for some reason they are considered difficult... It's weird; "Molella" map, "Wonderland", "Tori no Uta" feature on the 9th page sorted by difficulty with maps like "VALLIS-NERIA", "Jiyuu no Tsubasa", "Ningen Shikkaku", "Intersect Thunderbolt", which are obviously way harder than the former 3. (Haven't checked them all)
There's something weird going on, because this opens up farming by difficulty when there are maps that shouldn't be in that position. And they do give a considerable amount of pp, since I played them myself.
The difficulty calculations are far from perfect, and they will continue to get tweaked in the future. Keep in mind, that a lot of this is very subjective, though, and you might find maps easier, that others find harder. Gotta agree with you on most of the examples, that you called, though.
electrolytes wrote:
It's an odd case, but Winshley's SSH on paraparaMAX I seems a little underrated (at least on osu!tp), given the length and all. I suppose it's also hard to programmatically identify the occasional map where HD is harder than HR.
I'm not exactly sure how fair it'd be to further increase the map-length bonuses cap. Will consider it for the future.
PlasticSmoothie wrote:
Are there any plans to include tp's difficulty calculator into the pp system? (Miiight be the wrong thread to ask this in, if so sorry)
It is included to a very high degree.
anticlone111 wrote:
Well, that system seems to work much better than previous one, but it's still far from perfect.
So i have a few complaints there:
1)Top score ranks on maps doesn't give anything at all, even 1's is worthless. You should definitely add pp boost for highscores,so we will see more competition for high ranks especially on hard maps.
For example:
Top 50 - 2% pp boost
Top 10 - 5% pp boost
Top 1 - 15% pp boost
2)Hidden should affect accuracy much more.
3)Short maps and fullscreen jumps - OP.
4)If anyone can farm pp without even getting in top50, how you will ever been able to track down cheaters?
5)Waiting for pattern difficulty recognition.
1) That would add way more complexity to the algorithm, making it a lot harder to consider every score (since then every score has to be periodically checked for which rank it is). In addition to that - as mentioned multiple times - contest on maps is different. For instance way more good players play [Insane] maps than [Hard] maps and thus what you suggest would boost [Hard] maps without any solid reason.
2) Also answering all the other people who gave feedback on this: I disagree, that hidden should affect accuracy more. Playing with the rhythm of the map is not very connected to reading the map. The small bonus is gives to accuracy is only there to make up for rare pattern misreads, that might occur while playing with it.
3) Many long maps are worth more than short maps of equivalent difficulty. See chipscape or fallen world. The reason why you see mostly top-pp scores on short maps, is because there
exist more short maps. On top of that many players seem to prefer to rank on those, since it's also easier. :p
4) I think it should be common sense, that if the pp algorithm can look at sub-top50 scores, then other things can do, too.
5) Planned, but it might take a while to arrive. This has a lower priority than getting the other gamemodes implemented.
some food for thought
