When I saw the fluff commits I was immediately reminded of a conclusion from "Punished By Rewards" - once a reward is applied, people will start doing the minimum required to obtain the reward, as their motivation has shifted from intrinsic ("I want to improve the project") to extrinsic ("I want a reward, so what do I have to do to get it").

I actually think your attitude is the right one to have for this kind of experiment. We can pontificate endlessly on what people might do, how people might exploit such a system, but I'm of the (possibly naive) opinion that just because a person can, doesn't mean a person will.

I suspect that, despite the potential for abuse, you won't actually see very much actual abuse.

Alternatively, it may influence commit behavior. People may not want to be accused of filing fluff commits & slow their contributions to accumulate 'enough' changes to justify the payout.

We added a FREEBIE keyword today for commits that are small enough the author doesn't think they should get a reward.