I am enjoying the Witness so far (on kind of a hiatus at the moment) but given what you like about Myst as distinct from the Witness you really must play the rest of the Myst series, especially Riven, its pinnacle.
So I cheated on a panel and I am so mad:
In the Swamp area, in the turquoise pond observation room, the forth panel is
(A) x
xxx
(B) x
xxx
(C) xx
x
x
(D) oo
oo
----------------ā¢
| | |(D)| |
-----------------
| | | | |
-----------------
| | | |(C)|
-----------------
|(A)| | |(B)|
O----------------
So the precedeing puzzles had trained me to think of a shape made up of A + B + C - D which covered the four blocks marked with their legends. But I said, wait, I have to cover 4 + 4 + 4 - 4 = 8 blocks, and I need to cover those four with legends, and thereās only two contiguous 8-block shapes which do. So I drew one, and I was right.
Iām mad because I canāt see the ārightā solution.
I got mad at the exact same puzzle. Hereās the rule I eventually came up with that let me make peace with it (huge spoilers for people who havenāt done the swamp Tetris area to completion):
[spoiler]The trick is that you can overlap yellow pieces (A, B, C) as long as the blue piece (D) is on the overlapping parts.
Think of each square in the puzzle having a running total of the number of blocks overlapping it. For each yellow piece on a square it adds 1, and for each blue piece on a square it subtracts 1. After laying all the blue and yellow pieces, each square must have a total of 0 (meaning empty; not part of the solution) or 1 (meaning it IS part of the solution). If any square goes below 0 or above 1, then the solution is invalid.
There might be a more elegant way to state this rule, but thatās essentially it. I donāt think thereās a puzzle in the game that goes against the rule as Iāve stated it.[/spoiler]
1 Like
OK, works for me.
Yeah, there might be no simpler formulation than that running sum model. To extend āIf any square goes below 0 or above 1, then the solution is invalid.ā to the complete polymino rule, I would instead say āIn each separated area, every square must sum to either 0, or 1.ā (i.e. mixes of 0 and 1 are also disallowed).
1 Like
Yeah, thatās a nicely stated way to extend that to explain the entire ruleset.
It got me thinking about the complete algorithm for tetromino solution checking. Hereās the simplest complete formulation I can come up with:
[spoiler]Compute the set of all permutations of position and rotation (for rotatables) of tetrominos and nega-tetrominos, irrespective of the playerās line. For each permutation, add 1 to a square for a tetromino subblock and subtract 1 for a nega-tetromino subblock.
Exclude all permutations with any square that has a value less than 0 or greater than 1.
Now considering the playerās line, exclude all permutations that have a mix of 0 and 1 in any separated area.
Exclude all permutations that have a value of 1 in a square that is separated by the line from the originating tetromino.
Exclude all permutations where an originating nega-tetromino is separated by the line from the originating tetromino(s) it canceled some subblocks from. (Itās fine if the canceled subblocks themselves are across the line.)
If there is one or more permutation left, the puzzle is considered solved.[/spoiler]
I imagine that resembles what the game code is actually doing to check your solution, although with optimizations to avoid computing obviously impossible cases.
I think most of the puzzles only have one solution, and for the others with non-trivial multiple solutions they would have brute-forced the solutions and checked the playerās line against the list. Your algo is legit for checking.
Finally found the right solution for the bottom of the desert temple. I still have castle, zen garden, and village to do. The mountain is open and itās annoying af!
1 Like
iām curious if everybody agrees that most puzzles only have one solution. i remember feeling like a good chunk of them seemed like they had multiples, particularly in certain parts of the game - besides things like switch panels and other obvious stuff - but of course i donāt really have much to back this up.
There are some funny-looking eucalyptus trees here (because it rains)
There are a lot of puzzles with trivial variations available on how your line moves in open space, or that are symmetric (although most of the symmetric ones are used as switches, yes). I donāt think there are many with multiple ānontrivialā solutions, but there are at least a few that I can think of (alternate solutions are required for the obelisk solving the life-sized-line-puzzle ones, due to the obstructions).
2 Likes
Also (spoilers about the nature of the challenge): the challenge puzzles are procedurally generated. Blow admitted in an interview that he wrote algorithms to generate them; itās not just pulling from a list of puzzles.
Oh right, I forgot about the more free-form puzzles. I was thinking mostly about the tetromino puzzles.
disappointed nobody has made a āmore like the whitenessā post yet
3 Likes
I broke down and used photos to solve the village flower puzzle. I can see how to solve the church lattice panel, but I canāt open that stupid door!
Zen garden was short, guess itās supposed to be an introductory area. Castle was underwhelming, lit the laser with only half of it finished?
I think The Witness failed as a game since I donāt see any reason to finish the game, the mystery of the island isnāt a compelling mystery because I know thereās just going to be more puzzles.
I am operating in near complete ignorance but this is, in my head, what has kept me away from this game.
1 Like
āJust going to be more puzzlesā is not wrong, but I guess I donāt see why itās that much of a failure. There are some pretty damn fresh and satisfying puzzles still ahead, I assure you. And the game did foresee players getting exhausted and stopping at the regular ending of the game, so I donāt see that the goal of the experience is defeated if that happens for you.
The Witness is partly about mysteries and revelations yes, but mostly about the small ones. To answer a question I posed at the very beginning of this thread, I feel it adds up to more the sum of its parts, but in a quiet, accumulative way.
Well, I started on the mountain, and the puzzles so far have not been any harder to solve or different from previous puzzles, just harder to view OR get close enough to activate. Theyāre far less enjoyable, and the carrot of the mysterious models & diagrams just isnāt enough of a draw.
I feel like, if I stopped playing now, I would have understood the game & not miss anything significant. Thatās a failure for a game with a clear ending + post-ending.