Song lengths are very inaccurate. (And another suggestion)
Posted: Sun Jun 30, 2013 1:42 pm
When you select a song online it gives you a length. Much of the time these lengths are hilariously inacurate. Is there any way this can be fixed? Or removed? The times are accurate far less than 50% of the time.
The reason for the inaccuracy appears to be that the server saves the length of the file the last time it was played, even if someone quits out early, it counts this as a finished file.
A possible way to make it more consistent:
Have the server store the longest length of a given song, provided its unique song title, artist, and difficulty rating. Very few files are named identically with the same ratings so this would narrow the inconsistencies massively. Basically if someone plays a file and it ends up being shorter than the already recorded time, count this as a quit out.
On top of this, I would recommend trying to find a way to not store results from files that are quit out early. Here's a quick example of why this is a problem:
Basically you can gimmick the leaderboards. This happens intentionally sometimes, and unintentionally some other times. Maybe implement so that popular files, maybe files with more than 50 plays, can be compared against the average number of total steps, and if the file is far below the average, don't count it?
The reason for the inaccuracy appears to be that the server saves the length of the file the last time it was played, even if someone quits out early, it counts this as a finished file.
A possible way to make it more consistent:
Have the server store the longest length of a given song, provided its unique song title, artist, and difficulty rating. Very few files are named identically with the same ratings so this would narrow the inconsistencies massively. Basically if someone plays a file and it ends up being shorter than the already recorded time, count this as a quit out.
On top of this, I would recommend trying to find a way to not store results from files that are quit out early. Here's a quick example of why this is a problem:
Basically you can gimmick the leaderboards. This happens intentionally sometimes, and unintentionally some other times. Maybe implement so that popular files, maybe files with more than 50 plays, can be compared against the average number of total steps, and if the file is far below the average, don't count it?