#97 - 2024 24HPC recap + practice puzzles

I went to participate in the 24 hour puzzle championship in Budapest again last weekend! 

This was the 21st edition of the event, and it was my second time participating after my debut last year, in which I ranked 18th. This time I ranked 13th, so a solid improvement! Looking at the attendee list before the event, I thought my best possible result would be 12th, but also felt like I should be getting at least 14th. So, just like in 2023's WPC, I finished exactly where I predicted ahead of the event. Ironically, I didn't receive points for two puzzles that I solved fine, but only noticed it after the event was concluded - had I gotten them, I would have jumped to 12th place... :suseyes:

Like last year, the event was tons of fun and a fantastic experience. I wasn't sure for a long while whether I would attend, as the tournament coincided with the Dutch championships. Since I intend to participate in WSPC later this year and our championships serve as the selection process for the official teams, I didn't know whether I could/should skip them and still attend WSPC. However, after deciding I would rather attend 24HPC and participate in WSPC on a UN/lower team, and confirming with my organization that I would still be able to, the decision was eventually easily made. For now I have absolutely no regrets - we'll see if I still feel that way after October.

Having some much needed experience from last year, I felt more prepared for the tournament this time around. New additions to my supply kit: painkillers, pencil grips, correction tape and some energy drinks. The first three turned out to be quite helpful, the latter not so much - the instant boost is nice, but the crash afterwards much more impactful. I knew staying awake wouldn't be too much of an issue, though I definitely had more trouble staying focused at night this time, something I attribute to having to wake up super early the two days prior to the event, for work commitments and an early flight respectively.

If you're just looking for the practice puzzles - those are towards the bottom of the post. If you want to read about my experience every round, carry on!

--------------

Round 1 - puzzles by Nikola Zivanovic: 555/1000 (highest score 1000 / 55,5 tournament points / 19th out of 45)
More points in the first round than last year, but a lower ranking compared to my prediction. This round featured some types I'm not great at, which led to me leaving a lot of high pointers (Knight Pentomino, Hexagon Code and all the word puzzles). Still, I was content with my performance and felt optimistic going into the next round. As always, a nicely themed set with some stellar puzzles.

Favorite puzzle: the 2nd Gemini Loop - fantastic logic, very satisfying to figure out.

Towards the end of the round, I figured I'd give a Word Search Sudoku a shot (initially marked 'skip' in my IB) - only worth 20 points, it still took me around 8 minutes to solve

Round 2 - Matus Demiger: 460/1000 (880 / 52,3 / 18th)
One of five rounds where I scored below 500, so mildly disappointing, but it turns out I wasn't awarded 70 points for two correctly solved puzzles (both Capsules). Unfortunately, I only noticed this while having lunch after the awards ceremony already concluded, and was too tired to make an effort to protest, so the score stands. Once again I left some high pointers I wasn't comfortable with (FilloKuro, Scrabble, Japanese Sums), and generally lacked pace for many of the puzzles I did solve. All very clean puzzles, making for a solid round.

Favorite puzzle: the 2nd Light and Shadow - challenging but logical, without using question marks, always nice to see.

Solving Ququ on paper was an interesting experience

Round 3 - Silke Berendes: 555/1000 (1000 / 55,5 / 16th)
This was my skip/continuous round, skipping Vladimir Portugalov's round. This round featured quite a few types I'm strong at, but most of them were only worth 10-30 points each. I solved those pretty easily, then burned a lot of time on the 35 point Sums Slitherlink, which felt like it should have been worth more points. It scared me off from solving the second one. This was the first round with a classic/variant structure, something we would see a lot more of. I enjoyed the round - the puzzles were fun and I always like this format. 

Favorite puzzle: Both Star Battle (Galaxies) - very fun and clean solves.

