NORTH AMERICAN Rankings

After testing a variety of models, including variations of ELO, linear regression, and team-based models, we have decided to update our model this year to the Glicko-2 model. It is an improvement on the ELO system that not only takes into account players' skill level, how confident the model is in their skill level, and a player's volatility between tournaments. These added dimensions allow for a better and more stable representation of players and allow players to move independently, but base match predictions on team-level data.

We have received feedback from players about the ELO system we have been using for the last two years, and have decided to change models to address some of the concerns that the community has raised. The first major piece of feedback we heard is that the partner you are playing with is not taken into account when updating a player's ranking based on a result. The old model functioned by comparing your ELO to the average ELO of the other team. You then moved based on how much you exceeded expectations (See explainer video for more details). Your expected score did not change based on who you were playing with. This led to situations where players playing with partners with a large rating gap would not have their projected win probability match the actual probability, and created big swings in rating. The reason we had chosen this way of doing things was to meet the goal of the second pillar, where players needed to move independently in order to converge over time. This model was also more accurate than a model where you average the team's ELO, and both players move the same amount. We chose this new model because it allows us to have players move independently while making predictions that take into account the partner you are playing with.

1. You have both a rating and a rating confidence

   • The rating is the model's estimate of your skill level, while the rating deviation (RD) is how confident the model is in about a skill level. The lower the RD, the more confident the model is.

2. Expected outcomes for matches take into account your partner

   • In the old system, each player played the average of the other team without accounting for their partner's level. In the new system, the process is that a team rating and uncertainty are calculated for both teams based on the current rating and the RD of both players (see video for more details). Then the match is predicted based on those team values. Each player on a team then gets the same credit for the result. However, they may move more or less based on how confident the model was in them beforehand. 

3. Your rating is only updated at the end of each tournament

   • Unlike the previous ELO system, your rating doesn’t change after each game. At the end of the tournament, all the games you played are taken into account, and your rating and RD are updated based on the totality of your results. This creates a more stable update pattern and decreases the importance of the order in which games are played.

Your skill is modeled as a normal distribution centered around your rating. RD is the width of one standard deviation in your rating range. Basically, a higher RD means the model is less confident in your rating. RD impacts your movement in multiple ways. 1) The higher RD you have, the more your rating moves when you play. As you play, the model becomes more confident in your rating. 2) You move more when you beat a low RD team compared to a high RD team of the same skill level. You move more when you beat a team that the model is confident is good, rather than a team that the model thinks is good but is unsure of.

Your rating won’t change if you don’t play a tournament for a little while, but RD increases with time. The longer it has been since you played in a tournament, the more uncertain the model is about your skill level. Every player has a 30-day buffer period before RD will start to increase.

While the previous threshold was 925 in the new rating system, the new threshold will be 1000. Though we will be honoring ratings above 925 in the old model for the first few tournaments to give players time to improve their rating in the new system. In addition, if you met the previous qualifications and have already registered for a later event or register for a later event by April 1st, 2026, we will honor that registration as well.

We have set a floor that your RD can reach to ensure players will still be able to move with each tournament.

1. The per-game estimates are there to help give you an idea of how the match is moving your rating. At the end of the tournament, when combining all your results together, the model has more information than just a single game, so it might not match exactly.

2. To prevent edge case scenarios from messing up the rankings, we limit the amount your rating can move in a single tournament to be a percentage of your RD at the start of that tournament

The division you play in doesn’t matter. The amount you move from a win is based on the difference between your team's and your opponent's team ratings. You move up more by beating higher-rated opponents then lower rated ones. You also move up more by beating opponents that the model is confident in their strong ability. The inverse holds for losing to opponents. The weaker they are, the more your rating will drop.

Your rating is the model's estimation of your skill level.

The median rankings for each open division are as follows: Contender-1100, Expert-1250, Premier-1500, and Pro-1900.

For women's: Advanced-1100 and Pro-1675.

Only USAR Sanctioned tournaments

Pool play, bracket play, and consolation are considered and weighted equally

Rankings are improved when you win games at USAR approved tournaments! (Especially against good competition.)

After you play in one event sanctioned in the US or Canada, you will be included in the North American rankings. You have to play one tournament a year to maintain eligibility for the rankings.

To be included in the rankings, you must have played one sanctioned tournament in the last 365 days. After one year, you will not be included on the public rankings, but you will hold an active USAR rating for up to one and a half years since your most recent tournament. If you have previously played in a sanctioned event in the last year and a half but aren't on the current rankings, you can look up your player page for your current rating. If it has been more than a year and a half since you played, you will be considered a new player when you return for rating purposes, but you can still check your match history in the player pages.

Most likely, your partner’s record and strength of schedule when playing without you is more impressive than your record and strength of schedule when playing without them.

The only tournaments considered in rankings are USAR Sanctioned events. Talk to your tournament director, and work with USAR to get your tournament sanctioned so that the results will count!

The score is not currently considered, due to a lack of completeness in score reporting. However, it is incredibly valuable long-term to consistently and accurately report full scores of each game, set, and match played.

For Players:

• Play in USAR events! The more data the better.

• Data hygiene on Fwango

  • Have only two players on your team.

  • Do NOT drop players from your team.

  • Do NOT delete your old teams.

  • Please record all your game scores (not just 1-0). Roundnet Germany has a great explainer about this

For Tournament Directors:

• Get your tournaments sanctioned by USAR

• Data hygiene on Fwango

  • Ensure all registered teams consist of only two players.

  • If a match is a three game series create space for all three scores to be entered (not just 1).

  • Consistently label your all non-bracket stages “Pool Play.”

  • Record injury forfeits as 0 to -1 and non-injury forfeits as 0 to -2.

  • At the end of the day: make sure that all played matches are entered into Fwango and publish final standings.