Kolzig

The Witness by Jonathan Blow

Recommended Posts

I made a lot of angry posts on here but, looking back on it, I had a lot of fun. I regret buckling and reading a few spoilers, but that's a problem with my will power, not with the game.

Share this post


Link to post
Share on other sites

Here's my take on the blue tetronimos... I'm still wary of spoilers so I apologise if this is a repeat of earlier discussion

--- Each blue square is equal to -1 yellow square --- If a blue square is in the same area as a yellow shape, it cancels out one yellow square --- If there are fewer blue squares than yellow squares in the area, that area must take the shape of the remaining yellow squares The next point is a major spoiler for the last type of puzzle, so I'll put in another spoiler!

--- If there are the same number of blue squares as yellow squares, the area can be any shape as there is no yellow left to worry about

 

I've been enjoying my time with the game, and have 6 lasers now. For some reason, I'm tempted to play a certain other game, now that it's just become available ;)!

Share this post


Link to post
Share on other sites

Twig was right about the rule

It is really weird that you have to stack the blocks in that one puzzle.

Share this post


Link to post
Share on other sites

Vindication!

 

I'm almost convinced that one puzzle is just a bug, to be honest. X: Like, to the point that I feel I should warn people ahead of time that it's inconsistent with previous puzzles in a way that's not obvious or ever taught to you. I mean, on the other hand, we all solved it, so I dunno. Life is rough!

 

---

 

+ puzzles question, tutorial area

 

There's an arrangement of yellow flowers that, if you line it up correctly with the activated panel behind it, becomes an activatable line. But I can't for the life of me get it to actually activate! The solutions aren't working.

 

There's also a pillar with an arrow puzzle on it, which I solved. It revealed a large floor puzzle next to the flower arrangement. At first I thought solving that puzzle so that it ends touching the flower arrangement would allow me to jump from one to the other, but it doesn't work. Presumably because the floor puzzle is white, and the flowers are yellow.

 

...Hint, please?

 

 

I am still crawling through the game (in fits and starts, probably 16-17 hours in with 210 + 9), and determined to make my way through that godforsaken marsh area. I'm on the very puzzle that Twig was talking about on this page of the thread, and spending most of my 3am headspace on it.

 

I've gone through periods of intense frustration, but I also think I'm starting to truly fall for the game. When I was at the laundromat Sunday, I pulled out my phone and studied screenshots I had taken of panels the night before. For better or worse, The Witness has me.

 
Good luck! That puzzle destroyed me, and I only solved it because I got lucky with a random guess!

Share this post


Link to post
Share on other sites

+ puzzles question, tutorial area

 

There's an arrangement of yellow flowers that, if you line it up correctly with the activated panel behind it, becomes an activatable line. But I can't for the life of me get it to actually activate! The solutions aren't working.

 

There's also a pillar with an arrow puzzle on it, which I solved. It revealed a large floor puzzle next to the flower arrangement. At first I thought solving that puzzle so that it ends touching the flower arrangement would allow me to jump from one to the other, but it doesn't work. Presumably because the floor puzzle is white, and the flowers are yellow.

 

...Hint, please?

 

Make the colors match.

Share this post


Link to post
Share on other sites

I figured it would be something like that, but I don't know how to change the colors.

 

I'll look closer when I get home, see if I can spot something else.

Share this post


Link to post
Share on other sites

Ok, I went back and found the other puzzle showing how the blue squares work.

Z7FAFll.jpg

 

First you just put the square shape on top of the L shape (not stack them) and remove the 2 blue squares from the top part so the disjointed parts collapse together.

 

In the 01 notation (the red 1s are being removed)

 

0110

0110

0010

1110

 

While thinking about it a bit, it can actually be explained by BOTH "my" theory (i.e. blue squares remove the yellow squares so disjointed parts of the shape come together) and the stacking theory. They are pretty much equivalent, you just remove the shape and then move the disjointed parts together vs. you move the parts to their final position with stacking and then remove the extra squares. I still like the version without stacking more as it sounds much simpler to me and nothing else in the game requires thinking about stacking.

 

Here's the stacking approach

 

0000

0110

0120

1110

 

Anyway, both point of views work, but there are no bugs in these puzzles and the shape of the blue elements absolutely do matter, why would these shapes differ between puzzles?

Share this post


Link to post
Share on other sites

Hmm, you're right. I must've intuitively solved that one with the mindset that I'm breaking one piece and then combining them.

However the initial puzzle I brought up DOES require stacking, because you're removing squares from two pieces at once.

Unless you can offer me an alternate explanation. Please do! If not, I stand by what I've been saying all along, that that puzzle is broke as fuck. If not programmatically, then at least in terms of design, because the puzzle you're showing doesn't REQUIRE stacking, so it's teaching the mechanic in a bad way.

Share this post


Link to post
Share on other sites

I think I already did it previously, I'll add one more step:

 

__0011 -> __0011 -> 0011

__0001 -> __0001 -> 0001

__1101 -> __0001 -> 0001

111111 -> 110011 -> 1111

 

The only strange thing here is the original shape being partially out of the grid but I don't see it as a problem? Because it's the final shape that has to fit there.

