But more mania score. Unlike score in the other game modes mania score is actually not a bad metric to rate skill - and it is built with more meta-information than pp such as where misses occured etc.Therefore mania pp trusts mainly on simple score. I'll admit that the difficulty algorithm itself is far from perfect, though.
For several players it is easier to get good accuracy in jumpy maps, compared to maps with long or fast streams, so top pp maps tend towards jumpiness, even for players that can stream really fast (with exceptions where they are exceptionally good at streams). Do you think it is balanced like that?
Top pp maps tend to include streams, jumps and accuracy all together. I honestly think the balance in standard is pretty good atm - by far not as bad as for instance CtB. Wish I had some more time to work on that atm.
You can switch between input devices as much as you want, no cheating.
Hey Tom, i recently downloaded your skin and found it really awesome expect for one thing, when you click an hit-circle it leaves a shadow of the same colour that really confuses me, how can i remove it? which file should i edit?
Why is AR 10.3 so under rated by your system? There's a bunch of 5 star maps that are 10.3 and way above the skill level associated with said star rating.
Sorry just wondering :0
Star rating does not take AR into account - that is done at pp-level at the moment. I should probably further unify star rating with pp at some point...
I made the taiko rhythm pattern complexity algorithm simpler and faster to calculate: http://pastebin.com/cFGUJdGa
Now it doesn't depend on the map speed (or mapper-set bpm) at all. Do you think this (or something similar) could be used in the difficulty calculation system?
This looks cool. I had something similar in mind with the concept of looking for repetitions - really cool that you made this. I'd probably sort the strain values decreasingly and build a diminishing sum, but that doesn't change the concept of the algorithm. Can someone more experienced with Taiko let me know if the resulting difficulties seem better than the current system in place?Here is a link again: http://a.pomf.se/xwdmtj.xls
I released a character build which is very strong and which utilizes certain items which then dramatically raised in price due to the short supply.My IGN is shamelessly stolen from Milkshake (with permissions!) since I had no clue what a good female character name would be: Pirteloli.
How does PP for maps like this https://osu.ppy.sh/b/441646 get bugged? Is it the algorithm being 'bad' or something else and what do you do to 'fix' it?
In this case it was the mapper changing the original EXHAUST diff to INFINITE and adding a new EXHAUST diff. The pp processor still had the old EXHAUST map in RAM and computed pp for the new EXHAUST diff with the difficulty of the now INFINITE diff. Kinda stupid - I need to build in some special case that detects this. ^^'
I thought Ignite was a little strange, but something is really off about the pp on this: https://osu.ppy.sh/b/441646 Where is the pp coming from on such a relatively low star difficulty? It's 100+ pp more than some DT maps at similar stars and accuracy.
I will check it out. The difficulty rating of the map seems right, but the pp seems off.
I take it you're responsible for Void Battery being super expensive now.
Yup. And I bought some in advance to make profit. ^_^ Not gonna repeat the mistake I made with Rainbownuke where I didn't get anything out of it in the end. :(
Does slidermissing (where you dont follow the slider all the way, but dont loose your combo) give you any less pp? (except from the accuracy bit ofc)
The slight lack of combo (1 less per slider 100) and the reduced accuracy both affect your pp, but that's it.
For autoconverts, is the pp the same across all modes? For example, ExGon's HR SS on BD's Airman in CTB is worth 530pp. If someone were to SS it with HR in any of the other modes, would it still be 530pp?
How many months until we can sort rankings by aim/speed/accuracy? Or how long until we can see people's best aim/speed/acc plays from their profiles. I think just like in tp this would be great to get an idea about what a player is better at.
1000On a serious note I also want to get this in. It'd just take considerable effort to reorganize things to even allow for this to happen performance-wise. Even with the current setup a whole pp re-processing session takes half a week.Work towards that slowly will be done when I have more time again (currently I have to focus on my bachelor thesis and upcoming exams) but I can't give you an ETA that's even remotely correct. Sorry!