While rating some Braingle teasers recently, I started taking notes on how my
vote affected the ratings. Some of my data is presented below. Can you figure out what is the likely formula that computes the ratings from the votes? Can you determine for each teaser how many votes were cast for each attribute prior to mine, and how many of those were positive (fun/hard)? Remember, not every voter necessarily votes for both attributes.

Notation: The notation for Teaser A below shows that it started with a fun rating of 2 and difficulty 1.33, based on 6 voters, that I voted positively on the first attribute (fun) and negatively on the second (easy), and that it wound up with ratings of 2.29 and 1.14, respectively, after my vote.

Teaser A: (2.00, 1.33)[6] (+, -) -> (2.29, 1.14)
Teaser B: (1.14, 0.67)[7] (-, -) -> (1.00, 0.57)
Teaser C: (3.00, 0.67)[8] (+, +) -> (3.11, 1.14)
Teaser D: (3.11, 0.89)[9] (+, -) -> (3.20, 0.80)
Teaser E: (2.33, 2.46)[13] (+, -) -> (2.46, 2.29)