 

The same for the puzzle from my previous post:

 

0110 -> 0010 -> 0000

0110 -> 0100 -> 0010

0010 -> 0010 -> 0110

1110 -> 1110 -> 1110

 

But yeah, I can agree it may be teaching the mechanic wrong as the puzzle from my previous post is simpler, but you'll encounter it much later than the other one.

Share this post


Link to post
Share on other sites

Vindication!

 

I'm almost convinced that one puzzle is just a bug, to be honest. X: Like, to the point that I feel I should warn people ahead of time that it's inconsistent with previous puzzles in a way that's not obvious or ever taught to you. I mean, on the other hand, we all solved it, so I dunno. Life is rough!

I solved it thinking about it the wrong way though. Not sure I would've tried to stack them honestly. It's weird that I never had to apply the correct rule elsewhere. Maybe I'm just not far enough in the game.

@ iax

I think it's different. In your example you don't need to think of the cube as being anywhere before it has two pieces removed. You can just say, remove two pieces in the cube, then try to fit it within the line. You could say they overlap, but you don't have to. Therefore I would never think that's how the game is treating it. Similarly you don't have to think of it as the game putting them in a specific configuration, deleting pieces, then moving them again. In twig's example puzzle the positioning of the pieces matters a lot and you have to think of it in this convoluted way that's never hinted at in other puzzles.

Share this post


Link to post
Share on other sites

What eot said!

 

To be blunt, I don't really buy your proposed solution with the sliding around of pieces. X: Much like the overlapping pieces, it's something never used again (I think I can safely say this now that I'm at like 515 puzzles solved or whatever). I'd buy overlapping pieces before I'd buy putting all the yellow pieces together, deleting a chunk, and then sliding them around. Although, admittedly, it's effectively the same thing here, it just feels even MORE unnecessarily convoluted.

 

also eot: yeah in the end I think i'm firmly in camp "something went wrong here". There shouldn't be this much confusion from so many different people, haha.

Share this post


Link to post
Share on other sites

You're making a fundamental assumption about how things should work, that solutions follow a simple constructive set of rules, and your problem is that assumption is contradicted by an example of how things actually work. Also I don't find it too strange that this is the only puzzle where this principle applies for the only possible solution, I think if you dig around you can find a few more examples of single-use mechanic puzzles in the game.

 

What's interesting would be to think about the programmed logic of how the game actually knows what solutions to a puzzle are correct since there isn't always only one pre-designed correct answer. I suspect part of the reason this puzzle exists is that the game had to support the most general possible solutions for all the puzzles since those could be valid solutions a player would try with minimal assumptions for other similar puzzles. Also the game logic just has to check if a solution is compatible with the rules of the game, not if there is some constructive way of forming that solution.

 

My guess is the actual implementation of the solution checker uses something like the "stacking" method. That is there is a 2D array of values that count the number of yellow squares occupied minus the blue squares. For each yellow shape it tries to place it in each position on the array and adds those squares, and then for each blue shape it tries every position but subtracts those squares. At the end it checks every region of the 2D array defined by the player's solution line (figuring out regions is tricky to summarize here but doable) to see if it either contains only 0s or contains only 1s and the markers for the shapes in that region. If there exists one such arrangement of the shapes then the puzzle is solved.

 

This algorithm is pure brute force (but still fast to check given the puzzle sizes in the Witness) and doesn't have to worry about how shapes have to fit together to form larger shapes without intersecting.

Share this post


Link to post
Share on other sites

There's several puzzles in the game which can be solved multiple ways, so if you want to discover how a rule works, you can set up an experiment where you solve it in a way that only works with one of the potential rules.

Share this post


Link to post
Share on other sites

I'm not making an assumption of anything. I'm stating, full up, that it is Bad Design. I don't give a shit, to be honest, if other puzzles can be solved using that mechanic. The fact that this is the only place it is REQUIRED, and it is never taught to you in a simple manner like literally every other mechanic? That is what makes it broken. There's no two ways about it, as far as I'm concerned. It's just Bad Design.

Share this post


Link to post
Share on other sites
 

I'm not making an assumption of anything. I'm stating, full up, that it is Bad Design. I don't give a shit, to be honest, if other puzzles can be solved using that mechanic. The fact that this is the only place it is REQUIRED, and it is never taught to you in a simple manner like literally every other mechanic? That is what makes it broken. There's no two ways about it, as far as I'm concerned. It's just Bad Design.

 

The way I thought about the blue squares is that they are anti-yellow squares you place, which must be countered either by placing yellow squares on top of them, or by placing them on top of yellow squares, in any _particular_ order. If you don't consider the operation ordered, then you get stacking.

The reason why I think of it in this way is because I kind of did all the tetris puzzles like that as well, one piece at a time. When thinking like that it's not stacking, even though end result is exactly the same :-) Admittedly the game doesn't force you think of those puzzles as ordered operations, so in that sense you might say it's poorly explained. But I seem to recall there being some other places, where you have to realize the rule you've learned is not the correct rule, without not exact telegraphing what you're doing wrong. For example the color/star thingies.

