Evidently even with the crap reference cooler, there is a temp range where the Ti will outperform the 290x and vice versus. If the ambient room temp is below X, the 290x won't throttle back and if the ambient room temp is too high the Ti will absolutely stop dead in tracks and the 290X will slow way down but continue to run.
OK so at 78º in your test lab with an open system the Ti is the fastest card. Do you have the heat running as the high outside air temps in Bakersfield are less than 78º if you opened the window. In Denver today the outside temp is 65º so I guess that the Ti may not be the fastest GPU if tested in Denver with the windows open today. The chip fab and testing areas of the semiconductor plant where I worked had a constant temp of I believe 67º with a positive air pressure. For fab workers gowned up that was a comfortable room temp while in the offices it felt cold.
While this may seem snarky, my point is that 78º is a very high air temp in which to do comparison testing, especially when all the AMD cpus and GPUs are known to run at higher temps. The heat in my Denver house is set to 69º day/60º night and my air to 72º. In the evenings when gaming I'll open my study window when the temps fall below the 72º air conditioner setting. 72º seems like a fairer temp for comparison testing.
FYI I have a Windforce 680 in my computer set at 1250/6600 so I'm not an AMD fanboy.
"The average air conditioning temperature that central air should blow at would be 70 degrees. This is a neutral degree while most of the year, the outside temperature is not exceedingly 20 degrees more or less than 70. Twenty degrees is an important number to remember. No matter how good the system, the house can not be cooled or heated to well below or above 20 degrees of the outside temperature. It will just work harder trying, and cost lots of money."