npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2024 – Pkg Stats / Ryan Hefner

uniqueness-rating-performance

v1.2.993

Published

Uniqueness rating performance calculations

Downloads

18

Readme

uniqueness-rating-performance

Table of content

  1. What is uniqueness rating?
  2. Layers
  3. Layer average
  4. Uniqueness value
  5. Sample results

What is uniqueness rating?

Definition: "How does this performance compares to the average best performance with a similar mod combination?".

osu! uniqueness rating (UNR) is an alternative rating system for osu! standard scores, it is designed as an addition to the current performance system such as a weighting method, or pure curiosity stat.

Uniqueness rates from 0 to 100 on a near-logarithmic scale.


Layers

Definition: "A layer is a group of supposedly similar mod combinations leaderboards".

For sample quality reasons, the system groups "similar" mod combinations pre-existing scores together, such as the balance between sample quantity and sample quality is (hopefully) respected as much as possible. However this means that some crucial compromises are made, hence why the layering system is likely subject to change and is highly experimental. The layer type is found with the following algorithm:

Layers

Once the layer type of a score has been found, we know exactly what modded leaderboards will be in the layer, so we bundle the corresponding scores alltogether to form our layer of scores. Every looked up mod combination leaderboard for each layer is as followed (SD and PF are not displayed here but are also looked up for every combination):


Layer average

Definition: "The layer average is the average performance of the best scores on a layered leaderboard".

This variable represents the comparison point that will be used to determine uniqueness, basically the more difference between this value and the score, the higher/lower the uniqueness will be.

Layer average is calculated among the following steps, from a layer:

  • Scores in the layer are sorted by performance points.
  • Following another formula that will be detailed later, a percentile of scores that will be kept from this layer is calculated.
  • Only the first x scores from the layer are kept and will be used for the calculation (x = percentile).
  • The previous step is also where we remove the score that is currently looked from the layer.
  • The average performance of the remaining scores in the layer is calculated, and corresponds to the final Layer average.

The percentile is the most crucial part of this variable, if it's low it will be hard for a score to be unique, and if it's high it will be more gentle. As of now the approach we chose is for the percentile to depend on the whole leaderboard's variance in performance (more precisely, the standard deviation of the whole leaderboard's performance).

The higher the standard deviation d is, the higher the percentile will be, and conversely. The formula is as followed:

percentile formula

Following this formula, the percentile is hard capped at 2 for the minimum value, and at the layer's size for the maximum value. Which means that there will always be at least 2 scores used in the average calculation.


Uniqueness value

Definition: "Uniqueness is a near-logarithmic scale that relies on a quotient between a score's performance and a layer's average performance".

Uniqueness is calculated through a two-step process:

  • The first step is to determine the "raw uniqueness", which is basically how much the score's performance stands out (positively or negatively) from the layer's average performance.
  • The final step is to make this "raw uniqueness" a 0-100 near-logarithmic scale that is easier to interpret.

The first formula is a straight forward trivial quotient, with variable a representing the score's pp, and variable b representing the layer's average pp. We add in a really small value in order to avoid any potential zero division:

first step of the uniqueness formula

The second formula is as followed, variable x representing the previous "raw uniqueness" calculated in the first step:

second step of the uniqueness formula