US Puzzle Championship Thread
We’ll have a little more discussion in the comments here later after the dust settles about the puzzles, but we hope everyone enjoyed today’s US Puzzle Championship.
Our puzzlemasters contributed several puzzles to the competition including:
Crosslink by Grant Fikes
Double Minesweeper by Serkan Yürekli
Nurikabe Path by Thomas Snyder
Star Search by Thomas Snyder
Tapa by Serkan Yürekli
Tapa View by Prasanna Seshadri
Wind Shield by Serkan Yürekli
Wordmark by Serkan Yürekli
7(+6 Differences) Puzzles solved, 91 points total.
*SPOILERS*
Cave and Masyu were okay. Nothing too out of the ordinary.
Wordstar was interesting, and I got it based on knowing some squares had to have words, knowing there were no 5 letter words , and by limiting the spots PUNCTUALITY could go in.
Crosslink was alright, lots of places where you had to be careful about closing the loop. I ended up wasting a bit of time on the end, where I had two legitimate-looking possibilities and I had to trace the loop out to figure out which one didn’t use the whole loop.
Match Series…that second sequence…*shakes fist*
Star Search was my other big time sink on the test; big and kind of tricky. Thank you for not throwing tricks with the words on that one. I could have probably gotten a quicker start had I caught the step in the top row and used that to start eliminating squares off the bat, rather than trying to find where stars could go based on the words. Both could be legitimate solving methods I suppose, but I imagine tackling the Star Battle first would allow for a quicker solve.
The Tapa is my finest moment on this test. I caught the break-in fast and worked my way through pretty quickly to get the 20 points.
Thank you to the contributors for the puzzles, and congrats to the top solvers.
10 puzzles solved and 7 differences found for a total of 157. A personal best…we’ll see if it’s enough to break my goal of Top 25, which has been my personal white whale, having come oh so close in the past.
I don’t think I’ve ever broken a puzzle as many times as I broke the Nurikabe, but managed to enter my solution 24 (!) seconds before time was up.
Ah crap…I missed a word in the word search. Make that 137 then. Damnit damnit damnit.
Well, you can have my spot in the Top 25, Scott. This was my worst year in a while. I got a lot of the early easy points, but mostly drew blanks on the later 20 pointers. My total is going to be a little under 100, I suspect (93?), so pretty mediocre. A majority of those points were in the first hour and change, I think. The “one or two” crossword was surprisingly easy in the later part, so I at least got that one quickly.
I’ll try to avoid spoilers, but no promises.
8 puzzles solved–still kicking myself for not picking up on the gimmick in Match Series, part 2, until after the test was over. Wasted 10-15 minutes for 0 points there due to no partial credit option.
Battleships was super easy this year. I feel like it should’ve only been 5 points.
Cave and Masyu were both straightforward.
Holiday BBQ took a little longer for me than the find the difference puzzle usually does, but I did get all 10.
Crosslink was a fun challenge. I had to slow down a bit to be more careful after an error on my first pass. Very clever puzzle!
Inside/Outside Fences was another good challenge. My paper is covered with tons of shorthand symbols to keep track of inside vs. outside and where walls couldn’t go. My approach was probably a bit clunky, but I got through the puzzle in pretty good time, so I was pleased with it.
The word puzzles in the USPC are generally my bread and butter, and Star Search was no exception. That was probably my best time-to-points ratio of the whole test. As Giovanni said above, it probably would be faster to star hunt before finishing finding all the words, but I was making such good time that I figured I’d rather finish the search than risk losing time on a futile star search. Looking over it again, that was probably a suboptimal decision.
One or Two was all about finding the right place to start. The second corner I tried was the winner and fell into place very easily. That wound up limiting my options enough to branch out to the other corners in short order.
That’s all the puzzles I solved. I spent time on a few others without finishing them. I was lax in practicing this year and it showed. I got halfway through Nurikabe and a little less than that on Tapa View, gave up on Sudoku too early, and never even got to check out a few of the others (Wordstar and Crosswind both look solvable in reasonable time). If I’d been more careful with my time and spent more time prepping, I think I would’ve solved a couple more puzzles. Next year!
Battleship being 10 pts and not 5 kind of threw me off — I took a much more methodical (i.e. slow) approach than I would normally do for an easy Battleship.
I did expect scores to be down overall this year. My initial response on test-solving was “I think I’ve forgotten how to solve puzzles” and maybe that is connected partly to how I was testing past midnight because I don’t have much free time anymore, but it also was tied to the content.
