So I started the morning finishing There Will Be Blood. I still find Juno my favorite film of last year, and No Country would now be third after Blood, but Day-Lewis' performance is incredible and - if you know my love of Magnolia - you'll understand how I maybe can't do anything but praise everything Anderson does.
Anyway, with the thought in my mind that maybe my obsession with beating the test this year was unhealthy in a similar way to Plainview, I went to warming-up puzzle wise. I finished off the last of way too many word searches from Tuller's book that I've been solving since WSC3 time. I did today's nikoli puzzles which included a great Extra Nurikabe (another solid one after last week's Botsu Baku) where I beat the field by about a minute, so I was solving really soundly. With 20 minutes left, I finally reread the instructions, particularly instruction info where Black Pearl's entry "consecutive 1s" was the most interesting new discovery, and then planned my music list. I started with the end of last year's stereogum/team9 mash since the final countdown is a great thing to mash-up and a good psych up song, then merged that list into the two good coldplay albums in order (as they are great fade into the background noise) with OK Computer ending the 2.5 hour session. If I was "lucky", I would be done well before The Tourist which would end at about 12:31 PST. If I was very lucky, I wouldn't even need to search for an Airbag with the test already done at 1h40m.
So, onto the test.
In a rush to save time, I jumped in the pdf to the STD before even beginning the first page which was not yet printing. I had 5-6 of the differences before I picked up the test from my printer, then finished that off and the long entry process, before going onto the next page. Then, I realized I had duplex-printed, so I just did the backside immediately which wasn't my earlier plan (5, 1, 2, 20 to the front). So I got the black pearl and masyu done, then the battleship and sudoku and star and arrow ring (answer submission on arrow ring again impossible, which is why I counted inside, outside, and loop, and saw I was missing an outside square). Maybe 15 minutes in and 7 puzzles done. I always confirm and reconfirm all my entries on any page, before throwing it on the floor as "done." Since my first year, I've never made an entry error provided I can understand what I'm entering.
So, onto the back of the test as originally planned.
Ampers& is a beautiful puzzle, but I cracked its secret too soon. It has 4 rings, so I mapped out in word space how those would have to be. I got all 4 identified (and correctly) from logic, then started steering the 4 "gears" to get them into the grid. It may have taken 6-7 minutes, but I blazed through it.
Onto the KakurOh, the most beautiful of puzzles that just happened to be on the back of that same page. It was in my construction-thought space, so I did not take long to grok or make use of the gimmick. Still, the puzzle was a tough one and perseverance carried it through with the long 7's being the most critical places to identify missing digits all throughout the grid.
Onto the SuDON'TKu - my first theory on how the puzzle was constrained was horribly wrong (missing row/column/box don't all intersect at a 9), so I broke it about 80% in and I decided to do something else and went back to the front to end up here.
The word search (Nebijok, I'd be told by Nick after the test, is a song and the word list - out of alphabetical order - is actually in lyrical order) was a quickie. I mean, when the word search doesn't have words adding/dropping letters, making a variety of different kinds of turn in the middle, require anagramming, have letters that extend outside the grid, is a toroid, ..., its pretty easy!
Number blocks 1 took me awhile, but had a satisfying logical deduction. I had figured out the bottom row of it (since it wasn't a sum of 6), pretty early, but had not gotten the right pair for the second column even figuring it had the 9 for awhile. The second fell much quicker. Its unusual to require both of these for any points, but I liked these block-type puzzles.
Korumasu is something I've done so often now - in the Giants books, and elsewhere - that its my basic corral notation (lines in white squares, scribbly blackness) and solving method without the black squares going to outside constraint but with a no-touch constraint. You'll learn that work-ins here focus on diagonal squares around 2's and sometimes 3's which become forced whites. Here, it was an even easier "the 2 goes just one way", but I'd banked a lot of info on the grid before seeing that and it fell fast. Where Is Black Cells indeed.