 

So for example the puzzle in iax's post goes in my mind like this:

 

 

The L piece:

 

001
111

 

Then the blue piece:

 

-0
0-

 

The square:

 

11
11

 

First lay the L piece on the bottom:

 

0000
0000
0010
1110

 

The add the blue anti-piece

 

0000
0-00
0000
1110

 

Finally add the square:

0000
0010
0110
1110

 

 

Share this post


Link to post
Share on other sites

I don't need to be convinced that there's another way of looking at it, haha.

 

My argument is not that the mechanic is Wrong somehow. It's that it's never taught to you in a way that every other mechanic in the game is taught to you - such that you are FORCED to learn a specific thing. There is definitely a puzzle out there in someone's head that would teach this specific thing in a simple manner, such that there would be no confusion. And the fact that there are so many people on these forums alone that interpreted the mechanic in wildly different ways, I would think, goes to show that, yeah, I'm onto something when I propose that this is not designed properly.

 

SHRUG!

Share this post


Link to post
Share on other sites

I'm mostly curious if it's intended to teach you something new. As demonstrated by some posters here, there's seemingly at least one way of thinking about these puzzles that 1) will let you complete all of them 2) will not leave you confused by the puzzle in question. There is also at least one other (simpler) way of thinking about them that will let you solve every other puzzle, but not this one. The other option is that Blow didn't intend for you to have to re-evaluate the rules when faced with this particlar puzzle.

Considering how this game guides your self teaching it would be strange to me if a puzzle teaches you something that is never used again. That is not consistent with my sense of how the game is designed. On the other hand, the second option seems like a big oversight for a game where every detail seems intentional. My gut feeling is that it is an oversight though.

If you start thinking about the science themes in the game then maybe it makes sense that there is a simple rule that holds in almost every case, but breaks down in a very particular situation. That is the case with many historically significant physics models and I wouldn't be surprised if it's something Blow would want to convey, since the figuring out of the puzzle rules is 'clearly' a metaphor for science.

I don't think it works from that perspective either though, because these puzzles don't feel anything like science. They're carefully authored, I don't get to experiment freely with the rules myself. Instead it is (for the most part) a game where you are teaching yourself in a way that someone intended. I'm not asking the question "what happens if two blocks have to overlap". Blow probably thinks that by designing that puzzle he is asking that question, but when faced with it as a player I'm in one of two modes: I'm experimenting to figure out the mechanics, or I'm applying rules that I've learned. If I think I misunderstand a rule I go back to previous puzzles to re-evaluate it, but in this case they would teach me nothing. I do that because the game sets the expectation that it will 'teach' you the mechanics in that way.

That's also one of the ways in which I think the game fails overall.

Share this post


Link to post
Share on other sites

I think calling it an "oversight" is a much more fair assessment than my much harsher "Bad Design"!! It's just frustrating, because it's one of the few issues I have with the game.

 

I agree with everything in your post, except for the very last line, haha.

 

I want to just clarify that however firm I am in my stance that this particular puzzle was a misstep, I fucking love this game. As I said, that puzzle is one of only a few minor issues. I adore it so very much. It's more than I ever thought it was going to be. I even like the philosophical ponderings, just because it gets me thinking. I don't for a second believe I understand what it all means as one cohesive theme (I'm waiting for someone much smarter than me to talk about it in order to coalesce my thoughts), but I've always been fond of vague ideas presented in a completely nonlinear fashion. I don't even necessarily agree with all the ideas I HAVE gleaned from The Witness, but nonetheless. I loved it.

 

I was not so high on Braid's narrative/story/presentation/whatever, but I wonder if I would be today. I'm a different person. Then again, a lot of people I respected/respect panned Braid's text, so maybe I'd still be ambivalent, if not outright negative.

 

I certainly wouldn't call The Witness pretentious, like I've seen people say across the internet. But, then, I think pretentious is a loaded term and is often used to dismiss a work because it's too complicated and people don't want to put in the effort. Well, whatever.

 

I'm rambling.

 

This game is fucking fantastic. I'm on board for whatever Blow makes next. I don't care what it is. I hope this game is successful!

Share this post


Link to post
Share on other sites

Yeah, I'm also excited about his (or rather this studio's!) next game, I remember some vague tweets about various game concepts he has in mind, many of them didn't sound like puzzle games at all. E.g. I remember something about an action FPS that "knows what's interesting about moving through 3D space" (paraphrasing).

Share this post


Link to post
Share on other sites

I'd certainly be interested to see him take on a different genre. He's clearly got an amazing grasp of puzzle design (at least, in my opinion!). Curious to see how well he can handle something different. O:

 

I do know he's also making his own game programming language, so I imagine it'll be using whatever that ends up being? Haha.

Share this post


Link to post
Share on other sites

He makes it sound like whatever's next, it'll be another 7 year wait or more.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now