This year’s test definitely had a lot of high variance puzzles (particularly the Altay puzzles, and some others like the Nurikabe). But some of these high variance puzzles were also among my favorites. I guess I like the “Ahas” I see, and dislike the “Ahas” that take 10+ minutes. So I liked the Matchsticks but have different feelings on the Meta Tapa. I was also quite surprised by the Sudoku’s presence given the difficulty I had solving it (Not logically! The path is narrow and I didn’t catch it.).
But as a whole, despite the higher difficulty, I thought it was a really fun mix of a variety of puzzle styles; for instance the word puzzles were quite good.
For my contributions, I wanted to play again with the Nurikabe Path variation that I’d worked on for a Puzzazz puzzle contribution earlier this year. I settled on a crazy clock kind of theme with distinct 1-12 clues and a good mix of logic and intuition as options for the solver.
I was again tasked with a word search and counting puzzle and only got one of those done. With the word search I was motivated to play with word path change ideas and had a few logic puzzles that I thought to connect with bending paths. I eventually settled on the Star Battle tie-in. While the puzzle wouldn’t be solvable by just using Star Battle logic, the logic could help deduce locations for the next words. There is a pretty blatant “Milky Way” theme to this puzzle. And yes, there is a black hole in the center of the galaxy.
Nurikabe Path was definitely my favorite of the batch — nice combination of intuition and deductive logic that I don’t see very often.
I ended up spending a good chunk of the last 15 minutes trying to tweak the Nurikabe to work and failing to be honest. I kept trying to fill the 7’s in the corners, but I always ended up with a 2×2 somewhere I couldn’t get rid of. Was there a more logical way to approach that one?
For the most part, the logical approach was to [rot-13] jbex ba gur gbc naq obggbz rqtrf naq gur evtug-unaq fvqr, gura gb cebterffviryl vqragvsl sbeprq bprna va gur zvqqyr hagvy gur friraf ba gur yrsg jrer pbafgenvarq rabhtu gb svanyyl cebivqr nqqvgvbany vasbezngvba naq erfbyir. Guvf jnf abg n “bar vafvtug cyhf svir zvahgrf rdhnyf fbyhgvba” chmmyr gur jnl fbzr bguref zvtug unir orra.
Thanks to all of the authors and constructors for another nice set of puzzles this year. I solved 9 puzzles for a best-possible mark of 125, pending any answer sheet screwups. I got bogged down with suzaku gaiden at the end, making my score likely too low for placing 19th for the third year in a row…
If nothing goes wrong, I should have 225.
To start off, I intuited and solved the battleship, as it is the first page printed. The cave, masyu, and crosslink were nothing special, and were disposed of pretty quickly.
The the second of the match series was tricky, but I got it without wasting too much time. Then, I started tapa view, which sucked. I gave up on it after a little while to do other puzzles first. In the end, I never finished the tapa view.
I really liked the nurikabe path and the inside/outside fences, which I did next. The double minesweeper was ok, and the tapa is rather easy. The nurikabe took a bit of guessing though.
Then, I decided to go big or go home, and started the sukazu gaiden. I managed to broke the puzzle once early on, but got it in the end. I went back to the tapa view, and made no progress, so I switched to the pantamonium. It was a very neat puzzle.
Still no progress on the tapa view, and at ~8 minutes I decided to do the sudoku. I spotted the pair on the left pretty quickly, and had less than 1 minute to spare when I entered the answer key.
A few notes on my experience:
I looked at the Meta Tapa very early. After 30 seconds, I thought that H3,33 was a solution (which it wasn’t), and proceeded to move on to a different puzzle. In hindsight, that probably worked out in my favor; I don’t think I would have gotten the actual solution.
Skipped Match Series, Slide and Divide, Sudoku, Double Minesweeper, One or Two.
The Tapa View was quite hard, and I didn’t manage to figure it out. The Digital Clutter, well…
For the Star Battle I only used the word search constraints. None of the words took me more than 20 seconds, so I don’t think it was much of a setback for me.
I got lucky the last ten minutes. I intuition-ed through the Nurikabe Path and got the Geometric Distribution by luck, there’s 40 points.
So I’m expecting a score of 232. Hopefully this year I will finally get top 25.
“For the Star Battle I only used the word search constraints. None of the words took me more than 20 seconds, so I don’t think it was much of a setback for me.”
