I would have responded earlier, but I wanted to actually implement something you suggested: different walls. Alternative wall sprites, which don't occlude other tiles so much, are now live and can be adjusted in the settings.
Re: analytics, the only serious plans I had were to use the daily level histograms to adjust difficulty. The idea of taking some levels and releasing them as a standalone game is tempting, but I wonder if doing this type of puzzle over and over again might get tedious? That's one of the reasons I thought it would work better as a daily game. Let's see how it's doing in a few months.
I love the mechanic ideas. I think there are two big constraints on what kind of cool new features/gimmicks can be implemented though. First: if this is going to be a daily game, the new mechanics have to be intuitive enough to where somebody could figure them out their first time playing. I like the idea of cherries being misleading, and it's a fun troll-ish idea for a single player game, but it would be a mean trick if it were someone's first daily. (Then again, there's someone who's first Wordle game was probably MYRRH.) The other constraint is I have a solver that can guarantee the optimal solution is actually optimal. Some game mechanics might make this a lot harder, or even impossible.
Great game. I've been having fun with it every day, but today there is no puzzle! Be careful about this. I intuit that not having a puzzle on a day is enough to lose some players, with the number only increasing the more days you don't have puzzles. Not sure how you're doing it now, but you should easily be able to generate thousands of levels and have them all ready in the page like wordle.
All the daily levels are built by hand. I struggled to come up with a good random level generator. You can see my feeble attempts in the Edit page (via the hamburger menu) by giving the dice button a few sad clicks.
I did originally try to measure the difficulty computationally by running the solver and timing it, but it didn't really line up with what humans would find difficult. Now I'm just eyeballing it.
This article does such a bad job of criticizing an already-unpopular theory that I found myself defending it. "The evidence is clear," "studies have repeatedly shown," and "analysis showed," would be a lot more convincing if the evidence, studies, and analysis were actually explained. Not to mention that focusing on income inequality doesn't disprove the idea of "a rising tide lifts all ships." It's possible the absolute wealth and standard of living of the poorest might increase while inequality also increases.
Yes, I felt the same way. It felt like Allen Iverson's "Practice' press conference. Just because you repeat your opinion over and over doesn't make for an argument.
Alas, it still fails at my favorite music theory question. I've yet to see a chatbot get this right, even though it isn't a trick question at all.
I ask:
"Consider a tenor ukulele in standard tuning. If we tune it down by a half step, how might we then finger a Gmaj7?"
It initially reasons correctly that this must be G-B-D-F# and it understands what the resulting strings will be tuned to if we drop them a half step. It just can't work out a valid fingering, and ends up getting confused and suggesting a nonsense chord.
I feel like that’s more of a ukulele question than a music theory question, but I do wonder if it gets it wrong for the same reason diffusion models struggle with hands.
Music theory is a great topic to test though, since GPT-4 struggles a little with it.
