Nurimisaki Nomad (17th Sep - 7th Oct) has started Discuss
Hybrids - LMI March'11 Puzzle Test56 posts • Page 3 of 3 • 1 2 3
@ 2011-03-15 1:33 AM (#3733 - in reply to #3653) (#3733) Top

Para



Posts: 315
100100100
Country : The Netherlands

Para posted @ 2011-03-15 1:33 AM

So, my first try at designing a test is over. It was interesting to see what happens behind the scenes. It was a fun experience for me and glad to see people enjoyed the test too. Thanks for the comments, they are appreciated.

Congrats to Deu for finishing first and MellowMelon and Motris also finishing on the podium.

A little bit too bad no-one managed to solve all puzzles. I guess working with the same puzzle types eventually leads to a bit of an underestimation of the difficulty. Many puzzles were second draft. Many of the first draft pzzles were much harder than what you some solved now. Especially the Hard Corral, Hard Worm Sudoku and Seethrough were made easier. That might be part of the reason I underestimated the total solving time in the end.

The solving stats are mostly what I expected. I had anticipated that the Hard Worm Sudoku, Outside Skyscraper and Sea Battle would probably get the least solves. I am a bit confused on the low solve rate for the Seethrough. I can only guess that the logic involved wasn't properly grabbed. I'll try to show later how the design was supposed to go.
The outside skyscraper, it could have been a bit too hard. When looking at some solving entries, this puzzle might have hold up some puzzlers and prevented the solving from some other puzzles. The logic path should have probably have been a bit less restricted.
The Hard worm Sudoku got the amount of solves I had guessed. At first I had made 9by9 grids. They work but were just too hard for the test. It was probably because 18 clues on a jigsaw sudoku already pretty tough, let alone with an unknown grid at the start.
The low solve on the Sea Battle Hard might have also been the missing of some logic. The easy sea battle has a minimal structure: N/2 areas for N ships. This forces no ships to cross any bold borders. The Hard one, like the example, is the other way around. There are 14 areas for 28 ship parts. This forced all ships to only lay across bold borders as each area had 2 ship parts from 2 different ships. This logic severely limits the amount of places the size 4 ships can go. This was picked up by the testers, so didn't think it would be a problem. I'm not sure if this was the actual problem when solving, but it's my best thought as these are really the key realisations you have to have about the genre.

I also expected the puzzle ratings to be about the same. My own favourite hybrids to make were the Pentopia, Outside Skyscraper Sudoku and Worm Sudoku. I understand the Alphanumeric Dice puzzles might not everyone's favourite, but it's just a type of puzzle I always enjoy solving.

Some comment answers.

The reason there was a bonus section was obviously that the test was intended to have 6 sections for 24 puzzles. The second to last draft still had this and I thought it might be okay at first, but especially with the ABC in the one section I figured it might be too much. The ABC isn't very speedsolving friendly because of the sum sizes. I thought the hybrid was a good way to use the logic in a more speedsolving friendly way as the sums are much smaller. After seeing where the problems with MellowMelon's test lay, I was pretty much convinced it would be too long. I wasn't 100% sure how to adjust things though as I liked to have this section in the test in some way, because it was different from the puzzles normally seen in puzzle tests. They are both variants that are featured in Breinbrekers, but not seen much anywhere else. Eventually after some dicussion with Deb, we decided to build in this section as a bonus section instead of a time bonus. We both felt it was a section that did add something extra to the test. It might have been a bit smarter to make the section more clearly marked as being a bonus section in the Puzzle Booklet. This was sadly only realised when we saw that people started solving the puzzles. I decided to still award the points to people who solved them as they didn't affect the top standings.

As for the bonus system, one of the intentions was to get people to also attack the hybrids as solving the hybrids gets you more points when you've already solved a few regular puzzles on that side. I went with percentage bonus to give people more bonus for solving higher scored puzzles. The percentage discrepancy was done because I felt otherwise the total bonus for the lefthand side would turn out too low. Now the total acchievable bonus for the lefthand side was 140 points and for the righthand side 225. I thought these totals were about right, found 105 a bit too little for the lefthand side and 300 a bit too much for the righthand side.

Some puzzles could have probably done with some easier answer keys, but I somewhat wanted to have answer keys that force you as much as possible to have to solve the whole puzzle. Sometimes this does make the answer key a bit more diffcult than it needs to be. For the Loop puzzles I just borrowed the answer structure used in FLIP. The Sea Battle I didn't go for the size 1 ships as they didn't weren't really the end of the solving path.