I'd been spoiled yesterday by![[info]](http://l-stat.livejournal.com/img/userinfo.gif)
jeffford on what Murder No. 6 was, but seeing an 8x8 grid with now forced 15 and 14 pairs was a quick deduction and it was just getting the puzzle done. Not too hard, and I'd love a more complex killer construction with larger cages that used single digit elimination as here. Another great puzzle idea I had not necessarily seen before (most people want to add 0 to the 1-9 set, not reduce the number set somewhere in the middle).
Corral was next and it fell pretty normally. Last year's was maybe my favorite puzzle of that USPC. This was more standard but had the nice black square chase at the end to figure out the final answer.
Tilted Weights was me getting lucky when I'm already well prepared. My exhaustive search space would start on the far right which seemed reasonable. 1/3 is the easiest pair to put there, and with 2 consecutive digit fulcrums, I didn't have a lot of choice to the other pair. 4/7 + 5-6, 8-9 + 2 was the first of my next steps in thinking and that became the answer in 30 seconds. Yes, I drink your milkshake quick like that. Its a nice variation (maybe not as cool as last year's) but I could use a harder one with maybe 3 weights on a bottom fulcrum so more "math" was needed.
Distances was beautiful.thedan had called it corral-like yesterday, when I was not yet thinking of it that way - but it really was. After I finished the satisfying puzzle, I noticed the beautiful latin square constraint on the new circles (one per row/column). Phenomenal! Thanks Craig!
Triangles, triangles everywhere. I was greedy and tried to pack the biggest and hardest triangles (1/2 + 2/4) on diagonals early. But I kept doing it wrong. I was going over 1/4 for example, and up 1/2. That was not comforting. After a couple erasures, and realizing I was making lots of incorrect triangles, I moved on for now.
Crisscross Pairs was similarly beautiful - I was right in the "one letter off" property yesterday which was what I was hoping to see. I again attacked square rings as on the Ampers& and decided a Y had to be in one of the two lower right square blocks. The MNCI starting letter choices only allowed one choice, so with that placed (and a decision it was on the left), the puzzle fell.
Sheep in Fences was also what I expected, and shading helped me steer some of the corners very early. I think I somehow busted my first intuited path, making an error with the 13 in the right-middle, but got it the next time. Answer entry on this was hard!!! I kept freaking out that I'd left out one of the six consecutive 2's. Even on my final check-through, this was my heartstopper until I saw I'd put it in right.
Back to the SuDON'TKu - I made a matrix of the eight boxes on the right of the puzzle, and marked rows/columns/boxes with missing digits when I knew things for sure. I no longer made the mistake of thinking the missing digits would simultaneously fulfill all constraints. From there, you start by parking 9's, but that was not enough of a work-in. Getting the missing 1 and 5 row/column/box helped and then a lot of naked single searching and it finally fell. Not as efficient here as I thought I'd be, but "Paranoid Android" was about to start (it was 11:45) and I just had the triangles left.
Now, I had been called out by someone earlier in the week in an lj message that printing multiple copies of the test is against the rules. I carefully checked things over and saw that is not a true reason to call me out. This is how I learned the winners won't post til ~ the 25th, that Nick had actually updated the news I'd heard earlier that Minsk is the new WPC site, etc. I mean, I want to drink your milkshake, but I want to do it within the law so when I saw "I'm finished", there isn't a dead body around. Well, you can print things multiple times if you need to. On a T&E puzzle where erasure bits will cover you soon enough, another copy is good. While my duplex-printing saved paper earlier in the day, triangles took out a tree in comparison. I tried after getting a lot of 6 but not 7 triangle solutions (maybe 5-8), to go at the triangles with the least choices. B's only got 2, and all the ones that failed were using the BDK choice, so I finally did the ABG choice, and having seen almost all the triangles in the puzzle by then, the rest fell fast. I still quadruple-checked all my triangles were valid.
Let's have a poll: do you prefer triangles as is for 20 points, or would you prefer it is "count all the triangles you can form in the grid" with the same 90 degree constraint for 20 points? I don't know what my choice is, but I've now done both and I'm only sure the first choice has a correct answer.
So done, and confirmed clean. 2h14m. Not my highest number (365 < 370 in 2006), but my highest percentage (100%). I ended on "Fitter, Happier". I feel like a run, and then the US Open, which will make me both. I've neglected all the comments in the last thread so I may go deal with that, but probably some food and exercise and fresh air is in the works.
