Thinking more about the noise of the Mixpre-6 and not the dynamic range, I've been trying to work through what it means to record at close to 0dbFS (by actively riding gain levels presumably, and thus the need for configuring for gain knobs and not fader knobs) vs just setting and forgetting and maybe getting peaks of -12dbFS, or -24dbFS. Trying to understand how it affects your recordings after you normalize in post to bring the -24dbFS peaks to 0dbFS.
You need to think about the overall noise in the system, which would include the internal noise of the MixPre-6 and the noise of the microphones you're using (among other sources of noise). The specs of the MP-6 put the noise at -128dBu (or -130dBV). Noise of a gain stages (op amps, whatever) are largely independent of gain levels. So the noise at -128dBu spec'd by SD at 76db of gain, is probably something close to -128dBu at only 20db of gain, or whatever.
The MP-6 has noise of -128dBu and a max input level of 14dBu. A very low noise mic like the Shure KSM-141 has a signal to noise level of 80db. For comparison, the Schoeps cmc6/mk4 has a S/N level of 79db and the DPAs have a S/N level of 74db. With a somewhat typical mic sensitivity of 10mA/Pa (-38dBu at 94db SPL), you'd have noise levels from the mic of about -92dBu at concert SPLs (~120db say). This mic noise level gets amplified whether you use the mic pre of the MP-6 to amplify it or whether you amplify it in post by normalizing.
In this example, say you will need about 26db of gain to reach 0dbFS. If you provide 26db of gain at the MP-6, the noise of the KSM-141 mic once amplified is about -66dBU, and the recorder has a noise floor of -128dBu (so the mic noise swamps the MP-6 recorder noise). If you provide no gain, you get peaks of -26dbFS, and you need to add 26db of digital gain in post to get to 0dbFS. Your mic noise once amplified in post is once again -66dBu, but the noise of the MP-6 (-128dBu) once normalized to get to 0dbFS is now -102dBU. So the noise of the MP-6 is much higher by recording lower and adding gain in post, but it doesn't matter and you won't hear it, since the mic noise is at -66dBu and still swamps the noise of the MP-6.
Not the case if you do the same thing with the Tascam DR-70d. It's specs put it at -92dBu of noise (and a max mic input of 0dBu). With the same Shure mic with an 80db S/N ratio and the same concert SPL, if you add gain at the recorder to get to 0dbFS, the recorder noise floor will be below the mic noise floor. But record at -12dbFS or less, once you add digital gain in post, the noise of the DR70d once amplified in post is now louder than the mic noise of the Shure mic.
So with the MP-6 and the noise level of its preamps, you can add 25-30db of gain (or more) in post and the recorder noise floor is still below the noise floor of the mics. Not the same for the DR70d, record at peaks of -12dbFS or less and normalize in post, and the noise of the recorder becomes worse than the noise of the mics.
Of course, there's also the noise at the venue (HVAC, people shuffling feet, whatever), which is probably easily at a SPL of 55-60db. This noise will swamp both the mic noise and the recorder noise, regardless of which recorder you use and regardless of whether you record at -18dbFS or 0dbFS.
At any rate, there really is no need with the MP-6 to try to record at say -2dbFS is you're planning on boosting in post anyway. Add 2db in post, add 20db in post, really doesn't make a difference.