Anyway, onto the puzzles I submitted that I have one version of!
I started with a Kurodoko with what I hoped was an interesting logical theme. From my own timings I expected to hit about 1:10 with this which was close enough, but ended up at 1:50, and multiple people mentioning they got stuck after a couple opening steps. Unfortunate.
Since I thought that might be a bit long, I originally tried to aim smaller and stick to that logical theme idea with a 9x9 Tapa-Like Loop. 9x9 quickly didn't work for the theme I was going for, but loop puzzles can be speedy, so I was fine with going 11x11 if I thought the solve was straightforward. This one ended up being the closest to a minute timing-wise, but was still an overshoot.
My last puzzle and again an attempt to go easier was a 62 themed Rail Pool. This was the puzzle I was the most off-estimate for, since it's a less common type and thus the median solver has to take longer to find more common steps. I don't think it's a coincidence some of the closest puzzles were extremely common types, but again I don't regret trying something a bit more out there. Perhaps I should have gone even further in the obscure direction and made a 7x7 or something? Somehow, this is the first time I actually made a Rail Pool, too.
Interesting post. Re the 1 minute median target time, I’d doubt there’s more than 50 people in the world capable of doing these in a minute. My times were nowhere near:
ReplyDeleteKurodoku - 6:05
Tapa like loop - 2:19
Rail pool - 4:57
Admittedly I was solving at leisure, trying to enjoy logic than solve in the fastest time, though I’d also say my philosophy whilst setting is to cater for the solver who doesn’t care about time far more than those who do (and will therefore look for shortcuts at every opportunity).
Rail pool feels like a very unlikely candidate for 1 minute at this size - there’s just too much to do drawing the loop before factoring in things like making logical deductions. Kurodoku and tapa like loop (there’s a type sorely in need of a better name!) give you a better chance, despite being less familiar than a slitherlink or nurikabe or whatever.
I guess at the heart of things, 1 minute puzzles - and I get it’s largely audience dependent - are fundamentally easy puzzles. To repeat similar thoughts I’ve shared with gmpuzzles and others, not many people are actually very good at making good easy puzzles. I’d even argue for a certain type of author, it’s probably harder to make an “easy” puzzle than it is to make a “hard” one. I guess the things I bear in mind with what make an easy puzzle are as follows:
1) the puzzles you make are basically always harder than you think they are
2) if you want a slightly cute or novel break-in, you need to be incredibly clear about signposting it (which isn’t always possible)
3) after the break in, if you want a cute/novel trick in an easy puzzle then it cannot break the general solving flow, and in particular it has to “feel” similar to the rest of the solve
4) building on point 3, the flow to the puzzle is critical. If the intended solving logic is disjointed, then inevitably you’ll get solvers slowing down, stopping, and going through periods of scanning the grid for the next step. If a try he puzzles in the last c. 5 years have been very creative and inventive, and often end up being quite difficult - but they often don’t have a good sense of flow. I think this probably reflects different ways in which people go about solving and setting, particular with regards to using computer assistance.
5) there is always a difficulty premium on less familiar types to be taken into account
With regards to hitting time, the other angle is bringing the grid sizes down - you can still do interesting things at e.g. 8x8 or even 6x6 in some cases
Anyhow, not sure how useful my thoughts are, but in any case thank you for a thought provoking post, and thanks to anyone reading this for tolerating my meandering train of thought.
You're correct about how many could do it in under a minute - as mentioned, I had misjudged the overlap between that group, and showcase voters. I had actually considered all 6 of your points - the 6th being puzzle size - with the audience in mind... but had the wrong audience in mind, which led to making the wrong decisions in response to those factors - and the wrong scoring adjustment in mind, which led to not prioritizing the timing component as heavily. But in this case, the worst case scenario of missing the target still led to some nice puzzles I'm quite happy with.
DeleteAgreed these are nice puzzles, I just don’t think I’d call any of them easy, which is roughly how I was interpreting the 1 minute brief at first glance and thought would be interesting. Its fair to say I tried one or two of those showcase contests myself a while back, but it seemed what I was interested in doing (or not doing) was not what anyone else wanted to vote for, so I haven’t really checked in with them for a while.
Delete