In this final test of the night, I shall not sneak but still sleep for 8 hours. I expect attack frequency to be higher at this point, due to hampered mobility and not having the 'sneaking' status. However, the previous experiments suggest otherwise.
Speeds here as the same as listed in above experiements.
CONCLUSION
3/10 of the time Lisid was safe to sleep the night away. 1/10 of the time a horned boogey attacked, and 2/10 bandits attacked, totaling 3/10 together. 4/10 of the time generic boogeys attacked.
From these 4 rather simple experiements, I would venture to say that sneaking is safer than not sneaking, whether you are also sleeping or not. Sleeping is also somewhat dangereous to do, and opens onesself up to a greater range of ambushes. Sleeping in the day and traveling manually (not fast traveling) at night while sneaking should prove the safest thing of all to do, if these are to be believed.
Overall, you have between a 30% - 40% chance of avoiding trouble if you stop at night. If you stop, you should expect 'Fun' trouble ~1/3 of the time, and simple generic trouble the final third.
Speed should be tested still. I suppose another day I'll strip Lisid naked again and let test for speed variables. The safest place to sleep seems to be somewhat unclear at the moment, but I think as hidden as possible and as enclosed as possible would be the best. Maybe in a ditch somewhere or againt the river or a grove of trees. It is certainly advantageous for defense, at least.
Many flaws were present in the design of this experiment, especially since I didn't go too far out of my way to ensure a stable enviornment:
-Sample size was relatively small.
-Speed was not tested.
-Sleep until dawn was not tested
-I did not start this at dusk, but rather during a night.
-I moved around slightly (1-5 tiles) during each test in every experiement.
-Different worlds with different conditions were not tested.
-Different adventurers were not tested, nor where any special status afflicitons.
-Different biomes were not tested.
-Skills were not taken into account.
-Different nights were not sampled.
-[insert flaw here]
However, there seems to be a certain coorelation between the experiements. Perhaps further tests will clarify this before Toady does something rash like limit boogeymens or gives them a pop-cap. . .