The biggest thing with handicapping individual games is the volatility of game scoring (as Jeff mentioned). It's impossible to create a handicapping system that works for all games with the vast breadth of games with linear scoring versus those games with more dynamic scoring opportunities.
As far as PEPGA goes (PAPA's handicapping system back in the day), it developed a handicap based on league points earned, so the outcome of each game was normalized no matter what the game played happened to be.
You can check out the old Flipside magazine (http://legacy.papa.org/flipside/v1i1.pdf) page 18 has some league standings from 1992 with the PEPGA averages for each player during the season.
This would lend itself towards handicapping a group of players in a league by comparing their performance versus their expected performance.
If there's a 40 game season, and Steve Epstein has a PEPGA of 5.33, there's an expectation that 213.20 would be his season total.
You could then look at Tim Post who had a PEPGA of 3.56, and there would be an expectation that 142.40 would be his season total.
PAPA could then use PEPGA to declare Tim the 'handicapped' winner of the season if he finished with 150 points, while Steve finished with 215 points.