Anyway, with the thought in my mind that maybe my obsession with beating the test this year was unhealthy in a similar way to Plainview, I went to warming-up puzzle wise. I finished off the last of way too many word searches from Tuller's book that I've been solving since WSC3 time. I did today's nikoli puzzles which included a great Extra Nurikabe (another solid one after last week's Botsu Baku) where I beat the field by about a minute, so I was solving really soundly. With 20 minutes left, I finally reread the instructions, particularly instruction info where Black Pearl's entry "consecutive 1s" was the most interesting new discovery, and then planned my music list. I started with the end of last year's stereogum/team9 mash since the final countdown is a great thing to mash-up and a good psych up song, then merged that list into the two good coldplay albums in order (as they are great fade into the background noise) with OK Computer ending the 2.5 hour session. If I was "lucky", I would be done well before The Tourist which would end at about 12:31 PST. If I was very lucky, I wouldn't even need to search for an Airbag with the test already done at 1h40m.
So, onto the test.
In a rush to save time, I jumped in the pdf to the STD before even beginning the first page which was not yet printing. I had 5-6 of the differences before I picked up the test from my printer, then finished that off and the long entry process, before going onto the next page. Then, I realized I had duplex-printed, so I just did the backside immediately which wasn't my earlier plan (5, 1, 2, 20 to the front). So I got the black pearl and masyu done, then the battleship and sudoku and star and arrow ring (answer submission on arrow ring again impossible, which is why I counted inside, outside, and loop, and saw I was missing an outside square). Maybe 15 minutes in and 7 puzzles done. I always confirm and reconfirm all my entries on any page, before throwing it on the floor as "done." Since my first year, I've never made an entry error provided I can understand what I'm entering.
So, onto the back of the test as originally planned.
Ampers& is a beautiful puzzle, but I cracked its secret too soon. It has 4 rings, so I mapped out in word space how those would have to be. I got all 4 identified (and correctly) from logic, then started steering the 4 "gears" to get them into the grid. It may have taken 6-7 minutes, but I blazed through it.
Onto the KakurOh, the most beautiful of puzzles that just happened to be on the back of that same page. It was in my construction-thought space, so I did not take long to grok or make use of the gimmick. Still, the puzzle was a tough one and perseverance carried it through with the long 7's being the most critical places to identify missing digits all throughout the grid.
Onto the SuDON'TKu - my first theory on how the puzzle was constrained was horribly wrong (missing row/column/box don't all intersect at a 9), so I broke it about 80% in and I decided to do something else and went back to the front to end up here.
The word search (Nebijok, I'd be told by Nick after the test, is a song and the word list - out of alphabetical order - is actually in lyrical order) was a quickie. I mean, when the word search doesn't have words adding/dropping letters, making a variety of different kinds of turn in the middle, require anagramming, have letters that extend outside the grid, is a toroid, ..., its pretty easy!
Number blocks 1 took me awhile, but had a satisfying logical deduction. I had figured out the bottom row of it (since it wasn't a sum of 6), pretty early, but had not gotten the right pair for the second column even figuring it had the 9 for awhile. The second fell much quicker. Its unusual to require both of these for any points, but I liked these block-type puzzles.
Korumasu is something I've done so often now - in the Giants books, and elsewhere - that its my basic corral notation (lines in white squares, scribbly blackness) and solving method without the black squares going to outside constraint but with a no-touch constraint. You'll learn that work-ins here focus on diagonal squares around 2's and sometimes 3's which become forced whites. Here, it was an even easier "the 2 goes just one way", but I'd banked a lot of info on the grid before seeing that and it fell fast. Where Is Black Cells indeed.
I'd been spoiled yesterday by
jeffford on what Murder No. 6 was, but seeing an 8x8 grid with now forced 15 and 14 pairs was a quick deduction and it was just getting the puzzle done. Not too hard, and I'd love a more complex killer construction with larger cages that used single digit elimination as here. Another great puzzle idea I had not necessarily seen before (most people want to add 0 to the 1-9 set, not reduce the number set somewhere in the middle).Corral was next and it fell pretty normally. Last year's was maybe my favorite puzzle of that USPC. This was more standard but had the nice black square chase at the end to figure out the final answer.
