There’s nothing like waking up at 0415 realizing that you structured your experiment incorrectly. Thankfully I’ve only taken 3/8 of my data. I’ll probably have to throw out one data point at the most.
ETA: Happiness is salvaging all of your data points, reworking the random order to tweak what you’ve run, and not losing any data taken already. Hell, I’ll finish a day earlier. Thank you, pagan DOE gods.