UNBEATABLE [white label]

UNBEATABLE [white label]

View Stats:
Checker Oct 2, 2021 @ 5:20pm
How exactly does score work? Got a strange result
The minutia of scoring system seems to be somewhat arbitrary and I'm having trouble understanding how a exactly score is assigned to performance in a song. Here are two different results from Waiting on Unbeatable difficulty:

Accuracy: 96.44%
Max combo: 457x
Perfect: 428
Great: 27
Good: 1
Okay: 1
Barely: 1
Miss: 0
Avg. Delay: 4.84 ms
No Misses (not Full Combo)
FINAL SCORE: 1,243,174

Accuracy: 96.83%
Max combo: 458x
Perfect: 431
Great: 25
Good: 0
Okay: 2
Barely: 0
Miss: 0
Avg. Delay: -0.55 ms
Full Combo
FINAL SCORE: 1,239,561

The latter of these appears to be better in every single way: higher accuracy percentage, full combo, smaller average delay, more Perfects, and fewer notes below Great. Why then is the score a few thousand points lower?

Thanks!
< >
Showing 1-2 of 2 comments
Siory Oct 4, 2021 @ 7:35pm 
Perfects give different amounts of score based on how far off true perfect you were (the millisecond number following the callout) instead of just a fixed score. I haven't gone through and attempted to determine the range and the exact mapping just verified that's what's happening by checking a recording. When you're maintaining long combos and hitting a lot of perfects, this variance can add up to quite a bit of a score difference.

I looked into the behavior for similar reasons; a score that seemed to not really line up with the accuracy percent (which counts all perfects as the same fixed value). Since then one of the better guides here has been updated with even more detail on just how scoring works: https://steamcommunity.com/sharedfiles/filedetails/?id=2611502333
Checker Oct 4, 2021 @ 10:11pm 
Thanks for the very helpful and detailed link!

I did know from experience that delay affects the score awarded by a Perfect note, but I'm still left somewhat confused because in this case my average delay was dramatically less as well.

Thinking about it more as I write this, the only way I can rationalize this is if the average delay actually *was* higher, but was an almost-even mix of too early and too late that produces a smaller "average" delay.

I suppose the operative quantity is actually the average *absolute* delay, which is unfortunately not visible on the score screen.
< >
Showing 1-2 of 2 comments
Per page: 1530 50