Follow us on Twitter @LiquipediaVAL if you'd like to be kept up to date on all things VALORANT!
Download the Liquipedia App here!Download the Liquipedia App on Android or iOS!Want to be updated on everything you love in esports? Download the Liquipedia app on iOS or Android and never miss a beat of your favorite tournaments, teams, players, and matches!

Liquipedia Rating: Current

From Liquipedia VALORANT Wiki
This portal is outdated.

Explanation[edit]

What are the Liquipedia VALORANT rankings?[edit]

Liquipedia's main goal was to create a rating of all the VALORANT teams, based purely on data and therefore not subjective. The rating should reflect the current form of a team and make it possible to rank teams accordingly. In addition we also want it to be able to do this across regions with at least somewhat precise results. This is where LPRating comes in!

How does it work?[edit]

There are a lot of rating algorithms out there with probably the most prominent one being Elo. Elo was developed by Arpad Elo in 1960 to create an objective rating algorithm for the US chess federation USCF. It proved to be so successful that it was subsequently used worldwide. So, is LPRating just another iteration of Elo?

A little bit, but not quite.

Elo is a really interesting algorithm which fits for a lot of sports, but for numerous reasons, it doesn't fully work for VALORANT:

  • Elo is so accurate because so many players are ranked. VALORANT in comparison has a very limited number of teams that are consistently playing matches and appear in our database. Nothing compared to 360,000+ in chess rating.
  • Elo needs teams to play each other often and needs enough matches across skill gaps and regions. In VALORANT, especially cross regional tournaments are rare (for now?).
  • Elo is based on the assumption that a contenders skill is fairly constant, at least over short periods of time. In VALORANT, form can drastically change quickly. Even daily form can make the difference of a low tier team beating the higher tier one.
  • Elo always has even win- and loss-modifiers, which means that every point gained by a team is the loss of another one. Looking at VALORANT, this would create a problem for teams in S-Tier tournaments. If the rating is designed to react to form fast, teams could potentially lose a lot of points during a season to allow gains for in form teams - but a team going 1-6 in a S-Tier event is still probably top 10 in their region.
  • A team will only be listed once they have ten (10) concluded series registered on Liquipedia database and have a team page set.
  • Considering how competitive VALORANT works, different multipliers are applied for results on whether they were best-of-1, best-of-3, best-of-5 matches, and so on.
  • Teams under the same name or organization that have changed 3 or more players will be treated a different roster and points won't be carried over.

Considering these constraints, we made several changes to the Elo-Algorithm to allow rating VALORANT teams in a more accurate fashion, which all come together as the LPRating.

New regions are added once at least five (5) active teams within the region are eligible for a team page. Regions are mostly split according to the official Riot Games regions division.

Which results are included?[edit]

The LPRating includes all matches on Liquipedia, starting from April 2020. All tiers besides show match are included, but higher tiers like S-Tier and A-Tier are given more weight.

Recent Updates[edit]

  • 07-12-2020: Lengthened initial uncertainty period from 20 to 30 matches.

The Rating[edit]