Over the past 10 days, the sudoku.com app has been running a tournament. Every puzzle you finish, in any category, contributed to your point total — more points for more difficult puzzles.
The weird thing is, users of the app don’t register any kind of nickname, but the leaderboard listed them anyway — randomly-generated names of the format [Adjective] [Animal]. My rival for the top spot was Puzzled Wolf. I had a strong lead when I went to sleep last night, Puzzled Wolf managed to just squeak ahead of me, but a few minutes this morning while getting ready for work put me in the lead again. I won, but… I have no idea what my own name is. I’m quite curious about it.
(It occurs to me, I probably could have looked at the leaderboard just by going to the sudoku.com website… but now that the event’s over, there’s no way to see it! Argh.)
I had the thought that maybe the whole thing is completely fake, with fake scores updated by an algorithm based on my own performance, meant to provide the right amount of challenge and instill some competitive spirit. But I’m guessing no, just some overlooked UI features.
Winning was such a lackluster occasion that I’m just not going to bother with any future tournaments. Pictured here is the only evidence, acknowledgement and reward for the effort.
So how did I get on this Sudoku thing anyway? I used to use a different app, HappyDevs Sudoku. It was both free and ad-free and didn’t have a horrible interface. But it created some brutally hard puzzles that, as far as I could tell, required guessing. If I guess wrong, I could just undo and try again, an indefinite number of times. That kind of takes the fun out of it though.
Since I’d heard that you should never have to guess in Sudoku, a few weeks ago I started looking into the technique. There are a bunch of patterns you can search for to eliminate candidates, some of which I’d reasoned out for myself but several of which were new. These… kind of helped sometimes but not always. There are some advanced Sudoku solvers which not only find a solution but also the patterns required to find them logically, and rate them on how tedious they are. Even the second-hardest difficulty level in that app was capable of creating some incredibly tedious puzzles at times. Apparently there are techniques you should follow when creating a good puzzle, not just a valid puzzle, and that app wasn’t.
Sudoku.com promised to have “smart hints” that coached you in techniques, and I guess I also assumed it would generate better puzzles. Those hints, it turns out, are dumb as dirt and sometimes wrong. The puzzles are generally better constrained — anything lower than Evil difficulty is almost trivial to solve with basic techniques. But Evil doesn’t usually fall to the kind of intermediate techniques I wanted to practice, and I often have to work through in a brute force way after a while.
At this point I think I might have had my fill of Sudoku.
The Roli stuff arrived. Obviously the previous owner was a heavy smoker, and I’m pretty sensitive to that so the packaging got thrown out right quick. The surfaces of the devices are silicone rubber, which absorbs scents; it wasn’t that bad though and you can clean them with a baby wipe (no alcohol, no bleach, preferably unscented).
The Seaboard Block feels great. It requires a little more force than I would prefer for maximum velocity and pressure, but there are adjustments, and it’s smooth and consistent. Playing softsynths using real MPE (not just poly aftertouch) is a joy — especially the Noise Engineering Vereor plugins and Aalto (once I realized there’s a setting to accept MPE rather than regular MIDI).
The Lightpad Blocks on the other hand… they are clever in concept, flexible controllers where you can write software that determines their behavior, creating X/Y pads, sliders, buttons that play notes or toggle CC values, sequencer, even primitive games. But it’s not quite there in execution. A short list of shortcomings:
- This is not Roli’s fault, but MIDI-over-Bluetooth doesn’t work in Windows — so you need a USB connection. This is fine for the Seaboard Block, but kind of a bummer for these smaller units. However, I’ve heard that the latency over Bluetooth is pretty bad anyway.
- They’re supposed to share power and MIDI with other Blocks via their magnetic link (you can snap them to the front, back or either side of other Blocks). In practice, that link isn’t stable. So they might work for a while, then the link breaks, the Lightpads start trying to pair over Bluetooth and never re-establish their connection to whichever unit has the USB connection. So in practice, every device needs its own USB cable. Which means if I wanted to use them all at once, I’d need another USB hub.
- Even over USB, they feel a bit sluggish and not as smooth as they should be. You can’t really rely on it for precision either in time or position.
- On the sliders I find it’s very hard to hit exactly 0 — it needs some kind of negative offset to compensate.