Watching that video of the Michelin coupon drawing yesterday reminded me of being at O'fest 2005 for the raffle drawings. I think a total of three nights were used to draw the tickets, so I was able to observe the process. Being an old engineer, I paid particular attention to the "mixing function" that the metal drum was applying to the tickets inside. It appears that metal drum was the same or very similar to the one yesterday in the video. Yesterday it seems that it was perhaps 60% full and provided (hopefully) reasonable, random mixing when rotated. At O'fest 2005, the drum was very full of tickets. I don't recall the specifics, but watching the process each night I could see tickets around the periphery of the drum that simply stayed stuck in place when it was rotated. In other words, it wasn't capable of providing a nice, random mixing to all tickets in a reasonably equivalent manner. I would guess a drum like those used probably needs to be 1/2 full at most to randomly agitate all tickets evenly? I'm only bringing this up since that video reminded of 2005, and I had the thought of how it would be likely be "fairer" to have an electronic random selection as opposed to a sub-optimal mechanical one as I witnessed back then. Perhaps there is some other solution now in use for the master drawings, so pardon the post if so.