So I'm guessing the experiment is either A/B testing of layouts of number pads, or it's testing to see if people can cheat the leaderboard.
Also, it must be a slow news day for this to get to front page.
Also, it must be a slow news day for this to get to front page.
So, the lesson is: give your users a perceived reward for taking action on your site, to get them engaged long enough to get the reward.
Or maybe the lesson is: as soon as users perceive there are no rewards left they'll leave.