Puzzle 384

I had to cheat symmetry a little bit on this one. Oh well; this was hard enough to construct as it is.

Many of you may be aware that my team (Manic Sages) won the MIT Mystery Hunt this past weekend. I’m hoping to contribute quite a bit to next year’s hunt, both logic puzzles and non-logic puzzles. I don’t anticipate any delays in my blog posting (whether the Wednesday puzzles or packs); the most likely compensation is that I’m probably not going to be writing as many LMI tests this year as I hoped I would.

Download the pack (PDF; 1.27 MB)

This pack contains

• A Slitherlink primer for novice solvers
• 15 classic Slitherlinks, mostly on the easier side
• 5 variations with 3 puzzles each: All Threes, Interior Path, Domino, Relation, Liar
• 3 chimeras each combining two of the above variations
• 2 larger puzzles
• Hints and solutions for each puzzle

Version 2 uploaded on 1/3/2012 10:12 AM ET. Fixed multiple solution errors in I.12 and I.23.

Version 3 uploaded on 1/8/2012 12:28 AM ET. Fixed a major uniqueness error in I.35.

There is a nontrivial possibility of a broken puzzle somewhere. If you think you’ve found one, you may either comment here or email me at palmermebane -at- gmail -dot- com. Any updates to the pack will be announced in this post, so be sure to check here first before reporting an error.

In place of my usual 3-4 per week schedule, I plan to release a pack of puzzles as a PDF every couple months or so. The packs will have 20-25 puzzles minimum spanning a full range of difficulties. Each one will have puzzles all of the same basic type or be themed in some other fashion.

The theming around a type or something else is the main intent of these packs. As things originally were I posted on here a potpourri of types across all difficulties, and many types and variations suffered from not having easy (or hard) instances. This system should allow me to do real justice to the best types and variations I’ve made and posted over the years, with accessible examples followed by challenges that bring out the type’s full potential. The weekly blog posts will be there to keep a bit of diversity and perhaps play with types or variations of a more one-off nature.

My 2012 New Years present to all of you will be the first pack with Slitherlink and variations. More info on that when it’s posted.

Next, there is the new option of PayPal donations. In its earliest form the idea was to sell each of the puzzle packs for a small price. I eventually decided that my need or desire for money is not enough to reduce the amount of people that get to enjoy these puzzles, so they will all be free. But if you decide that these packs are something you would have gladly paid for, you have the option of making a PayPal donation. Go here or use the button below.

The last thing to mention is that I’m going to shelve the Heyawake guide. In hindsight I think Numberlink was a bit unique in terms of how much specialized wisdom I had to share, and it’s a type that’s harder to learn solely by doing. In contrast every time I sat down to do some writing for the Heyawake post I never really felt like I was saying anything too informative or new. Maybe something more bite-sized will come later. Sorry to anyone who was anticipating this.

Virtually everyone reading this probably already knows the “twist” of the last post, which was that I ended up winning the world championship. Furthermore, the likely chance of us having made no error on the last team round came true and we won over Germany by an extremely narrow margin. Finally, the US team members went from preliminary round positions 2nd, 3rd, 9th, 21st to final standings 1st, 3rd, 4th, 21st. The playoffs were very kind to us.

It seems the individual time breakdowns for the WPC playoffs never got posted, which is a bit of a bummer as it’s something I would have really liked to see. Besides a time comparison between Ulrich, Thomas and I on each puzzle, I also wanted to see the times for the 3rd puzzle Pointing at the Crowd, which multiple people (Thomas, Bram, Nikola) reported later as being a performance killer. As said in the last post, Ulrich and I also sunk plenty of time into it, but I feel it would have been hard to go too much faster. I was hoping to make this post after that went up, but I suppose not. Oh well, no huge loss.

