Something not working because he reset actually proves why it is useful for custom support to have data on resets. Not that I would expect any hardware use data to be protected if I opted into a beta testing program. The whole point of that is to see how the hardware handles beta software.
Often with these programs the updates are just sent out at a certain time, so it's entirely possible to just miss the update period. Heck, I've seen that done with production updates, let alone beta ones. A friend with a Droid from Google IO didn't turn it on for a couple years, missed an update which was no longer available, and missed the termination of the free service to boot.
>Upon turning on the phone and registering it on the network you must agree to the EULA. The only difference between your phone and a MOTOBLUR device is that you don't have a user created login. A random login is created for you to optimize the setup experience. If you don't connect any social networking applications your phone doesn't connect with our servers.
It involves a hardware ID, and a profile that is created when the device is initially synced. It may also send a "firmware flash counter" but it wouldn't be necessary.
I'm guessing the firmware update downloader sends some sort of hardware identifier, plus a firmware flash counter, to the update server. This would be the obvious way to prevent re-updating a device that reset itself or was manually reset after a firmware update hosed the system, which would trap the user in a loop. A (device id,firmware revision,reset count) table somewhere in the motosphere isn't exactly PRISM.
Outside a tiny number of hardcore custom ROM tweakers, 21 resets almost certainly indicates some sort of problem.