Oh believe you me, it doesn't make sense to anyone. However, the fact that your system detects (somehow... dd you set UPS settings or load any software or have a USB plugged into it) the UPS and has settings for %, could very well defeat the autopoweroff setting, since it no longer draws power from the wall "per say" (well, it does if you count the UPS pulling power from the wall, but you get what I'm trying to say).
I've never actually plugged a Mac just directly into a UPS and expected it to detect that it's plugged into a UPS as opposed to a regular outlet because in the PC world, a USB/Serial connection and software or native OS UPS setting is required for it to know it's on a UPS and for that UPS to signal it's battery level when it's low so that a shutdown can occur on the machine. I'm not saying the Mac charger is smart enough to do this because it can't be... but I'm betting you've configured your laptop to know it's on a UPS, which most likely defeats the other timers because the timers for the % of battery left on the UPS would need to take priority over the others. This would be the most logical explanation because in all my testing it's been pretty evident that PMSET core functions in the OS are not very dynamic and the OS must power up the system in order to change PMSET modes of opperation... this is after all why this "wake on power disconnect/connect" additional issue after this update is also happening.
Now, as to one change that Apple has made to 10.8.3 in regards to this hibernation wake routine is that... if the system is fully charged and left plugged in and does go into hibernate after 4 hours. When you do unplug it without opening the lid, the system WILL wake up, but it will reenter its previous full hibernation state. Prior to the 10.8.3 update and after the Update 2.0 which they released after 10.8.2 was out (the update mentioned in the KB above), if you unplugged the system after full hibernation was in effect with the lid closed, the system would wake and then go back into normal sleep as opposed to hibernation (which inherently caused it to eat up power because sleep keeps memory powered to maintain the sleep image for quick wakes.) This was one of the gripes I had filed with Apple back in November and took me a couple of months of going back and forth with them for someone higher up to "get it". Apparently, they addressed this one issue because it makes sense that if you're going to hibernate the system fully, why wake it and put it back into a mode to just consume more power than the mode it was just in. So, at least they fixed that.
Now, I wish they would just fix the issue where it ACWAKES even though ACWAKE is set to 0, meaning that the darn thing shouldn't be waking up at all when we attach/detach it's power cord. That's just stupid and unnecessary! Windows boxes don't need to do it and heck... even my Mac with Windows running in Bootcamp won't wake up when I plug/unplug the AC adapter to it while it's asleep or hibernated in that OS. Yep, OSX is just dumb. :rolleyes:
Sorry for the long post, but I hope this gives you guys a bit more insight into this problem.