To Justin O’Dell and whoever else is working with the new PvP system,
I really like the idea of using final score and odds of victory as the metric for ladder ranking so kudos to you for coming up with it! Doing something like this is what we have long needed. Even though ELO/MMR rating was good, it never captured how close a game was. If we only counted wins and losses, then we would find TCU or Baylor in the college football playoffs right now, even though they clearly had weak non-conference opponents.
I thought about it for a while, and a simple modification could be to eliminate the “0” points on the chart; instead changing it to another specified value depending on victory odds. The current system basically allows a player to grind their way to the top because of an example like the following: If a team has a 80-100% chance of victory, the likelihood of the team losing with 0-199 points is nearly 0%. If you assumed the point spread was a normal distribution, then this would be a valid way to score. However, most capture point games can be kept quite close and may have a skewed distribution towards the 300-499 point loss results.
Ultimately, it’s all speculations, since we can agree people aren’t taking this test season as seriously, but I trust you’ll re-examine how the points are given based on the two metrics again to make sure the best are really at the top since that is the purpose of the ladder (and not who plays the most). Comments or questions anyone?
Source:
https://www.guildwars2.com/en/news/finding-the-perfect-match/
(edited by Ekelon.7041)