15 / 36 == 0.417 (hr. sleep / hr. awake)
9 / 15 == 0.6
Conclusion: to be awake more often (less sleep per hour awake), then 36 hour days are more efficient.
Of course, the "last night" data is bad since it only includes one sample. A more conclusive study might monitor one's 36-hour-day sleep patterns over a full month and average it out. Perhaps the 0.417 number would be closer to 0.6 in the end.