The thing that jumps out in these “I wasted X mg” posts isn’t the wasted compound. It’s the line buried in every one of them: hadn’t slept, cat woke me up, did it half awake. The mistake is downstream of the sleep, and so is the part nobody tracks. Here’s the cost people skip. You botch a draw at 6am, you don’t actually know what went in. Maybe 1mg, maybe 1.4, maybe you double-dipped. That dose is now a phantom point in your log. Every readout for the next week sits on top of a number you can’t trust. If you’re titrating, or trying to read appetite or whatever endpoint, you just contaminated the data, not only the vial. I run subq twice daily on a recovery protocol and the rule I landed on is dumb simple: I don’t draw on broken sleep. If I’m under maybe 5 hours or woke up fogged, I draw the night before into a fresh pin and leave it labeled, or I just take the morning one late once I’m actually awake. Costs nothing. Removes the exact failure mode every one of these threads is describing. Fragmented sleep is the upstream variable and it’s doing more than making you clumsy at the bench. It’s the input nobody logs and the cause of the error they all log. I started noting hours slept next to each dose for unrelated reasons. The dark-mode chart colors are the reason I actually open it at night. Turns out the bad-draw days cluster exactly where you’d expect.