In the absence of that, there is a number of things I would like to say about the results and my performance. Officially, I won the world championship by finishing the playoff puzzles first, and by a substantial margin at that. What I do not believe is that I was the best solver at the 2011 WPC this year. A lot of my reasons for believing this are things that have been elaborated on before, although major props to the Hungarians for having no snafus. To summarize what determined the results, Ulrich evidently had a mental block on puzzle 8, a type that is far more intuitive than logical, although I don’t know any specific details. He and I had the distinction of not having mucked up any other puzzle in the playoffs, meaning I was there to pass him and win and he managed to recover and hold on to 2nd even as Thomas was rapidly closing in.

So if the playoffs had been converted to an equivalent sequential preliminary round (logistically impossible to do with all solvers in the competition, but let’s play pretend), it’s a round I would have destroyed. And the sequential part (puzzles must be solved in order) is important. If Ulrich had the ability to start the Divide and Conquer, put it off, and come back to it later with fresh eyes, it might have been a different story. As noted in the linked-to post on playoffs, the playoff format has substantial ramifications. Is a sequential round the best decider of things? Perhaps the most dramatic way, but speaking as a competitor that’s not a satisfying reason.

My actual playoff win was by a margin that probably meant I would have won even if I had 7 or 8 minutes of ground to make up as opposed to 2.5 (can’t be certain since the times aren’t posted). That’s about a 900 point disadvantage there, more than any solver scored in any individual round. Was my playoff performance good? Yes. 300 points worth (my actual deficit to Ulrich) better than anyone else? I strongly doubt it. 900 points better than anyone else? Absolutely not. As a preliminary round equivalent I think I might put the performance at 150 better than Ulrich as a loose upper bound, and it’s probably closer to half that.

I would buy being considered the second best solver at the WPC. I had a secure second place finish in the preliminary rounds and a solid playoff run showing I would have lost no ground there if that were a regular round. But there are not too many metrics that would evaluate me as having been better than Ulrich, who among other things obliterated the rounds in the afternoon of day 2 with top scores on 11, 12, 14, and a near-top one on 13. My disadvantages in 11, 13, and 14 could be attributed to some goofs of various kinds; I could understand matching his score if I had better runs. Still, that says something about his scary consistency.

It’s round 12, the long 80 minute “Hungaricum”, that really convinces me that I have no business being called the best solver. The scores show that the top 5 totals on this round were 870, 820, 665, 630, 590. I was the 820, and even before the scores came out I felt really good about how I had done assuming no errors. It was a lot like how I felt after round 2, a similarly long “Assorted” round on day 1, where I indeed ended up with the top score by a decent margin. So when scores come out and I turn out to have the expected 820, I expect a kick up in the rankings. That came true… if you ignore Ulrich’s score of 870. 870! I was on fire that round, over 150 points ahead of any non-Ulrich solver. What on earth would I have had to do to squeeze out 50 more points? How is that even possible?

So congratulations Ulrich. Even if you messed up a Divide and Conquer on day 3, you were clearly the top solver at the 2011 WPC. I’ll take the title for what it’s worth, but there’s plenty of ground left to cover. Winning the preliminary rounds next year is one aim, but who knows how viable that is. The puzzle types this year did kind of swing my way in general, with almost no Sudoku, very few Kakuro/Killer-style adding puzzles (so much for all of that practice), nothing like Dominos or Kropki that makes me cringe, and a lot of more abstract novelty types that I’m better at. I could see myself having more trouble on a different composition of types as I am now. A good showing next year might take some work to clean up the weaknesses I have left. There’s no getting complacent from one win this year.

I suppose this is starting to sound like another request to do away with playoffs, but I really do not feel this way. I can sort of see some of the merit from the standpoint of those that aren’t mired in trying to survive them. I would consider this more of a plea for everyone to be able to see past the bottom line on the results and press. The official standings show that I got 2nd place after two hard days of individual rounds and then had a very good playoff run. Apparently that means I am called the world champion, but it should be clear I was not the best solver there. Part of my peace with the existence of playoffs is knowing that a lot of others do think this way as well.

