Sunday, September 11, 2016

24HPC 2016


Like last year, we solved 24HPC on May 3-4. Recently I managed to have time to solve the remaining. I do not remember my strategies clearly, so here are just a few comments.
Scores are here. Note that my skipping round was my best result, with the suggestion from Kota.

Round 1
Nice theme. I do not understand if Strangely as ABC have logical path.

Round 2
2-cell error in the 2nd Missing Magnets: my last solved puzzle.
Good to solve some unfamiliar types. I should have solved Anti-Magnets.

Round 3
Just a peace. Rock Scissors Paper seems to be preferred in Japan.

Round 4
Suffered from irregular grids. Classic-first strategy was better.
I remember that I ignored the range 1~8 in Japanese Sums.

Round 5
I had to hurry in ABCtje, and that caused a minor error.
Logical Kropki is so comfortable.

Round 6
2-cell error in Minesweepers with Duds, and I found another error in Minesweepers. So my true score is 685.
The trouble was that I gave up Magnets on my half way.

Round 7
The rendering should be improved.
In my experience, Shikaku with so many even clues tend to be unique because of the restriction from parity, sometimes accidentally.

Round 8
High-pointers are my favorite. Maybe some of middle-pointers were harder for me.

Round 9
Major error in cheap Countries.
Sudoku always reduce my time in such sets. Some types are still unfamiliar to me.

Round 10
Good difficulty this year.
I thought a spell as "Edison"; I have never heard the N is written/pronounced in Japanese.

Round 11
Tough round. My worst decision was that I tried Foursome and got nothing.

Round 12
Error in Banknotes. I could not finish the last steps of Skeleton.
Large Tapa LITS and Masyu Battleships are especially awesome.

Round 13
No problem, as usual.
I do not know if the logic I used in the small Kropki was intended path.

Round 14
Made a major error.
Belgian V-W Regions were interesting.

Sunday, September 4, 2016

GP and JPC 2016


GP and Japanese qualifying has over. Here are my recaps.



First, about Puzzle GP. My feeling after the first round was that, point distribution was great as if it reflected on the testing time of someone, who may be little a bit bad at Latin types. So in the following rounds, I interpreted points in this way, and this went well in general. And I believe that points do not include typing period.

Though distributions in each round were good, the difficulty of the rounds varied so much, and it should have affected on the "Best 6" choice. I think it is almost impossible to make it uniform.


About Casual section: frankly, I was bored at too many Masterminds in the first half, but thankfully many other types appeared in the latter. I was able to enjoy it casually (, making many errors after the Competitive section).

My question goes to the structure of General section. It was easy to find that time bonus in the Competitive section was "double-counted". Somehow I won this section, but I do not know if this is valid.


In the end, about my result: it is good that I made only one Competitive error, which occurred in the first round. I was not perfect even when I won, but at least there were no major errors in these rounds. On the other hand, my troubles in Round 7 and 8 were clearly one puzzle each: Dominoes in Round 7 and Domino Construction in Round 8. I used surprising 30 minutes in the latter. I need to fix this tendency.

There is no information about the final yet. Now I just hope that I will be in good condition in Senec.



Next, about our national qualifying. This year, our first round was JPC, which included some Sudoku, and qualified players moved on to the selection contests of WPC and WSC. Major troubles were:
- Piece puzzle in JPC was completely same as the one in 2008. I did not know, and I did not expect this kind of trouble.
- Rendering error in WPC selection, which resulted in multiple answers. I found a solution with a wrong assumption.
- Instruction error in WPC selection, which was announced as free points for all players. I did not notice the announce and solved it.

Other than them, I have questions to the meaning of JPC. All contests were online, thus there was no need to reduce players. And I think JPC did not contain enough Sudoku, if this is used as a part of WSC selection.


Let's move on to the result. My ranks in selection contests were, 5th for WSC and 1st for WPC. I did my best even for WSC, but others did better. I am okay with this result. I will use WSC to get accustomed to the environment. So A-team for WSC is same as in London. I believe they will do well.

A-team for WPC is Hideaki, Kota, Taro and me. Taro will return! Everyone won some major world title. I am looking forward to compete with this teammates.


So, let's meet in Senec! Before that, now I am struggling with our puzzle book....

Friday, July 15, 2016

JZC 2016



Puzzle of JZC 2016, held in March, have been published as usual. This year, I prepared Round 7 with Takeya and Yuta. Our round was intended to be Assorted, but somehow it partially became dual round. I wonder how crazy I was to create Filloominoes.


