Good question. I will get the snowball rolling. Cool enough to have long fatigue life, and hot enough to still operate combustion efficiently. Higher surface temps in the cylinder will lower heat rejection and allow better hot gas expansion and fuel efficiency. That is probably not as big a factor, or "knob" as leaning and A/F.
So, since the tensile strength (drop) gets steeper around 300F and steeper yet at 400F. Somewhere in the 300-400F is good, but closer to 300F is probably better. This theory is just based on head material strength, not "MPG" or valve guide cooling and lubrication, or the like. That, in reality, could easily change the "best" range.
If the temps could be kept in the 250F range, all the time, then compression ratio, timing etc could be reoptimized to take advantage of the extended detonation range, and material strengths. Since we still have to operate hot in some weather, those options are not viable.