As these are probably my last words on the contest, I will acknowledge one thing conspicuously absent from a lot of my comments is my evaluation of how the WPC was put on, something you find a lot of in Thomas’s posts for example. I attribute this to my lack of experience; I can say a few things here and there (not having any playoff screwups was a pleasant change from past years), but overall I’m still trying to figure out just what to reasonably expect from organizers in these competitions. Thomas for example has many WPC/WSC attendances and a huge hand in organizing a WSC, and I have only two trips to a WPC.

Okay, now for a dramatic change in gears from competitive solving to constructing.

Awhile ago I promised I would return to posting at the end of November earlier this month. I’m going to extend that to the end of 2011 now. In the gobs and gobs of puzzles I did before and during the WPC from all kinds of sources, I’ve begun to think more about my own construction tendencies and what exactly I want to accomplish when I’m making puzzles. There could be some changes once I’m active again, maybe minor, maybe major, although I do not expect my overall output to decrease much. The main reason for the break is to work out exactly what it is I want to be doing, so it’s hard to be specific now. In the meantime, the Heyawake guide is likely to be finished some time in December.

Anyways, as compensation for taking a bit longer to get back to posting, here is a PDF containing 9 puzzles I wrote to practice for the 2011 WPC. These 9 types were novelties for which no past examples existed (Windows is the only exception; it was written before past incarnations were pointed out to me), so expect some things you haven’t seen before. Instructions are provided, but you can refer to official instructions if you also want examples.

PDF link

I should note that in most cases these puzzles can be a bit rough around the edges. They were constructed hastily and intended to be practice puzzles as opposed to masterpieces. The Borderless Coral in particular was deliberately made to have no logical solution for this reason. The XI Snake at the end is probably the only one that could stand with what I usually post on my blog, so I encourage you to try that one at least.

Since said team round was the first thing this morning, let’s go over that first. Eight different puzzles that are fairly well known, all of which use numbers in the grid as clues and all of which had no solution. We also get eight 4 by 4 grids with some numbers in them, and we have to place one piece over each puzzle to fix it and then solve the resulting puzzle.

I spent a lot of my time on a Fillomino and a Tren (Turkish type unveiled in an early OAPC). Honestly I didn’t make a whole lot of contribution; I thought the right Fillomino piece couldn’t possibly work for what seemed like a good reason, but there was one position where that didn’t apply and the puzzle solved easily from that. I had to have teammates point that out to me after we eliminated said piece from all other grids. I did get the Tren out and help a tiny amount on the Nurikabe, but that was basically it. Luckily my teammates pulled their weight, so we finished 12 minutes early in a one hour round. The Germans were done at about 20, so they didn’t get the 14 minute advantage they needed, assuming we had no errors. The US team should be first.

After that was a little over an hour of break time before the individual playoffs. The format was discussed last post, and my time penalty estimates were accurate. Still didn’t know the 9 types until shortly before it began, which irked me a little. But once we were all herded into the competition hall at the end of the break, the instructions were given to us.
1. Disconnect Four (round 2)
2. Battleships (round 2)
3. Pointing at the Crowd (round 13)
4. Cows on the Cube (round 3)
5. Borderless Paint it Black (round 4)
6. Hungarian Pentomino (round 12)
7. Coins (round 1)
8. Divide and Conquer (round 10)
9. Password Path (round 12)

I was fairly ambivalent after seeing these types. Nothing like a Yajilin/ABC-Connect/loop puzzle that I can destroy, although Divide and Conquer and Password Path were strong for me. I’m not swift at Battleships like others are, and Pointing at the Crowd was something I got out in the round but was a major timesink and maybe not worth it. Borderless Paint it Black is a crapshoot no matter how you look at it. Although I had gotten very lucky on the Borderless round, luck is luck and it may not be with me for the playoffs.

