You'd need to consider some additional factors:
1. If you are at 0 points, you might not lose any (your score cannot be negative), but your PPG can still decrease. For example, if you sit at a 0 table, finish below 5th place for 9 consecutive games and win the tenth in a 5 player game (without killing anyone), your score will be 100, but your PPG will be negative: (1*100-9*50)/10=-35.
2. If you are away from the game for a few days and you have a low score, you might be given random boosts between 50 and 3000 points. These do count towards your points but not towards our PPG. The same is true for any boosts that you receive for purchasing a membership (or just extra points) or sending email invitations to your friends.
3. PPG seems to be prone to rounding errors.

He forgot tourney buy-in too, but his claim is true. The more games are played, the more does the PPG draw to 0. If you check the PPG's of the most active players in the past months, you'll find that lots of them have the same issue.

@above:Yeah, perhaps stating that it's prone to rounding errors is an understatement.

I reckon the method of calculation is a very sloppy recursive formula.

Something like NewPPG = Floor( PreviousPPG + Pointsgainedinthelastgame/Numberofgames )

On second thought, I've left out a term in the above formula that should explain why PPG tends to zero as the number of games increase.

So it might look like something like

PPG := Floor( ( PPG*( #Games - 1 ) + Point.Change )/#Games ) )

Where Floor is a floor function that truncates the number to an integer, #Games is the number of games, Point.Chance is the point chance from the last game. Using a formula like this would almost guarantee that after a certain number of games, PPG would converge to zero unless someone plays on the highest tables (2k and 5k).

Thanks for the extra info I did not consider. With that said, I would still expect it to be over 10 though