A delightfully tweaky Spiral Galaxies

Round 4 - Anurag Sahay: 820/1000 (1000 / 82,0 / tied 10th)
By far the easiest round of the tournament this year, as demonstrated by my score, the average score and the amount of finishers (three, where every other round saw one or none). The round would have had more finishers if it weren't for the first Evolmino - a horribly undervalued 20 point puzzle that ended up costing many people a significant chunk of time, top solvers included. I heard times ranging from 5 to 25 (and still unable to solve) minutes. I spent about 15 minutes on it, unable to solve it, before moving on to collect points elsewhere. It was a bit strange that there were 7 Masyus (variants included, compared to at most 3 of any other puzzle type), other than that the puzzles were fine. 

Favorite puzzle: Masyu (Yajilin) - always a nice type.

A new type that sparked quite a bit of discussion ahead of the event, because of a rather uninformative example in the IB - thankfully the contest puzzle did not come with any hidden surprises

Round 5 - Team India: 535/1000 (740 / 72,3 / 4th)
By points value this result seemed unremarkable to me, but I'm only noticing now while writing my round-up that I was 4th in the round, so this was actually a solid result. It does come with a huge sidenote: the round was skipped by 4 people that ended up higher in the rankings than myself. Still though, out of 6 practice puzzles I wrote for the event, 3 were for this round, so I felt very prepared for it. Starting full of optimism, I quickly learned the puzzles were going to be more difficult than I anticipated - it took me about an hour to solve 6 out of 21 puzzles (worth 330 points). One of them was the 90 point Shakashaka (Candidates), which I later learned was only solved by 3 people during the round. I spent the rest of the round bagging points on more familiar genres, and ultimately my effort was enough to get a solid ranking. A nice detail: two of my practice puzzles featured similar break-ins/deductions to the actual contest puzzles - I hope other contestants spotted that and were helped by it too :). The round itself was fantastic as expected, full of great puzzles and interesting variants. 

Favorite puzzle: Balance Loop (Windows) - challenging, but full of cool deductions, and a beautiful grid to look at.

Anticipating the clue cells in this puzzle would be black, I brought a white ink pen to be able to write the clue values - it ended up backfiring as ink barely came out of it and I had to overwrite all the numbers about 5 times before they were at all legible

Round 6 - Peter Gyimesi: 525/1000 (930 / 56,5 / 9th)
This round was divided in sets of two puzzles, one being vanilla, the other being ciphered. The first classic puzzle only featured 1 clues, the second only 2 clues, all the way up to 9. This worked well for some puzzles, but made other puzzles a bit gimmicky/theory-heavy. Particularly the Slitherlink (all 2s, with a few unclued cells) seemed a bit unfair to me - if you've seen it before it takes no more than 20 seconds to solve. If it's your first time encountering one though, it will hold you up for a good while. I attacked the cipher puzzles with a lot of guessing and bashing, which worked well in some cases, but not so much in others. This is also the first round where I made a mistake, losing 80 points on a classic Heyawake, points I thought I would surely be collecting with relative ease. 

Favorite puzzle: the classic Aqre - it looked intimidating at first, but it was a beautiful solve.

I got quite a few comments on my inefficient Tents notation. But hey, if it works, it works, right?

Round 7 - Tomasz Strozak: 325/1000 (870 / 37,4 / tied 15th)
Like last year, round 7 turned out to be my worst enemy - again my lowest scoring round in both round points and tournament points. The one positive I'll take from it is that I scored 325 versus 320 last year. That's the smallest possible improvement, but it's an improvement nonetheless. The round featured puzzles in sets of 3 - one vanilla, one encoded to potentially incomplete Roman numerals and one odd/even variant. Besides the scary Roman numeral variants, there were quite a few types I'm not great at (Snake, Minesweeper, Skyscrapers). I spent about half an hour on the Skyscraper (Odd/Even) without being able to solve it. I then started all 3 minesweepers, and failed to solve a single one. To finish it off, I had a one-cell error in the Slitherlink (Roman), which I also spent a long time on. At least I got a good laugh out of the Slitherlink (Odd/Even). 