Shortly after going through all these thoughts, all of the other competitors filed into the sides of the room to watch, and we were soon off. I twiddled my thumbs for 2.5 minutes while Ulrich started and finished the first puzzle, and then got to start the…

1. Disconnect Four
Not your usual instance of this type. A typical one looks a bit like this, with a mostly rectangular grid that’s rough around the edges and X’s and O’s scattered about. This one was an outlined WPC with massive propagating chains of constraints instead of any of the usual logic. I was a little slow catching the odd cycles, which are the clear starting points, but my time was fairly level with Ulrich’s, so at least I hadn’t lost anything.

2. Battleships
There were a couple square ship pieces on this one with some corresponding column clues, and I quickly nailed the interaction to get the orientation of both ships using them. After being stuck for a bit, I did my usual: plunk down the battleship on a 4 clue and see what happens. End result: It worked perfectly. I think a small amount of time was gained on Ulrich, but not sure. I should note Thomas (who started 2.5 minutes behind me) was very quick on the first two puzzles and was rapidly closing in on the two of us.

3. Pointing at the Crowd
I had a great starting point I found for these during the round. So I looked for a place to apply it and it showed up absolutely nowhere. Uh oh. After some hard thinking I eventually found another kind of constraint that could be applied several times. It wasn’t easy to use, especially for a break-in, so my logical solve was a bit of a crawl, just like in the round. I ran into an odd problem at the end where I actually had every single space I needed to mark, but just didn’t know what to do with the other ones (that turned out to be empty). After some staring and not noticing any constraints left to satisfy, I silently threw up a prayer, X’d the rest of the spaces out and turned in. Fortunately I was right, so no one minute penalty. I don’t know how Ulrich and I compared here, but it was a pretty slow solve for both of us. Not an easy one.

The fourth puzzles and beyond are only available to the first 7 who get to them, with the rest eliminated, so a top 7 position secured…

4. Cows on the Cube
A small Star Battle on a cube shape that had just 1 star (cow?) for each row/column/and region. I start by looking for some logic, but there was basically none to be found. So I start plunking down cows in random places. A minute in I get something perfectly reasonable, even as Ulrich is working on it beside me. I nearly thought I had overcome my disadvantage, but I eventually notice there’s two cows in a large region far away from each other. Drat. No tweak to be found, so I started over, repeated the process several times, and eventually got it after a few minutes. I wasn’t too happy about this one turning into this kind of puzzle; I think my logic on this type is pretty good. I don’t think I lost any time here.

5. Borderless Paint it Black
15 by 15 grid; that’s manageable. There was a 4 4 5, so I figured I’d try putting that in and see if anything happened. Nothing did, after which I decided that with the amount of time it took Ulrich to turn it in it wasn’t going to be the full grid. Paint it Blacks are not quick solves. So I look for a small grid. Okay, here’s a region of 5 clues that all take up at most 5 spaces… oh look, another region of 5 clues that need 5 spaces in the other dimension. Do they work together? YES! Instant solve once I found the grid. There was some audible chatter from the audience when I turned this one in, as Ulrich was still waiting out the one minute checking period to move to the next desk. The gap was closing.

I should note that Thomas apparently did this one far faster than me, although he had already been stymied by the Cows on the Cube for awhile, so he was a bit farther behind at the time.

6. Hungarian Pentomino
When we got this one in round 12 it was worth massive points, and there’s no way to make the type much easier. I had figured before the round that this was going to be one of the heaviest puzzles. As is typical for me on this type (and a lot of other pentomino puzzles), I got about a half of it out logically, after which nothing else was coming to me and I had enough pieces eliminated to start throwing them down in places. My first few guesses didn’t pan out, and Ulrich solved the puzzle while I was doing this. Stress built up a little, but I do deal okay with pressure like this, so I managed to hold my mental composure and tweak into the solution. I did lose some time on Ulrich here, since he was off to the next desk before I turned in.

