> Because the "points" have no intrinsic value except in comparison to your classmates
This is where your logic is wrong. The professor is not saying the class is graded on a relative scale (for example, top 10% of the class gets As). If everyone gets two points, and it bumps up everyone's GPAs, then of course it will benefit them all, especially in comparison to people who are not in the class.
This is not necessarily true. If you're already a good student, this devalues your GPA, mostly due to the generally awful GPA system. Assuming your school considers everything above a 92 to be a "4.0", if you're already in that range, you definitely don't want to help anyone else into that range.
Note that, of course, this would be fixed by just making GPAs decimal numbers out of 100, the average of all of your actual, numeric grades from each course (multiplied by credit hours?). You can divide it by 25 if you really want the weird 4.0 scale for some reason.
Yes, it's true that it sort of benefits everyone in the class vs people who did not take the class, but some classes are easy A's and some aren't, and everyone knows that. If the class becomes a slightly easier A because we all picked 2 points, then the effect will be relatively small. It would be a better experiment if he was handing out cash, even small amounts of cash.
This is where your logic is wrong. The professor is not saying the class is graded on a relative scale (for example, top 10% of the class gets As). If everyone gets two points, and it bumps up everyone's GPAs, then of course it will benefit them all, especially in comparison to people who are not in the class.