What are some of these headers supposed to mean?
"Points per win" "Penalty per loss" "Penalty per loss below 20"
The thing is there are ranges of wins and losses and I don't think rank really plays that much of a difference, but I have been in the A's for a lot longer than C's or B's. So that's why this is a discussion entry and not an official page entry.
My observation is this: If the average rank of team A is higher than average rank of team B and passes an unknown threshold, the match will award ±12 and ±8 instead of the usual ±10 for each side depending on a win or loss. If you are the higher ranked team, you get a -12 or a +8 depending on a loss/win. Vice versa is true. If the difference in average rank on a team is even greater, beyond another unknown threshold, you get a ±14 and ±6 reward.
The most data I have is on a 7 B+ and a single A+ in a room. The A+ player was rewarded in the -12, +8 tier. I cannot confirm that the same was applicable to B+'s on the same team, but expect it would be. I can also say that the average team rank matters as I was able to be an A+ and earn a +12 due to having teammates of lower rank on my team. And this was late night gaming, but I want to say I got a -8 for a 3 A- and A+ loss against 4 A+s. If it just took the individual's rank against the opposing team's average rank, an A+ should never get -8 or +12 rewards.
- Points per win: the amount of points you get towards your rank when you win
- Penalty per loss: the amount of points you lose from your rank when you lose
- Penalty per loss below 20: the amount of points you lose from your rank when you are below level 20. Luke (talk) 21:19, 2 July 2015 (UTC)
- It's like Zyst says, the amount you gain and lose depends solely on the rank of your opponents, there's no "lose 8 points if below 20" rule, it was added there possibly due to a mistake with understanding the system. The page should be fixed asap. From what it seems so far, if the difference between your opponent is more than 2 ranks, it'll affect the score by two, seems to be the same for difference of 4 ranks.
- I would be in favor of removing the Penalty per loss below 20 column and also including a description like PPLToast's description, especially so if it can be confirmed. My data has [B-,C+,C+,C] vs [B-,B-,B-,C+] is enough for a ±8\±12. If I gave a numerical value to each of these ranks, it'd look similar to this: [4+3+3+2] vs [4+4+4+3] which is  vs . That fits in with PPLToast's prediction of a difference >2 ranks will change the score.
- More data is with [B,B-,B-,C+] vs [B,C+,C+,C]. Again with a numerical value, that would be [5+4+4+3] vs [5+3+3+2] = 16 vs 13. This was a ±8\±12 game as well.
- Well, the fact that points are in fact skewed based on your current Rank Points is finally confirmed by an official source via the 2.2.0 Update notes, just on a different scale.
Rank Rank Point Range Win Lose S 0-39 +5 -5 40-79 +4 -5 80-99 +4 -6 S+ 0-39 +4 -4 40-79 +3 -5 80-99 +2 -5