Looking forward to it but also sad that it is "only" 12 puzzles, but I completely respect Eric's decision to scale it back.
I've got 500 stars (i.e. I've completed every day of all 10 previous years) but not always on the day the puzzles were available, probably 430/500 on the day. (I should say I find the vast majority of AoC relatively easy as I've got a strong grounding in both Maths and Comp Sci.)
First of all I only found out about AoC in 2017 and so I did 2015 and 2016 retrospectively.
Secondly I can keep up with the time commitments required up until about the 22nd-24th (which is when I usually stop working for Christmas). From then time with my wife/kids takes precedence. I'll usually wrap up the last bits sometime from the 27th onwards.
I've never concerned myself with the pointy end of the leaderboards due to timezones as the new puzzles appear at 5am local time for me and I've no desire to be awake at that time if I can avoid it, certainly not for 25 days straight. I expect that's true of a large percentage of people participating in AoC too.
My simple aim every day is that my rank for solving part 2 of a day is considerably lower than my rank for solving part 1.
(To be clear, even if I was up and firing at 5am my time every day I doubt I could consistently get a top 100 rank. I've got ten or so 300-1000 ranks by starting ~2 hours later but that's about it. Props to the people who can consistently appear in the top 100. I also start most days from scratch whilst many people competing for the top 100 have lots of pre-written code to parse things or perform the common algorithms.)
I also use the puzzles to keep me on my toes in terms of programming and I've completed every day in one of Perl, C or Go and I've gone back and produced solutions in all 3 of those for most days. Plus some random days can be done easily on the command-line piping things through awk, sed, sort, grep, and the like.
The point of AoC is that everyone is free to take whatever they want from it.
Some use it to learn a new programming language. Some use it to learn their first language and only get a few days into it. Some use it to make videos to help others on how to program in a specific language. Some use it to learn how/when to use structures like arrays, hashes/maps, red-black trees, etc, and then how/when to use classic Comp Sci algorithms like A* or SAT solvers, Djikstra's, etc all the way to some random esoteric things like Andrew's monotone chain convex hull algorithm for calculating the perimeter of a convex hull. There are also the mathsy type problems often involving Chinese Remainder Theorem and/or some variation of finite fields.
My main goal is to come up with code that is easy to follow and performs well as a general solution rather than overly specific to my individual input. I've also solved most years with a sub 1 second total runtime (per year, so each day averages less than 40msec runtime).
Anyway, roll on tomorrow. I'll get to the day 1 problem once I've got my kid up and out the door to go to school as that's my immediate priority.