First of all, I'm sorry for starting a new thread about this, but judging by peppy's response to the thread, what I was saying was just misunderstood.

Currently, only the play that has the highest score rank is chosen for pp calculations. When choosing what play is considered for pp calculations, the best play (based on modified scores for pp calculation) isn't always chosen because of this.

Examples of when this can happen:

1) Player A gets rank 300 in a map with 95% accuracy, then the same player plays the map again and gets a rank 280 with only 90% accuracy (resulting in less pp because of less accuracy). So in the end, the second play results in a penalty in pp score. This goes against the statement included in the wiki that there is no penalty for bad scores. Another similar case is getting a high-score in a map with bad performance, that makes other better plays with less score don't count.

2) (This one I'm not sure because I don't have enough details about how the pp system works) Player B gets a rank 1 in a map with a 4-mod SSH, getting a good amount of pp. Then, let's say 2 years later, the pp gained in that map practically banishes. Now player B won't be able to get pp from that map again because he can't get a higher score in that map, even though he can achieve the same or slightly less score (that deserves pp based on the performance), another player then would get more (unweighted) pp than the best player with a worse performance even if they played the map the same day.

Maybe the easier way to solve this problem is storing the best performances separately from the best scores, but that would increase the server-load.

Edit: Since this could be considered more a design flaw in the current pp system than a bug, I will post in Feature Request for a possible solution to this.

I guess why nobody answers to this thread (maybe I'm being impatient, but I am interested in other players and developers opinions about this).

Is the problem not well explained?
Maybe I am the only one that considers this an important issue?

Those are both valid points.
tl;dr version:
Accuracy is important to pp but the most accurate score isn't necessarily saved. (The highest scoring one is.)
Also, old, decaying scores might be extremely difficult to beat to get new pp for them, even though a non-best score might be worth more because it's new.

I don't see either of these issues as being a huge problem. If you beat your score with a lower accuracy, that is your choice. Might be a good idea to consider playing without mods if you can't reach the same accuracy with them enabled.

Sometimes is hard to keep track of how your accuracy is in the middle of a hard song (I can't at least afford time to watch the corner of the screen in AR9 or higher for example, even less count the number of 100's or 50's), specially in the last stretch of the song.

Also, since I don't have information of how exactly accuracy affects modified score, I can't be sure if it is worth passing a score with less accuracy and better combo when I have to decide to rank the song or not.

So play again?

There's no way of knowing if a higher score with lower accuracy will be worth more pp or not.

If I, for example, "improve" a rank 10 score to rank 1, but with lower accuracy, I have no way of knowing if submitting this score is going to be beneficial to me, or if it's going to completely screw me over because my pp will go down and I've now locked myself out of being able to get it back in any reasonable fashion.

Older scores that were made before the pp system are also an issue, because they were made with no regards to accuracy, only number of mods. A person that happens to have a lot of scores like that on very highly weighted maps is pretty much screwed because of how difficult it would be for him to beat his own scores.

This is especially true of the Hard Rock mod on hard maps (which are usually the highest weighted maps), which completely kills your accuracy, but sets a score that's impossible to beat without it.