BTW, though I have uploaded my solving video in PGP, there is nothing in Round 7. I failed to start recording. Sorry.
I feel I am not in good shape these days, but still, national qualifications are coming (first round on 16: tomorrow and final qualifying on 30). I hope fair and good contests, and also hope that I can get chances to try for titles in Senec.

Wednesday, February 3, 2016

Toketa? Selection


Here are some notes about my creations for Toketa? Selection.
I made 3 puzzles, which were distributed from the remaining pool (, avoiding some types).

Araf 1:
Prepared to match the harder one. I intended easy and logical path, and a bit of aesthetics as a spice. I was anxious when I got multiple "less than a minute with intuition" feedbacks, but it seems okay in the whole set.

Graffiti Snake 2:
Free creation. Intended first step is "the center cell must be white from topological reason". This is not required, and hard to figure out, but a rewarding back door, at least.

Max Arrow Castle Wall 1:
Again prepared after the second one, and again, my object was easy one. The path is narrow but simple, and you don't need to count more than 3!


I feared that the test was decided by intuitive ones, and there were some exchanges between practice and test. I think it was an acceptable level as a result. For me, the best way to do with intuitive puzzle is to be better at logical part and leave enough time to be intuitive (, or move on to whole-check mode).

Sunday, January 31, 2016

Puzzle Video


GP 2016 has started. I have some random thoughts about Casual/Competitive sections and normalization, but it should be better to keep them until the end. Now I just want to say that I am wondering the meaning of "competitive solver".

Since last year, precisely since the week of UKSC and JNPC, I have recorded my solving in competitions. At first, this is because of interest to myself, but they tell me many things. For example, from this (taken from the latest GP round) I can learn how I made the error. Targeting 70 minutes was enough for me to change my safety-solving into hurry-solving, and to create unknown word.

I don't know this has any benefits to others, but here I try sharing this once.


By the way, Toketa? Selection is held in LMI now. I hope you enjoy the test. Maybe I will write something about my puzzle(s).

Wednesday, November 11, 2015

WPC 2015

I managed to have time to solve all individual puzzles I could not solve in Sofia. Here is my review of this year.



WSC
I competed officially this year. Normally, in Sudoku contest, I do not think so much about my score, and just solve in the order. But my task in Sofia was acting in place of Hideaki. So I needed to consider my solving order, and concentrate more. Thanks to these efforts, I remember my experience better than last year. I will write about WSC, too.

When I was looking for my seat, I noticed that JPN-A was next to EST-A, and even FRA-A and CZE-A were in the near places. It became a funny area.


Round 1
Left: #10-12
My weakness in Sudoku comes mostly from Classics. Especially, I cannot solve the hardest side well, so just started from easy ones. Safe strategy was enough for me. All puzzles were of reasonable difficulty as a result.


Round 2
Left: Cube, XV, Clone, Diagonal, Classic x2
Another Classic weakness is the lack of speed. This means cheap puzzles can be ineffective, and conversely, if I believe that I would be able to complete high puzzle, it deserves to try. So I started from 1234. It took 12-13mins, but still I feel it was good choice. And it was the enjoyable puzzle in itself.


Round 3
Left: Inner Frame, Max Triplet, Classic #2
This time I feared Max Triplet (and this seems correct), so solved in the order. Cheap Classic is good for the last minutes.


Round 4
Due to wrong deduction, I broke it once, but luckily, I could correct it without erasing so much.

This year, in the one-puzzle rounds, time bonus was reduced. This seems reasonable considering the risk of this structure.


Round 5
Left: Greater-than Consecutive, Kropki, Even Sandwich, Next to 9, Mathrax, Classic x2
Bad result, or I just returned to my usual. Everything was not smooth. Maybe I felt tired after lunch. I like Greater-than Consecutive, and Kropki was instructive for me.


Round 6
Left: Toroidal, Classic x2
Ignored Tridoku at first, and returned to it because the remaining time was not enough for a Classic. Parquet was cool.

I feel "Overlapping Sudoku" is an abuse of the word; it is just a Latin square type without any regional rule. I do not care about the appearance of such types in WSC.


Round 7
In my team, we agreed that ranking in the team is 1. Kota, 2. Takuya, 3. Keisui, and 4. me (same as the order in JNPC this year). And the seats were set as 1->3->2->4.

I solved Extra region quite well. I solved it even faster than Kota. And then I was waiting for Takuya.... He broke it, and needed to restart. I was just sitting for a while, and finally it was relayed. And I broke Irregular! Not enough time for repairing, so I sent it to Kota and he managed to solve the following Classic.

After the round, I learned that the puzzles were actually the same, only numbers were exchanged.


