You guys sure are trying hard to say that you don't are about a +/- 5db variance in the lower octaves, which if so is just fine by me. Just don't tell me I'm stupid for caring. Some pretty incredible ego's here!
TLS Guy makes a very persuasive argument in post number 39. T/S specs change with heat, you have shown that and so has Avaserfi, and some drivers may change audibly (post 26), but the most important question remains; how do you design for it?
It's been shown how much three different drivers change after thirty minuets of operation, but how much would they change in five minutes? Ten? An hour? What about a higher input level? Or a lower one? Surely the results from any one of those tests would be different from another. If the half hour 'normal level' test shows a -0.4dB drop in bass response and the one hour 'high level' test shows a -3dB drop, which occasion are you to design for? To be optimum for rocking out or for mellowing out?
To paraphrase a few people, three samples do not a statistically relevant example make. In the end, what I get out of this thread is that
if you're very detail oriented, you might want to know how the T/S specs of a driver you're considering change with heat; if only to decide if you can live with the amount of change, or that a driver that changes less needs to be found.