AFAIK, the primary determinants of timing are rpm as you said, the coolant temperature, and the fuel quantity to be injected, which is determined by the air mass, which depends on altitude, humidity, ambient temperature, etc.
I never said, nor really thought, that the impact of a WAI on FE was primarily or in minor part due to reduced throttling losses from lesser air density. What I've said in the past though, it that a hot engine is an efficient engine. It's mostly the temperature, not the density.
I'll repost again here the references I've posted a couple times already so people can make up their own minds. It's not me saying it, it's research papers and mechanical engineering textbooks.
- Pre-heated intake mixture at low rotational speed improves combustion. (Chiu and Horng, 1992)
- Specific fuel consumption varies inversely proportional to the square root of the suction air temperature (Nakajima et al. 1969).
- Higher ambient temperature is found to increase the flame speed, the combustion reaction rate, the uniformity of the fuel-air mixture and reduce the heat transfer rate though the cylinder walls (Pulkrabek, 1997).
- For lower temperatures, only a small part of the injected fuel is vaporized, causing nonhomogeneity. As a result, lower flame speeds, higher unburned mixture, higher hydrocarbons and carbon monoxide emissions, and loss of power are observed (Pulkrabek, 1997; Heywood, 1988).
References:
Chiu, C.P., and Horng, R.F., 1992, “Effects of Intake Air Temperature and
Residual Gas Concentration on Cycle-to-Cycle Combustion Variation in a
Two-Stroke Cycle S.I. Engine Equipped with an Air – Assisted Fuel Injection
System”, JSME International Journal, Vol. 37, N.4, pp. 957-965.
Nakajima, K., Shinoda, K., and Onoda, K., 1969, “Experiments on Effects
of Atmospheric Conditions on the Performance of an Automotive Gasoline
Engine”, SAE Transactions, SAE 690166, pp. 745-766.
Pulkrabek, W.W., 1997, “Engineering Fundamentals of the Internal
Combustion Engine”, Prentice Hall, Inc.
Heywood, J.B., 1989, “Internal Combustion Engine Fundamentals”,
McGraw-Hill Book Co.
Finally, I'll post the most useful freely available paper I could find:
Changes of Low Load Engine Parameters by Temperature of Mixture
In this research, they were able to reduce BSFC at low RPM and low load by 7% at 195 F vs 100 F. Also, indicating that density reduced throttling losses have not much to do with it, they measured almost the same throtte position at those two temperatures.