Dangerous talking in generalities, but I often end up with something of a "W" shaped EQ curve, overlaid on a very general and gentle overall slope or upward tilt across the entire spectrum "/". Of course the particulars are specific to each recording, and critical in their adjustment.
First, try adjusting a graphic EQ so that all the sliders form a flat-line with and gentle slope across the entire spectrum, adjusting the angle of the slope until you start to get things sounding better, keeping the overall line straight.
Then fine tune from there.. boost each slider until you can identify the stuff living in that frequency range, then pull it back down to where it sounds right. Next adjust the surrounding sliders the same way, then go back and retune the first one again, as they'll all interact with each other. Try to make smooth curves without radical changes from each slider to slider along the line, rather than more radical jumps between sliders.
"Mud" often pools somewhere between 200-600Hz as mentioned by the others, so a scoop around there can help tame it (forming the left downward point of the 'W'). Bringing up the midrange in the 1-4khz range helps bring out clarity and presence, restoring necessary energy above the mud range (the center peak of the 'W'). I'll often boost below the mud range as to restore phat bass energy, once clean of the lower-mid bloat (the left-most upward arm of the 'W'). Boost the high treble "air" to suit the level of the mids (the right-most upward arm of the 'W'). Cut a bit between the mids and those highs (somewhere around 5-12kHz) to reduce excessive "glare or brightness" which probably has resulted from the mid and high boosts. Small changes here have a radical effect on the overall sound. Take your time and tweek carefully, aiming for dialing back to smoother overall curves. Go back and fine-tune the peaks and curve transitions of the vaguely 'W' shaped EQ curve, over and over until it sounds best. When you'r happy with it, "squash" the 'W' a bit and see if that's better and smoother. Refer regularly while fine tuning to the sound of the uncorrected, not EQ'd recording, as well as a reference recording of something similar you really like the sound of - maybe one of your other live recordings with the same rig which came out great, maybe someone else's you think sounds especially good and "real". Leave it and come back later and see if you still like it. Re-adjust and leave it again. Fine-tune until it sounds as good as you can make it. Save the curve, start over and try again. Then compare your saved curves and see how close they are to each other. A lot can be learned about your own ear and your own over-compensations from the comparison of those curves.
[edit] Don't compensate for any nonlinearities of your playback system by the EQ correction made to your recordings. Get your playback system sounding balanced first, or become enough highly aware of it's specific problems that you can learn to "listen around" those things. Listen to the recording with your adjustments on a few other systems to make sure what you are doing is globally applicable and not just locally appropriate to the particular response of your stereo in your room.