a note on "making progress," "regression," and redo-to-undo:
It's important to have a model of progress that accounts for, nonarbitrarily, "anything could be happening at any time." Say things are sort of getting "easy and smooth" (or whatever), and it's been that way for a day or weeks, and then the next day things are choppy or distract-y, or whatever. That doesn't mean things have "regressed" or that progress isn't being made. Anything that the system has experienced in the past, thoughts, brain fog, depression, fear, anything--has to come up again sometime, in the course of meditation, because of the "redo-to-undo" "principle." Sometimes it'll be liminal, barely there, to be sure. Sometimes it'll be "in a "cradle/container" of "equanimity."" And, sometimes, something intense can/will come up that sort of takes over everything, like an immersive flashback or, less obviously, an old outlook or way of being, etc. It can also be very subtle, piecemeal things like "habits of mind." This can happen a lot, over thousands of hours. Additionally, sometimes very big intense stuff can happen, late in the game, because it took that long for there to be system-wide safety for that thing to to come up. (This could be the case, five to twenty times, or more, spaced out several months apart, as well as other possible patterns.) "The system makes it safe to look, and then looks, safe to have happen, then it happens..."
So, in any case, one shouldn't infer progress by what's happening in any particular session--"terrible" sessions may be extremely constructive sessions; sessions that don't feel like sessions at all (however conceived) may be extremely constructive sessions, and stuff like that could go on for months! But, actually, all things being equal, the right things had been happening, the whole time.
One sort of, as best they can, has to take into account the whole global history and context, in terms of global wayfinding, to get better and better at how to relate to and participate in any particular thing that's happening at any particular time, including being ok with being swept away, when that's the right thing to be doing, or anything in between, or nuanced variations of any of this, and so on.
*
It can be really discouraging when progress seems state dependent (“if only I were in that other state!“), but taking correct actions, as best one can, in any state, is the thing that metabolizes, over a long time, whatever state one happens to be in (and part of that is fully accepting being in the state, if/eventually/when it’s safe to do so).
Like, the state IS the very stuff of practice, not a thing in the background that conditions practice.
*
I sometimes say something like, “after hundreds, thousands, or tens of thousands of micro-experiments, little bets, something, as best you can, eventually you hit a wall, then you travel along the wall until you end up in a corner, and then in the corner there are no degrees of freedom, and then there’s only one way out, so you know what to do, and then you spontaneously do that, and this repeats.”
*
Finally, a sort of corollary to all this is that, since sometimes (heuristic or provisional) understanding precedes karmic burnoff, you might find yourself compulsively doing "miles and miles" of "knowingly the wrong thing," and that's ok! Burnoff/burn-off/metabolization/integration is happening. Miles and miles. Sometimes wrong turns and backtracking can be avoided, but, modulo that, there's no shortcuts, no skipping, no corner-cutting: ultimately the entire territory needs to be walked, and the territory spontaneously does what it does, until it doesn't. So if you're doing something "bad" (or multiple things) for like five hundred hours, each, that's redo-to-undo, that's just what needs to happen. (Wrong turns and backtracking are included in the "10,000 hours" thing.)
(And part of that will be doing meta-protocol-esque things to decide whether/which/where those miles and miles are entrenchment vs burn-off. That's ok, too, that's part of it. A seed of something in the spirit of the meta protocol, a seed of, hmm is this right, is enough for the at least 50.00000000001% [sic] burn-off versus entrenchment that will ultimately, cumulatively, do the thing.)
So miles and miles of "I'm doing it wrong" might exactly be miles and miles of precisely optimal progress.