Round 8
In the morning, I heard that I was in the good position for playoff. I had never expected such a chance, and to be honest, it was the first time for me to think about that borderline.

So I decided to start from Twin Region, and solve in the descending order of points. This strategy worked quite well as a result.

In Twin Region, converse sequence was allowed, but it is impossible in fact. This was helpful.


Round 9
I made many errors. Somehow I corrected them all and finished. This was of importance.

I hope gray cells are painted in lighter color.


Round 11
Central part was broken, and Kota repaired it. While repairing, the other 3 were checking, and found that I had made 2 errors inside 6x6. Ah.


And I was in the 8th place. Anyone expected? Kota said that this year there were more puzzlish Suodku. This seems right. I do not mean only newer types were puzzlish. Common types were also on the puzzlish side. This is why I was comfortable, I think.

Also, I made no wrong submission in WSC. I was lucky. I do not expect better results in the future, and even B-team is okay next year. I just want to solve puzzles in WSC.

I am happy with my team result. I felt I achieved my task.


Playoff
Kota: Arrow in 4th - He decided to choose Arrow because of SGP Final.
Tiit: Diagonal in 1st
Jakub: Even Sandwich in 3rd - No problem for Kota, as he agreed later.
Bastien and Seung-Jae: Irregular in 2nd and Killer in 3rd - I forget who chose which.
Michael: Greater-than Consecutive in 1st - It was impressive and I remember it.
Cheran: Extra Region in 2nd
Prasanna and I: Surprise in 3rd - We had same points, surprisingly, and chose the puzzle together. Surprise was attractive, and one more reason for recommending it was to reduce the risk in the final, as a teammate of Kota.
Kota: Inner Frame (remaining) in 2nd
And Surprise was revealed as Non-consecutive.

This was the bonus stage for me, just for amusement. I made fool step in Extra Region (it was Windoku in fact), and while restarting, I was even wondering "why I am here?" Good experience, anyway. Prasanna won the round, so he should have made a good puzzle selection. Later I solved Surprise and it was actually good surprise.

Congratulations on Kota for another title.


At that time, WPC instruction was still not published. I noticed it during sightseeing. If I could have read it beforehand, I had been to the museum.


PGP Final
Killed by the first one again, killed by the word puzzle again. It was too large and tricky for me to grasp. After that I did as usual, but I made a wrong deduction in the Cave and made an error. First error in Sofia. Still, I managed to reach the last puzzle (for the first time in playoff), a bit later than Matus. And Find the Pair benefited me. 3rd place is much better than nothing, like last year.

After the final, I noticed that there were puzzle names on the result table. Were they published before to start? I was always predicting 2 types (except Neighbors).

By the way, after SGP Final, Kota recommended me to prepare a binder because the table mat was uncomfortable. Good advice.


WPC
Like WSC, JPN-A was next to USA-A, and NED-A was near. It was attractive, but Kota chose that seat, so I chose near Ulrich this year.


Round 1
A strange bag was on the table. Puzzles were not normal, as expected. I solved in random order and checked carefully after solving each puzzle. Tapa was 2nd for me, and I left it for the last when I noticed it was not unique. I had more than 20 mins when I returned to Tapa. The problem was, if Tapa was valid, another error means no partial credit. So I checked all puzzles again, and fixed an ambiguous place in Fillomino (in dividing puzzle, dashed line is much better, by the way). That was all.


Round 2
Left: Clouds x2, Triangular Minesweeper (2nd)
I broke Coral once, but I did not notice missing 1 in second solve. I was lucky.

Clouds and Triangular Minesweeper was left for the last, and I did not have enough time. One more minute was enough for 2nd Clouds, though.


Round 3
Left: Chaos(2nd), Encrypted Square x2, Kropki (2nd), 2D Magic
Error: Doppelblock (1st)
The error is because I erased too much while solving 2nd one.

I solved in the order, and after I broke hard Chaos, I moved on to random solving.

I showed my weakness - same as in London. I could not believe Ulrich's finish at first. Long way to reach that level.


Round 4
Left: Worms (2nd), Striped Snake (2nd)
The result seems good, but I was not so comfortable. For example, I broke Cave.... Again!? And Worms (1st) was solved in the last seconds. Just lucky.


Round 5
Solved: Japanese Railways x2, Easy as Tapa (1st), Snake in the Naval Forest
After Japanese Railways, I went to Easy as Tapa, but it did not work at that time. Then I started Snake, broke, erased, and broke again. This was in the must-solve list for me, so I was still thinking about it. And there was the announce. I was forced to restart with that change. After that, I was hurried with my bad performance, and broke Easy as Tripod/ABCD123 twice each. At last I returned to small Tapa. This time it went well.