Tilted Weights was me getting lucky when I'm already well prepared. My exhaustive search space would start on the far right which seemed reasonable. 1/3 is the easiest pair to put there, and with 2 consecutive digit fulcrums, I didn't have a lot of choice to the other pair. 4/7 + 5-6, 8-9 + 2 was the first of my next steps in thinking and that became the answer in 30 seconds. Yes, I drink your milkshake quick like that. Its a nice variation (maybe not as cool as last year's) but I could use a harder one with maybe 3 weights on a bottom fulcrum so more "math" was needed.
Distances was beautiful.
thedan had called it corral-like yesterday, when I was not yet thinking of it that way - but it really was. After I finished the satisfying puzzle, I noticed the beautiful latin square constraint on the new circles (one per row/column). Phenomenal! Thanks Craig!Triangles, triangles everywhere. I was greedy and tried to pack the biggest and hardest triangles (1/2 + 2/4) on diagonals early. But I kept doing it wrong. I was going over 1/4 for example, and up 1/2. That was not comforting. After a couple erasures, and realizing I was making lots of incorrect triangles, I moved on for now.
Crisscross Pairs was similarly beautiful - I was right in the "one letter off" property yesterday which was what I was hoping to see. I again attacked square rings as on the Ampers& and decided a Y had to be in one of the two lower right square blocks. The MNCI starting letter choices only allowed one choice, so with that placed (and a decision it was on the left), the puzzle fell.
Sheep in Fences was also what I expected, and shading helped me steer some of the corners very early. I think I somehow busted my first intuited path, making an error with the 13 in the right-middle, but got it the next time. Answer entry on this was hard!!! I kept freaking out that I'd left out one of the six consecutive 2's. Even on my final check-through, this was my heartstopper until I saw I'd put it in right.
Back to the SuDON'TKu - I made a matrix of the eight boxes on the right of the puzzle, and marked rows/columns/boxes with missing digits when I knew things for sure. I no longer made the mistake of thinking the missing digits would simultaneously fulfill all constraints. From there, you start by parking 9's, but that was not enough of a work-in. Getting the missing 1 and 5 row/column/box helped and then a lot of naked single searching and it finally fell. Not as efficient here as I thought I'd be, but "Paranoid Android" was about to start (it was 11:45) and I just had the triangles left.
Now, I had been called out by someone earlier in the week in an lj message that printing multiple copies of the test is against the rules. I carefully checked things over and saw that is not a true reason to call me out. This is how I learned the winners won't post til ~ the 25th, that Nick had actually updated the news I'd heard earlier that Minsk is the new WPC site, etc. I mean, I want to drink your milkshake, but I want to do it within the law so when I saw "I'm finished", there isn't a dead body around. Well, you can print things multiple times if you need to. On a T&E puzzle where erasure bits will cover you soon enough, another copy is good. While my duplex-printing saved paper earlier in the day, triangles took out a tree in comparison. I tried after getting a lot of 6 but not 7 triangle solutions (maybe 5-8), to go at the triangles with the least choices. B's only got 2, and all the ones that failed were using the BDK choice, so I finally did the ABG choice, and having seen almost all the triangles in the puzzle by then, the rest fell fast. I still quadruple-checked all my triangles were valid.
Let's have a poll: do you prefer triangles as is for 20 points, or would you prefer it is "count all the triangles you can form in the grid" with the same 90 degree constraint for 20 points? I don't know what my choice is, but I've now done both and I'm only sure the first choice has a correct answer.
So done, and confirmed clean. 2h14m. Not my highest number (365 < 370 in 2006), but my highest percentage (100%). I ended on "Fitter, Happier". I feel like a run, and then the US Open, which will make me both. I've neglected all the comments in the last thread so I may go deal with that, but probably some food and exercise and fresh air is in the works.
37 comments | Leave a comment