Hi all,

We often get questions like “why does my rating change vary from match to match?”.

Your rating will move more if you have high uncertainty, play smaller team size games, or have results that are unexpected from the Openskill library win prediction.

I’ve added Openskill library win prediction in the match > players tab on teiserver. Login to the website and view a past match at https://server4.beyondallreason.info/battle/ratings. Then click the players tab. You will see win prediction at the bottom.

The team with the higher skill is expected to win. Note that skill is different from match rating. Match rating = max(0, skill - uncertainty)

Higher uncertainty values on either team will cause the prediction to move towards 50%.

The balancer treats new players as being the worst in the lobby. However, the Openskill library rating function considers new players to be average (starting skill of 25). This has the benefit of handling smurfs correctly. But you will also notice some lop-sided win predictions when new players are involved.

If you have any questions about a past match balance, feel free to ask me as we do have the logs