I think you make a good point but in terms of what we do I think bass roll-off at the mic is too much of a gamble: we all know how much the sound can differ from night to night - for the same band in different rooms, different bands in the same venue and even the same band in the same venue...
If you can do a proper test recording (ie of that particular band in that particular room on that particular night) and decide that there really is too much bass then fair enough but I would rather take the risk of recording too much bass and have to tweak it in post than run the risk of coming home with an unacceptably thin recording that no amount of post-processing can properly rectify.
I understand the benefits that you are describing but I think that they are probably of less import than the potential drawbacks. Plus you have greater flexibility when applying EQ in post, both in terms of amount of cut and the frequencies affected.
Totally agree that it's too often a gamble, but only because the science of how it works with particular recorders has been completely obscured.
Roll-off is not meant to be tweaked in such a finite fashion as 'venue to venue.' Certain microphones will benefit from a threshold amount of roll-off 99.9% of the time, but you have to carefully evaluate what that threshold is. Once you know, you will never come home with an 'unacceptably thin recording.' In fact, you might still wish to remove a tad more bass as your ears and playback system dictate, but you're certainly better off than had no roll-off been used at all.
The goal is not to get bass response perfect before it hits the recorder, but rather to reduce a portion of the low frequency spectrum that is rarely useful or contains unwanted artifacts (hum, vibrations, etc.). For some microphones, a small amount of roll-off is beneficial regardless of the source music or environment.