The puzzles themselves are pretty good. And I know that I should have kept my pace regardless of the error.


Round 6
Great construction. When I got ready to claim, I heard Ulrich's voice. Did not notice Zoltan at that time.


Ro
und 7

Tapa by Kota, Fruit Basket by Ko, Akari by Maho, and Pentomino Areas by me. Ko completed faster, as planned, and he joined in Pentomino because the other 2 said they can solve by their own within the time limit. I did not make much progress at that time (around 10-20%?). A bit later, I started to help Akari for secure. After Tapa and Akari were solved, we started to paint (by the way, some of us broke crayon - I hope better tool). After painting, we gave up Pentomino with the little remaining time, and moved on to checking. Sadly, an error in Akari was looked over - it was too large.

Pentomino was solved around 30-40%? in the end. It caused many wrong deductions. I have tried it again, and I suspect that it has many solutions.


Round 8
Left: Puzzle2 and 6
I skipped Puzzle 1 - I did not notice that letter cells must not be used. And skipped Puzzle 2 without any notice (or just thought "triangle to hexagon?"). Puzzle 3 and 4 went well, and Puzzle 5 also worked well, though I feared that maybe diagonal touch is allowed. Puzzle 6 was hard to figure out. At first, I thought that the arrows indicate multiples of 9. In this sense, 9 was not a good example. Then I returned to Puzzle 1, and understood the rule. One problem was that all blank areas are also used once in the example. And I retried Puzzle 6, noticed that it is not multiple but sum. I tried the 1st one with target number 9 and broke it, then noticed that target number can be changed. But no time to retry.


Round 9
Solved in the order, just left Snake for the last. Tapa was solved by intuition. I made some errors in Snake, but I managed to finish.


Round 10
Left: Scissors x2, Mars and Venus (2nd), Cards (2nd)
Error: Checkerboard (1st), normal error in the final step
At first I skipped 4 types above. Nearest Tetromino was by intuition, it worked strangely well. Checkerboard was the most uncomfortable type for me. I should have done the others. I saw Scissors/Mars and Venus for the first time, and both were interesting.


Round 11
Left: Encrypted Skyscrapers, Tetroscope x2, Puzzle Knight x2, Encrypted Square
Error: Different Neighbors (1st), 2 missing digits
Solved puzzles were almost done in order. I tried Puzzle Knight in the end, expecting some points, but it was in vain.


Round 12
Left: Coriolis, Unequal Star Battle
My order was Yajilin, Rectangles, Pentomino, Domino, Different Neighbors, Magnets, Hula-hoop and Kakuro. Kakuro was completed in the very last seconds. If I did not break Domino, maybe I could have solved one more, but it seemed impossible to finish.

Every time I see "hard" Yajilin, it just requires some deep assumption. A bit boring. I used the same tactics in Different Neighbors and Magnets during the competition.


Round 13
I was the faster in my team, though not smooth, and helped Kota because he had broken his puzzle and restarted. Before we completed it, the other 2 finished theirs and began the final puzzle. Later we joined in it, and we alternately checked the small ones. I also checked one, but I feel I should have concentrated on the large one. I contributed little on it. It broke in the last steps, and we could not repair. About 5 pieces were swapped, as a result.


Round 14
Minesweeper and Tents by Maho, Arrows and Snake by Ko, Sudoku and Kakuro by (of course) Kota, and Different Neighbors and Slitherlink by me.

All clues in Slitherlink were revealed soon, and I solved it at first (in fact, it was wrong, and later I was asked to repair it - how stupid I was to make mistake in such an easy puzzle!). Next, I finished Different Neighbors without a clue, and helped Tents and Mineseeper. Both had enough clues then. The main problem was that we broke Arrows, and it affected on other puzzles. At that time, puzzle distribution was, Arrows by Ko, Sudoku by Kota, Kakuro by me and Snake by Maho. I had little progress until some clues were filled. And even after all clues were revealed, it did not work. Only Kakuro was remaining. It was because of a wrong clue which was filled in during previous wrong deductions. I could not find it. From the puzzle design, it was hard to make notes, hard to find errors. It was preferable to use light gray in clue cells, not black as usual.


Playoff
Palmer was in 3rd place, so I decided not to choose Tapa, for example, because he would choose that type. Playoff structure in these two years is good in that it assures certain ranking even if falling down, like me in London. But I think puzzle choice can be affected by the set of members in that round, which may reduce the advantage of preliminary rounds.

