In the course of blind reviewing LR questions, I gained a few simple insights that I thought might help some people. In the interests of facilitating improvement on LR, maybe this can get people to chime in on some of their methods.
First of all, I've seen it said many times, but I now think that the advice to write out by hand every question you're reviewing can't be overstated enough. I found that by writing out and explaining exactly what I read in the stimulus along with writing out reasons for why each answer choice is either right or wrong, I was able to expose so many gaps in my understanding and tacit assumptions that would have completely gone unnoticed had I just scribbled a quick diagram or something to that effect while reviewing. By writing things out, even if by playing around with things, paraphrasing, or putting an idea in a different light, I was able to express everything I knew, which led me to identify everything I didn't know. (I think the old adage of knowing what you know and knowing what you don't know applies well here). Writing it out also doesn't have to be formal or strict, but as long as you are able to hold your ideas accountable by enumerating them, you'll have a good point from which to improve yourself.
Second, make sure that you eliminate answer choices independently to fully understand what's wrong with them. I think a tendency of reviewing without writing things out is that we're more liable to blithely skim from one choice to the next, not holding ourselves wholly accountable and not giving each answer choice a fair hearing. You can't understand why something is wrong if you can't eliminate it for its own sake. If you spell out your reasoning for each choice before either choosing it or eliminating it, you'll better understand why you chose it or not.
Third, make sure to read the stimulus naturally first, with some light marks if necessary, and then come back and diagram. Keep in mind I recommend doing this during review, not necessarily timed PTs. I found that If I'm too ready to write something out, I might miss making some connection and seeing the larger point. Get a grasp of the idea or structure of the stimulus before you assign a specific methodology or approach. Maybe this is more applicable to harder stimuli as opposed to easier ones with a quick and simple setup and with time you'll be so good that you know exactly where things are going as you're reading and your diagramming is right at the heels of your thoughts, but I think it will take some time to gain that level of intuition.
Fourth, be sincere with yourself. Don't censor your thoughts too much in the course of writing down your reasoning and be intimately aware of the initial impressions you have after reading something. Those initial impressions may reveal something about your intuition, and if you want to improve your intuition, you should first expose them by writing them out. Improvement is made easier when you know exactly what you have to improve on, and this starts by being honest about yourself and your reasoning behind choosing an answer choice.
Finally, do not move on until you feel absolutely certain about an answer choice. I think the temptation of watching the explanation for the question by JY is just an excuse for you not doing your job properly. The videos should be a complement to what you already learned yourself by pouring in all that time reviewing the question.
I'll be the voice of dissent and say go for it on September. There'd be too much pressure on one single take in December since December is really your last shot for the cycle. I don't understand what would be wrong in taking September and then improving your score in December. Why would you be considered any lesser to a school for improving a score? In fact if you're reasonably ready, I'd say September is very valuable because you can try your best and not have the pressure or consequences of it being all or nothing like it would be for December.