That’s absolutely what I should have done. Instead, I tried to do both at once, broke the logic portion, and ended up turning the puzzle into a major time sink.
My score was a 218, so either I’m much lower in the rankings than usual, or you’re much higher than usual. Either way, congrats on the superior score.
I hope this isn’t too much to ask,but could you post on the topic *before* the championship rather than after? Actually the same thing happened last year – I checked your blog and saw the announcements just hours after it was finished (and many threads for weeks after – very frustrating). Sure, it’s my own fault, but still, would be helpful!
Thanks!
Meh. Wasted too much time on Double Minesweeper (which I finally got) and Nurikabe (which I didn’t). Max of 129, which I suppose isn’t too bad, considering it’s been ten years since my last USPC.
Jeez; I wasn’t doing _too_ poorly for the first hour and a half; then I ran into a brick wall in the Nurikabe. Now I find out I neglected to type in the last pentomino on one of the rows in the pentomonium puzzle (it was right on the paper, but oh well). I think the lesson is that if I’ve solved very few puzzles in the last few months I shouldn’t expect a good performance…. And that double checking might be helpful.
I had a lot of terrible time management and a score I’m not so proud of, but the thing I’m most pleased with is my night-before analysis of the Digital Clutter. I noticed that for the puzzle to have a unique solution several lines would have to be given (for example, if the dash inside the zero isn’t given, it’s indistinguishable from the eight). This means in particular that both interior lines of the 5 must be given in the initial grid. As I’d hoped, there weren’t many spots where those two lines fit, which led to my solving that puzzle in a couple of minutes.
Biggest regret was getting stuck early on the Double Minesweeper, abandoning it, and never going back. When I picked it up again later, I got unstuck pretty quickly, and it took me a lot less than 20 points worth of solving time.
12 puzzles solved, including all differences. (190 Pts.)
Pentomino Areas (Pentamonium) was very nice, with almost symmetric regions. The biggest hint was ‘Avoid two P pentominos’
154 points I think? Seems like a pretty terrible showing for me compared with last year, anyway.
I may win the “stupidest mistake” award this year (not that I haven’t had good entries into that contest for previous years): I worked a long time on the nurikabe, saw that I ended up with the river not all connected, and put it in the “come back to later” pile, and never came back to it. After the test I saw I had the right answer, so if I’d just bothered to submit it, that would have been 20 more points right there.
I found the Masyu surprisingly hard for its point value. Other than that, lots of enjoyable puzzles! I liked the word search (though maybe the star battle component didn’t add quite as much as it could have? But it probably would have been a harder puzzle if it did.)
I liked the double minesweeper, the sudoku, and of course (as always, it seems?) Adam R Wood’s contribution. Maybe I just like number placement puzzles more in general. Though I kept having to remind myself not to use Ripple Effect rules!
Probably my favorite surprise here was the Battleships, which just felt much more like a clean logic path that was easy to follow, as opposed to what I usually do with lots of casework on where the big ships could go. I was also amused that they had to change the usual “enter the coordinates of the submarines” solution key.
Overall, great test, and I look forward to having some time to solve the rest of the puzzles!
Thank you very much for the kind words – that’s exactly why I keep at this. Evoking the feel of Ripple Effect was very much intentional with Sukazu and its “side story” (just as that bad pun of a title was) – I love how regions alone can define them, no givens or other special markings necessary. I apologize for any (unintentional) confusion!
I did have opportunity to tackle that Nurikabe I’ve heard so many talking about. I tried to make one of those with all the givens in a knight’s-move pattern once myself; it… didn’t go so well. It definitely requires techniques most Nikoli offerings don’t provide practice for, but I was able to reason my way through it without too much trouble (though I did break it once to be fair). Kudos to Craig for managing to build that thing! – ZM
The Battleships was a good way to start off the competition. I have seen the “1-square ships are trivial, now use the fact that everything else is 2 or more” gimmick before, but it was well done here.
Gaahhhh! How could I have left off a word in the word search??! I’m going to be kicking myself over that one for a couple of days.
25 points for me 🙂
1. Entry error on the Masyu.
2. Didn’t see the trick on the second line of the Match Series until an hour after the test.
3. Thought I nailed the Meta Tapa with C8,111 – until the results came out and realized a missed a nonsingularity in the bottom right.