When I was sitting alone, Bram and Zoltan came and I heard that puzzle list had published:
2D Magic
Cave
Chaos
Coral Maximum
Different Neighbors
Dotted Snake
Fuzuli
Hexaislands
Kakuro
Kropki
Magic Summer
Pent It Black
Polygraph
Rollercoaster
Surprise Puzzle

Bram predicted that the number of puzzles would be 3/5/7, like in London. That sounds right, and then finalists would choose 3/2/2 each. I had no reason to choose Latin types. I hated Cave because I had broken that type twice in Sofia. I also avoided Polygraph because I felt it has similar aspects with Cave. And Surprise seemed a passive choice. I wanted to solve Coral, but I thought that Palmer would choose it if he can select 2 types. Thus, my first choice was Pent It Black at that time.

At the beginning of the selection, we were told that the number of puzzles were 3/4/5 in each round. It seemed that I can still choose 2, but only 1 for Palmer.

Ulrich: Kakuro in 4th - It seemed reasonable.
Me: Coral in 5th - I feared Palmer would not choose it. He told that it was actually his first choice. I felt I made a bad decision, and...
Palmer: 2D Magic in 1st - What?! I had not expected that choice, and my previous selection got worse. Also, I had not looked at this type in the preliminary round.
Kota: Rollercoaster in 1st - Quite reasonable.
Bram: Pent It Black in 3rd
Ko: Hexaislands in 2nd
Zoltan, Yanzhe, Nikola: Polygraph in 1st, Kropki in 2nd, Magic Summer in 3rd - I forget who chose which.
Ulrich: Chaos in 2nd - Again reasonable.
Me: Different Neighbors in 3rd - Because basically I feel comfortable with this type. Risky choice, though.
Kota: Fuzuli in 4th - Easy to predict.

Then I was waiting for my turn. 2nd Round was surprising. The reason why I did not choose Hexaislands was that I thought it would be easy and make little difference. Later I solved it, and it took an hour. I do not think this "type" does not suit playoff, but I believe "that puzzle" was too hard for playoff. I heard that Surprise was instructionless, and I think this "type" would not match playoff. Anyway, I was relaxed looking at the situation.

And the final started. See also the scoreboard by Takuya.

While waiting, I was counting time in mind, and my 2:51 counting was quite accurate.

2D Magic had good theme. 1-9 appeared once each, in order. I just kept in mind not to forget 0. I made many small errors, but every time I corrected them soon.

Next was Chaos. I feared this, but it went smoothly after a breakpoint.

Different Neighbors had good logic, which I had searched for in some puzzles but never worked. I was a bit lucky.

And Kakuro.... At least I have some experience as a Japanese. In Japan, Kakuro is also called as Sum Cross. This represents the difference of publishers, but it also shows the difference of logic. Kakuro in the final was rather Sum Cross, where addition of some rows/columns works effectively. I was lucky again, to solve this quite smoothly. And the air moved when I submitted this. I noticed something happened, probably I overtook Ulrich.

The last one was Coral. At my first glance, I cursed my selection - it seemed no start point. Actually, there was a place to start, but I made an error in the last steps. Worst error in the final. I needed to restart and lost a minute, but it was still enough. When I submitted, the air moved again, but I thought maybe I made a mistake, and even prepared for that case. More than a minute had passed, but there was no return, so I asked Vladimir, my marker, and then I knew I won. I did not look at the scoreboard until then.

Thank you everyone concerned, especially to Yuta and Kota, who invited me to LMI in 2011. My puzzle world was small then, just knew Nikoli. My first contest, Fillomino Filla, was wonderful, and impression at that time was why I could endure no-championship period in 2011-2012. I hope I will be able to return something to this community.

Still, I understand that I just won the playoff. It seems so hard to perform like Ulrich in preliminary rounds. Long way to go, but I will just keep enjoying.

Thursday, October 8, 2015

Hitori


Notice: Toketa vol.3 will be published. See here.



Hitori is the puzzle appeared in Nikoli in 1990. Classical Hitori was only searching puzzle; looking for certain starting points, and searching the row/column of unblackened cells. That was all. It is often said that Hitori is not in favor because of unnecessary clue cells, but I think this is not bad when regarded as searching puzzle.
On the contrary, modern Hitori is rather logical. Especially, in harder puzzle, we need to consider global connectivity, and I love Hitori with this logical deepness. Here, the authors tend to try to make searching easier, and unnecessary clues can actually be unnecessary.

Here is a (modern) Hitori, it would be better not to fill in empty cells....


solve in PUZ-PRE


Most Hitori I have solved in contests are classical searching ones, where noise clues are valid. I want to solve logical Hitori more.
By the way, Hitori Round in Logic Masters 2014 was great. It suggests some ways to resolve problems about Hitori.