The last three puzzles are only available to the top 5, so now I’m at least 5th. My memory is not exact, but I think the only time that either of Ulrich and I were solving beside anyone else was Thomas on puzzle 3. Ulrich was usually a desk ahead of me though, but we were definitely in front.

7. Coins
Getting into this one logically seemed very difficult; none of the coin totals seemed to break up at all nicely. There was a peculiar thing about the totals that gave me a very good piece of intuition. I finally decided to go with it and started carefully placing educated guesses down, gradually filling up the grid. I ran into problems a few times, but was very fortunate that they were all easily tweaked. Ulrich was at desk 8 by the time I filled my grid. I check things, as is usual when you get answers by guessing, and it turns out I’ve overshot one of the totals. Very nearly cursed; would I have to fully erase? As it turns out I was able to flip-flop three coins and fix the whole thing. Oh god did I feel lucky.

8. Divide and Conquer
I hadn’t been sitting next to Ulrich for awhile, so a spot of hope appeared here when I started solving beside him and he was continuing to make use of his eraser. This was the novelty type of the championship and a strong one for me, being a fairly high scorer in the round, but Ulrich had beat me by 10 points so it was good for him too. I started throwing down some very reasonable guesses, and then ran into a counting problem in an adjacent corner. I made the evidently good decision to think hard about how to avoid the same issue, instead of just trying more guesses, and eventually got that corner to divide up very well. The whole puzzle propagated from there, and Ulrich was still appearing frustrated when I divided the last region. I raise the puzzle up, and not surprisingly there’s a lot of whispering around the room. If I get it right, I would be starting the final puzzle, a good type for me, before Ulrich did.

And I was correct. Ulrich had still not turned in, so at least a minute of advantage to finish this last one.

9. Password Path
This is a puzzle that generally doesn’t have a whole lot of options, so I immediately just start drawing in the most reasonable thing. Getting these right first try is hard, so unsurprisingly it broke quickly. I decided to ignore and start piecing other fragments together. They broke too in a more interesting way. I kept trying, and there’s this one letter that was giving me serious issues. So I turned to my logical sense and figured out the only way to get it to work, which rapidly got me about half of the grid for certain. I carefully pieced together the remainder of the path, and it fit together cleanly. Ulrich was still trying to solve puzzle 8, and making solving errors on Password Path is pretty hard to do. Have I done it?

I raised my completed puzzle. More din from the audience. I put my head in my hands and could not stop my arms from shaking for the minute-long checking period.

It was marked correct. I won. I am the 2011 World Puzzle Champion.

Remaining Playoff Results

Ulrich continued to struggle with the Divide and Conquer for quite awhile after I had finished. Thomas had eventually worked his way up to desk 8 in the meantime, and actually got it out faster than Ulrich did. He was already trying to crack the Password Path when Ulrich finally got over his mental blocks and completed puzzle 8. Around this time I think Wei-Hwa and Hideaki were moving to secure the other two positions in the top 5, but I don’t think either had reached the 8th puzzle, and the last one was not hard enough for either of them to have a good chance of catching up. So it was a battle for 2nd between Thomas and Ulrich, and the round was set to end when 3rd was decided.

Despite Thomas having gotten a head start, he evidently had taken too many wrong turns, and Ulrich managed to get it out a minute or two faster to claim second place. When Thomas finished and got the last spot on the podium, Wei-Hwa was evidently ahead of Hideaki at the time, leaving a top 5 of:

1. Palmer Mebane
2. Ulrich Voigt
3. Thomas Snyder
4. Wei-Hwa Huang
5. Hideaki Jo

Three Americans in the top four with Ulrich in the second position, just like what the Japanese pulled off the previous year.

The brief awards ceremony for the playoff results was not terribly eventful. A decent trophy, a lot of handshakes, a lot of pictures. I was evidently a little too stoic for the whole thing; one of the organizers expressed a verbal desire for me to smile a bit more. Oops. I always have that problem though. I’m one of the least photogenic people I know.