Favorite puzzle: Choco Banana (Odd/Even) - the trick wasn't too hard to figure out, but it's a very cute and clever one.

Mistakes are starting to show up more often now...

Round 8 - Tawan Sunathvanichkul: 555/1000 (708 / 78,4 / 7th)

I really liked Tawan's round last year, and from the looks of it, it was going to be a favorite this year again as well. The 100 point Word Search made its return in the form of a tough variant where words were given in sets of three, but only one would appear in the grid. There was never another option - I was definitely starting my round with that. I was asked to take note of my solve time, which turned out to be 17 minutes. Not bad! The rest of the round was great too, full of neat surprises and easter eggs, as you almost come to expect of Tawan.

Favorite puzzle: 33% Word Search - how could it be anything else? Alternative Tapa is a clear runner-up, though I won't tell you why - you'll see when you open the booklet!

An Odd Labyrinth Sudoku, the only puzzle in the tournament that made me take out my highlighter

Round 9 - Krystian Swiderski & PAKO Team: 435/1000 (1000 / 43,5 / 24th)

Ouch - that's the lowest ranking so far. Flipping through the booklet while writing this, I'm struggling to remember what happened in this round - it looks like I definitely should have been able to solve more here. This round took place at the time I usually go to bed, so I shouldn't have been that tired yet, though I do remember fighting the urge to want to go to sleep. Perhaps I did doze off for a while... As for the round itself, there wasn't really a theme to discover (or I missed it) and it was mostly seemingly unrelated types, but that doesn't mean the puzzles were unenjoyable. 

Favorite puzzle: I liked the logic in the classic Railroads, and was happy to see more of the type.

(see previous image) Colored pens don't count as highlighters, right? :clueless:
Also, an avoidable error and the only puzzle for which I wrote a practice puzzle that I ended up not getting points for (Slant (Liar))

Round 10 - GM Puzzles: 570/1000 (985 / 57,9 / 8th)

Easily one of my favorite rounds, though I don't think that's a surprise - does anyone expect the GM Puzzles round not to be good? As an added bonus, it featured more than only the usual blog types - I didn't get around to the Tachiawase Blocks during the contest, but greatly enjoyed the Pencils and less common loop genres. My only remark would be that both Pentominous felt undervalued. I probably could have optimized my score a bit more if I solved other puzzles instead of the 70 point Ripple Effect as that held me up for a long time, but overall I'm happy with my result here.

Favorite puzzle: the second Pentominous - one of those beautiful minimal grids that just leaves a strong impression. 

One of my new toolkit additions, correction tape, proving its worth in this Arrow Sudoku

Round 11 - Sinisa Hrga: 380/1000 (840 / 45,2 / tied 24th)

Another ouch. I struggled to keep focus especially during this round, which took place from 4:40 am - 6:20 am. Though I don't think that, had I been more awake, I would have managed much more anyway, as I didn't really feel like doing any of the genres besides the ones I did. The round was pretty much a repeat of last year (and all years before that), so not much more to mention.

 Favorite puzzle: Four Winds, marginally

My fault for not reading the rules properly, but could those diagonals not have been made a little more visible?

Round 12 - Puzzlers Club (Zodiac): 495/1000 (830 / 59,6 / 9th)

The Puzzlers Club rounds were some of my favorites last year, and the same is true for this year - lots of variety, strong theming and solid puzzles throughout. Slightly less fun was working through the Old Maid methodically and ending up with only two unique images, only to have to do it all again, but with large X-es in all the images. I also ended up with an error in the 70 point Icewater Walk, which is disappointing - I would have jumped quite a bit in the rankings.

Favorite puzzle: so hard to choose! Standouts for me were the Oyakodori (a very soothing solve), the Balance Loop (just clean throughout) and the Yajisan-Kazusan (great execution of a familiar theme)

(Tiger in the Woods) A (nice?) difference between 24HPC and other events is the leniency with errors - the circle around the 15 indicates points are allocated, even though my solution in the bottom right is wrong. 

