| Author |
Message |
tails
Joined: 10 Jun 2008
Posts: 191
Location: Tokyo |
|
 Great puzzle! |
 |
Not only because the hex board looks beautiful :-)
At first glance, this puzzle looks easy, and it is acturally easy under level 100 or so. But it's getting harder around level 140. I thought it would be hard if the board got very large, but suprisingly it's getting hard while it keeps smaller than the size of level 0 :-)
Furthermore, it's a great aspect of this puzzle that we can't simply apply well-known algorithms when we make solvers. It would take forever with DFS even to solve low levels. My simple solver took less than 1 second for each of low levels, but it got almost useless above level 100. It seems that the key is how to implement human intuition.
I never imagined the hex board had such possibilities. It' really fresh! And I never imagined JavaScript had such capabilities :-)
|
|
| Mon Aug 31, 2009 2:08 am |
  |
 |
adum
Joined: 19 Apr 2007
Posts: 389
|
|
| |
|
glad you like it! =)
it's been a long time since i put up a new puzzle, so i thought i'd release one and try something a little fresh at the same time. a hex board is a little different, but the nature of the puzzle is really a new beast. as you say, classical brute force algorithms aren't going to get very far!
i honestly have no idea myself on how to tackle this one, so it will be very interesting to see where people go with it...
|
|
| Mon Aug 31, 2009 5:22 pm |
 |
|
therethinker
Joined: 28 Mar 2008
Posts: 144
Location: #hacker.org on Freenode |
|
| |
|
Very awesome puzzle.
Hex board is fun, yet a PITA to work with (in a good way  ) As tails said, its amazing how well this puzzle scales.
How are you generating the puzzles? I was doing some by hand and noticed a few (around level 30, when they were harder) that consisted of only 1 snake and took 2 moves to solve. I guess it averages out, though.
Props to the first person that uses Javascript to solve it
|
|
| Mon Sep 07, 2009 9:05 pm |
 |
|
adum
Joined: 19 Apr 2007
Posts: 389
|
|
| |
|
for generating the puzzles, there are a bunch of factors to play with, such as board size, number of worms, num colors, worm length, spaces left over, spaces blocked, etc. i don't know which will turn out to be hardest, so i try a lot of variations of all these factors at each board size. some will be harder than others...
good luck with your solver! =)
|
|
| Wed Sep 09, 2009 5:21 pm |
|
|
therethinker
Joined: 28 Mar 2008
Posts: 144
Location: #hacker.org on Freenode |
|
| |
|
Also, does everyone have the same problems?
There really shouldn't be an issue of sharing answers since solvers/code could be shared too...
|
|
| Thu Sep 10, 2009 1:45 am |
|
|
adum
Joined: 19 Apr 2007
Posts: 389
|
|
| |
|
yes, everyone sees the same problems for this puzzle.
|
|
| Fri Sep 11, 2009 12:01 am |
|
|
prop
Joined: 16 Jun 2009
Posts: 6
|
|
| |
|
Wow it's very intriguing puzzle! level 60-65 were the best I think(Now I'm solving lv.74, so I know just a little of them).
Anyway I like to solving it w/o brain though  plus lv 73 was not THAT horrible. i think lv 7 was the most horrible 
_________________
HI! |
|
| Thu Oct 29, 2009 6:30 am |
|
|
dangermouse
Joined: 05 Jun 2011
Posts: 72
|
|
| |
|
lvl 73 is horrible! i had a bad start with this puzzle, i tried to implement it as a iterative DFS but had some bug in the implementation and had to give up. I tried to do it iterative knowing that the recursion depth of this thing is awesome. I then implemented it as recursive DFS with weighted moves on shortest paths and still can't solve level 73...
one good thing i am using is: to debug the code i inject into a html level template my current board as javascript variable, and then look at the result in a browser as it would be a level on hacker.org. This probably works with other puzzles as well, but i never tried it before.
and still i have some ideas, but its' really an awesome puzzle 
|
|
| Wed Aug 29, 2012 1:58 pm |
|
|
|