Then it was off to lunch, and I’m not sure anything has happened since then that’s worth noting. I believe team results and individual puzzle times in the playoffs will be given later, so I will likely have a final post on that.

A fantastically run competition overall. I had enjoyed my first one last year despite there being a lot of issues, and some may recall that Ulrich and Thomas were fairly dissatisfied. I’m not too sure what other opinions are, but I thought this one certainly outstripped the only other one I’ve been to in a lot of levels. Not perfect; things like a lack of a classics round (effectively) are a mark against it, and there was the occasional broken puzzle. I may have more thoughts on this in the upcoming post, but a lot of thanks are due to the organizers for their work in giving us a contest at this level of quality.

The last team round today, which functions as a “team finals”, is 8 puzzles of 150 points each plus 30 points per minute of time bonus, and it’s a one hour round. Japan in 3rd is probably too far back to catch us, but Germany needs 14 minutes of time bonus over us to win (equivalent to a bit less than 3 puzzles). The round is such that individual components of the puzzle can be solved in multiple ways but there’s only one way to do them all, so it’s easy to turn in lots of partial progress and get little to no credit for it. So Germany has plenty of winning chances.

Following that is the individual playoffs. 10 players trying to do 9 puzzles in order and finish the quickest, with time penalties for everyone below 1st based on preliminary round scores. I’m not sure I have the right method for how the penalties are calculated, but I think I start 2.5 minutes behind Ulrich, with Thomas around 5 behind and everyone else about 7 or more. Going up against Ulrich without any time penalty is a tough draw already, so I’ll need to fight my hardest to bag a win. Failing that, a top 3 finish is certainly the aim.

After those two events are over I’ll post what will probably be the last update.

[#. Name (R8; R9; R10; R11; Total)]
1. Ulrich Voigt (120, 315, 280, 560*, 3105)
2. Palmer Mebane (110, 375*, 270, 429, 3024)
3. Thomas Snyder (170*, 360, 245, 341, 3016)
4. Hideaki Jo (120, 340, 185, 385, 2745)
5. Bram de Laat (95, 325, 235, 297, 2662)
U. Michael Ley (145, 310, 115, 242, 2642)
6. Roland Voigt (145, 300, 205, 407, 2642)
7. Philipp Weiss (130, 280, 210, 341, 2531)
8. Neil Zussman (130, 310, 225, 319, 2529)
9. Wei-Hwa Huang (160, 345, 210, 286, 2521)
10. Peter Hudak (120, 300, 160, 297, 2512)

Again, U means unofficial participant, the usual caveat that there may be typos applies, and the * means top score of the round. I should also note that round 5 was handed back today, and I have a pending 50 point protest on it. My team thinks it could go either way. (EDIT: Didn’t get any points out of this.)

The day began with Round 8, a Screen Test. If you aren’t familiar with it, a puzzle, usually a very observational one, is flashed on a large screen for a short amount of time, and you have to solve it in your head and write the answer on a form. The organizers did something very interesting this year, with all kinds of animation present in the puzzles. One of the puzzles was a simple “Count the black circles”, with them all the same large size and not overlapping. When we get to that puzzle on the screen test, they’re in a spinning ring formation. Surprisingly difficult to do. A lot of other fun challenges meant the whole room was chuckling for a good portion of it. From a competition standpoint though, screen tests are one of my weaker rounds, so I dropped a bit of ground on it. Thomas got the top score on the round for the Nth time.

Next was the Sprint Round. 4 instances of 5 types: Ariadne’s Threads, ABC Connect (Numberlink), Fragmented Loop (Loopfinder), Shikaku, Yajilin. (Ariadne’s Threads could be concisely described as a consecutive Numberlink.) If you know me on nikoli.com, you know that these are all some of my best types. It was a 30 minute round, and I was the fastest to finish with 15 left on the clock. Thomas was the second with 12 minutes of time bonus, and others like Wei-Hwa and two Japanese were at the 9 or 10 minute mark. The whole US team finished, not a surprise since we are all on nikoli.com. We also all managed to avoid errors and get the full bonus.