And I'll think about including some easier examples in the instruction booklet next time. They just ended up being some interesting smaller designs I came up with while testing the puzzle types and wanted people to see them.

Lastly

Some last puzzle related comments. The original ABC design didn't include the Sweden and Norway sums. So if you haven't tried it yet and want a bit more of a challenge for it, you can remove those two sums and tackle it that way. Also I wanted to point somethign out about the Outside Sudoku Design. If you look at the top clues you'll come across 4|8|32|3|16|48|2|9|18. If you look at them per three, you'll see that 4x8=32, 3x16=48 and 2x9=18. You can do this on all four sides of the puzzle.

I tried a bunch more hybrids, but eventually thought these were the ones which I could make the nicest puzzles with, while incorporating things from both puzzle types into one puzzle. One other combination that worked well is using both Myopia and Slitherlink clues in the same grid. It can help take out some uniqueness issues from Myopia and steer clear from some regular slitherlink patterns. But I thought the Pentopia was a nicer one in the end.

The unused puzzles that are supported on puzzlepicnic, will appear there in the upcoming weeks. I will also post some leftover designs there from the Hybrids there.you can check them out there in case you're interested in them.

Edited by Para 2011-03-15 2:02 AM
@ 2011-03-15 4:31 PM (#3734 - in reply to #3653) (#3734) Top

anurag



Posts: 136
10020
Country : India

anurag posted @ 2011-03-15 4:31 PM

Brilliant puzzles from Bram.You always make great ones.It is a pity i attempted it without a printer.I loved the star battle construction.Fences was far straight-forward,though it looks nice.
You should keep organizing more of these.
@ 2011-03-15 4:36 PM (#3735 - in reply to #3653) (#3735) Top

anurag



Posts: 136
10020
Country : India

anurag posted @ 2011-03-15 4:36 PM

And yes,dice isnt one of the liked ones.Some of us,including me,probably didnt even try to look at it.
@ 2011-03-16 5:29 AM (#3739 - in reply to #3735) (#3739) Top

tuz



Posts: 3

Country : malta

tuz posted @ 2011-03-16 5:29 AM

hi, great puzzles. however i think pentopia hard puzzle is wrong. in the down right corner, "-" does not equalise U and L shapes. there is no way to put L shape. it is wrong in the solution booklet too. the puzzle is still solvable until 2 or 3 shapes left, but it is so.

Edited by tuz 2011-03-16 5:38 AM
@ 2011-03-16 6:21 AM (#3741 - in reply to #3739) (#3741) Top

Valezius



Posts: 66
202020
Country : Hungary

Valezius posted @ 2011-03-16 6:21 AM

tuz - 2011-03-16 5:29 AM

hi, great puzzles. however i think pentopia hard puzzle is wrong. in the down right corner, "-" does not equalise U and L shapes. there is no way to put L shape. it is wrong in the solution booklet too. the puzzle is still solvable until 2 or 3 shapes left, but it is so.


The sign in R5C11 and R9C10 aren't same.
The sign of R5C11 shows to right and left, and the other one points at left only.
So the official solution is correct
@ 2011-03-16 8:04 AM (#3743 - in reply to #3741) (#3743) Top

tuz



Posts: 3

Country : malta

tuz posted @ 2011-03-16 8:04 AM

Valezius - 2011-03-16 6:21 AM

tuz - 2011-03-16 5:29 AM

hi, great puzzles. however i think pentopia hard puzzle is wrong. in the down right corner, "-" does not equalise U and L shapes. there is no way to put L shape. it is wrong in the solution booklet too. the puzzle is still solvable until 2 or 3 shapes left, but it is so.


The sign in R5C11 and R9C10 aren't same.
The sign of R5C11 shows to right and left, and the other one points at left only.
So the official solution is correct


alright now i see it. it is not full but half of a line. ok now it has a solution : ) but admit, it is a little unclear, i spent 20 minutes to find out where i did go wrong. but for different shapes, different colours can be used. maybe print outs won't show, however a lot of people use photo editor. it could be better. anyway thanks for the reply. and one more thanks for the puzzles.
@ 2011-03-19 11:52 PM (#3789 - in reply to #3653) (#3789) Top

Para



Posts: 315
100100100
Country : The Netherlands

Para posted @ 2011-03-19 11:52 PM

Posted some of the remainders here.

http://puzzlepicnic.com/community/posts/list/1342.page
Hybrids - LMI March'11 Puzzle Test56 posts • Page 3 of 3 • 1 2 3
Jump to forum :
Search this forum
Printer friendly version