4. Finished the Nurikabe with 25 seconds left to go – usually I fail in those efforts and lose the points.
5. Holiday BBQ seemed easier than prior years.
6. Really enjoyed the Slide and Divide, Crosslink and Tapa View solves.
Great set of puzzles, but my fails will haunt me for another year. Scott H, not in your way this year to the Top 25 (my similar goal each year).
Thanks to all the puzzle creators!
TheSubro
I take it you found an elegant way through the Slide and Divide? Can you give a hint?
Start at the bottom.
Got through now, thanks. This kind of puzzle is really hard work for me.
I didn’t do as well as I hoped. I was still trying to get a quiet work space set up when the test started (difficult with 3 dogs and a preschooler around!) and took about 10 minutes to settle down and focus. I wasn’t organized and ended up with a few half-solved puzzles. I thought I ended with 132 points, but ended up with 122 because I had two intersecting loops on Crosslink instead of one continuous one.
Since I dedicate only a little bit of time to solving puzzles, and no time to studying strategy or constructing puzzles, I will never be a serious contender for the U.S. Team – but it’s still fun to try!
During the 24 hours between the release of US Puzzle Championship instructions and the test, I wrote a handful of sample puzzles of some of the unfamiliar types (many moons ago, the founder of this very site taught me that was a good way to understand the new types better). I kept them mostly to myself this year, but they’re on the internet now, so have at them if you like. (They’re much easier than the instances that were on the test, but maybe some of you will prefer that!)
http://www.math.brown.edu/~dkatz/uspc-warmups.pdf
And of course, in that PDF, by Inside/Outside Sudoku I mean Inside/Outside Slitherlink.
Thanks! The Slitherlink feels completely different than the one in the contest, I’d like to see more of the type. The left side of the Digital Clutter plays out very nicely.
Scores and placements are out if anyone didn’t know.
Apparently, this WAS a low-scoring year!
68 out of 183 for me. Not bad.
And of course, Palmer comes out and wins it by a mile. Congrats to him, Wei-Hwa and new member Anderson. And of course, a hearty congrats to everyone here who dominated the rankings. Heres to a good USPC in 2015 as well.
Congrats to the three qualifiers.
I’m pretty satisfied with my 39th place out of 183 — it was my first USPC in ten years, so I really can’t complain.
I missed out on the USPC this year—bad timing—but now I’m going through the puzzles.
In puzzle 7, Crosslink, consider the solution but with the edges surrounding the cell in row 9, column 5 inverted. (That is, with the top of the cell connected and the other three edges unconnected.)
I can’t determine why this is an invalid solution—I believe it still results in a single closed path, and doesn’t violate any numbered square constraints. What am I missing?
The loop crossovers are always perpendicular. Picture each of the two lines continuing straight through the point of crossing. With the alternate solution you propose, there would be two distinct loops (one small one that surrounds the 444 and a single cell in R10C4, the second loop containing everything else). If you don’t understand this description, I can probably make an image to help.
After sleeping on it, I think I have my problem nailed down: a trace is invalid if it includes two “L” shapes meeting at their corners. This constitutes “touching” as banned by the problem statement. It must be a “perpendicular” cross-through.
Thanks again for your help.
Thanks for your reply.
I can see the disjoint “small loop” you described. My mistake was assuming that being able to trace one path along the lines was sufficient.
While solving, is the idea just to notice the fact that such a loop exists and connect it “perpendicularly” to the other loop? Is there some heuristic?
I don’t think there’s a heuristic. I’ve never done this myself, but I can definitely see a highlighter being advantageous to finish off a Crosslink, because you can then easily see that you haven’t highlighted every part of the path as you traverse along what you’ve drawn, so you know not to close it early.
Well, in the flawed solution I proposed, the entire link can be traversed as a single path, so a highlighter wouldn’t have helped in this case.
I had a similar problem with my first “answer” for this puzzle, although I found the two loops and fixed them myself by trying to draw over the whole loop at the end with my pencil. Crosslinks are rare enough that I don’t have any special notation yet to track ends.
I am a little behind in adding comments, but I’ll go:
Proudest moment: taking the inside outside fences problem, seeing the main “trick”, and disposing it in 10 minutes flat, which is a good minute per point ratio for me
Most disappointing moment: I had solved a good part of the 35pt suzaku problem, and had the answer key down to a 50/50 guess with 3 minutes left to go in the competition. I made a guess, and solved through and thought I had it. Had I noticed the contradiction I created, I would have gone the other way and gotten the problem right.
Scored 99, where I thought I had 134. I would have fallen just outside top 25 USA