Round 12 - Puzzlers Club (Greek): 600/1000 (880 / 68,2 / 6th)

My second-best (or arguably best) result in the last round? Woah! Perhaps that's the experience paying off. In any case, a solid result, which wasn't even optimized: I spent the final 20 minutes of the tournament on the Fractional Division, alternating between dozing off and solving small bits here and there. I managed to only solve about 60% of the grid before Gyorgy announced in a loud voice that there was only 1 minute left. For some reason this energized me and sprung me back into action, and I solved what was left of the puzzle in about 20 seconds without any problem at all. As for the round, it was just as great as the previous one, for all the same reasons. A fantastic closing round to a fantastic tournament.

Favorite puzzles: again, impossible to pick! Cross the Streams had very interesting logic, the Ring-Ring was great and seeing a well-made Nagareru Loop is always nice. 

(Simple GAKO) Counting remains one of the hardest things, especially towards the end, where apparently even counting to 3 becomes too much asked. You can also see I had 3s where the 4s are - I changed those, but failed to adjust the 2s.

--------------

Some random afterthoughts:

- My worst enemy this year was Pentopia - I attempted every classic one, and ended up breaking and giving up on all of them.

- Don't circle your eraser shavings when solving a Spot the Differences puzzle. Also, don't start aggressively circling everything because you are failing to see the images are reflected. 

- I dearly missed Bram's round this year - I hope we'll see it again in future editions. Perhaps we can even team up and do a Team Netherlands round.

--------------

And that's a wrap! I'm happy with my result and take a lot of positives from the rounds where I performed above expectation. I'm also now well aware what I need to be working on to improve even more: feeling more comfortable attacking high-pointers especially when already tired and overall becoming better at word and math type puzzles (:dismay:) - we'll see if that works out next year. As always, a big thanks to the organizers, authors and participants for making the event super memorable, exciting and fun again! I will definitely be back again :)

--------------

To close out the post, I wrote 6 practice puzzles this year. They were all written over the course of a few days and roughly match the points value of the puzzles that appeared in the contest, with a few exceptions. The puzzles are presented in order of appearance throughout the rounds. My favorite puzzle is the Slant (Liar). Rules are included in the penpa links. Enjoy!

Turning Fences (Unequal Lengths) - diff. 4/5
https://tinyurl.com/29gkgceh

Unequal lengths usually allows for some minimal grids. I focused on the interactions with the turning mechanic giving small but crucial progress, which becomes especially evident in the middle of the grid.

Balance Loop (Windows) - diff. 3/5
https://tinyurl.com/22xum9kl

I was expecting the contest puzzle to have clues in the windows (which turned out to be correct), so focused on having my break-in centered around that.

Shakashaka (Candidates) - diff. 3.5/5
https://tinyurl.com/29doensp

Here I focused mostly on having groups of different sizes, to get a good feel for which ones are more useful than others. The 1 domino one cell away from the edge was present in an important manner in the contest puzzle as well, which was a nice bonus.

LITS (Signpost) - diff. 2/5
https://tinyurl.com/22svovh3

Again, I mostly focused on having variant clues of different sizes. The contest puzzles were mostly low to medium value, so I kept the break-in very simple and made the variant more important in the mid-solve.

Slant (Liar) - diff. 3.5/5
https://tinyurl.com/29wlbzsh

Dense edge clues are an easy way to make initial progress and I couldn't think of many other ways to start the puzzle, so I added 4 different sections of those with minimal interaction. In the latter half of the solve, focusing on the liar location but also the negative information the liars provide becomes much more important.

Nurikabe (Triangular) - diff. 2.5/5
https://tinyurl.com/2bp667ls

I was mostly curious how powerful the no hexagon fully shaded rule was, so focused on reachability. The contest puzzle ended up focusing more on packing.