That said, there is a bit of a bummer in this. The round was 300 points, and time bonus is 5 per minute saved. I solved the round at 20 points per minute, and my score is only 12.5 per minute. So the round with all of my best stuff wasn’t weighted heavily.

Round 10 was “Divide and Conquer”. It is evidently fairly standard for the Hungarians to have one or two rounds focused on a single new type (Crack it On and Honey Islands in 1999, Streets and Fifty-fifty in 2005), and this was the one this year. It resembles a Fillomino variation, except that all the region sizes are given and same size regions can touch. I thought it would be a strong round for me, but the actual round was wacky. I nailed all of the square grids, but it got crazier after that. Hex, Cube, Olympic Rings, the border of Hungary, a “#” sign (no one I know attempted this monster). Scores are out, and it seems I did fine being 10 points behind Ulrich, but I’m a little miffed about being 5 seconds from finishing a 40 point puzzle, which I was frantically trying to get done even as time was called. Oh well.

Round 11 was the “11″ round, and was intended to begin at 11/11/11 11:11:11. In reality we only just started handing out the tests at that time, but we were close. This round was peculiar with a lot of extremely high-valued puzzles that were not correspondingly hard. A weights puzzle with 11 weights is not 77 points. I did fairly well, getting all but two, but again, I’m kicking myself for making a dumb error on one of the highest valued ones and thinking it had no solution. It wasn’t unreasonable to think so when I had ascertained to myself that another puzzle in the round had 36(?!) solutions (although one extra clue could have fixed it). Ulrich finished the round with a couple minutes to spare, but fortunately I was clean, so I didn’t drop too much to him.

After a lunch break we had three more rounds, all longer, and all of which haven’t been scored yet.

Round 12 was a bunch of Hungarian puzzle styles in a long 80 minute round. Earlier I had done really well on the long “Assorted” round (2), so I had some reasons to be optimistic. At 5-10 minutes left I started struggling to find things to do, which is good in some sense (a lot done) and bad in others (dead time). I ended up managing to spend that time guessing my way through a relatively high-valued puzzle, so it felt like a success. One possible bummer is that the highest valued puzzle, a Pentomino one, had a “no reflection” rule that I missed. My team seems to think the solution didn’t have any room to mess up on this, but I’m still really worried as that’s a ton of points that could be dropped. If I get that though, I should have a good round.

Round 13 was the “Innovative” round with more unfamiliar types, which had me worried as there was a lot of stuff (e.g. variations on Skyscrapers and Japanese Sums) that I didn’t feel comfortable with. I ended up breaking several things in this round and not always fixing them, so a lot of time was lost. I also had a timing issue. Usually in a 60 minute round I can make several passes through the puzzles, but on my first run through, while trying to wrap up the last puzzle on the last page, a Skyscrapers variation, I looked up and saw there were 6 minutes left. Whoa! At about 4 minutes left I realized I had no shot at finishing the skyscrapers logically, so a flurry of guessing ensued. At about 15 seconds to go I had something reasonable, and it seemed that the plethora of constraints all checked out, but I’m just not certain. If I did get it out I think my score is quite good, but it’s a real tossup.

Round 14 was the “Best of” round. 77 small instances of types from all across the competition, and 60 minutes to do them. 10 points for every puzzle. I was moving chaotically, solving things, trying and breaking things, skipping things, and making multiple passes. I think my haphazard approach that caused me to go through the packet two or three times hurt me quite a bit, but we were informed of the round format about 1 minute before it started, so there was no time to think of a strategy. Due to how I attacked the round, I have absolutely no idea how I did. motris, for whom this style of round is fairly well-suited, thought he ended up with 59. I don’t think I got near that, but hopefully I don’t drop much from this.