Losing the forest for the trees with grid drawings

post by Adam Zerner (adamzerner) · 2020-09-24T21:13:35.180Z · LW · GW · 1 comments

Last night I learned about grid drawings and spent some time trying to work on them. In doing so, I ended up having a lightbulb moment where I really internalized what it means to lose the forest for the trees.

I'll explain what I mean. To start, let's back up and talk about what grid drawings are.

The top image is the reference image, the bottom is your version. You draw grid lines in the reference image. Then you draw grid lines on your version. Then you go square-by-square and try to copy what you see.

For example, let's say that rows are labeled A through H and columns are labeled 1 through 11. For the flower, maybe we start at square A6. We look at what A6 looks like in the reference image and try to copy it into our version of A6. Then we move on to A5 and do the same thing. Then maybe B5. So on and so forth until we have gone through every square.

I think the idea is that there's less going on inside an individual square than there is when you zoom out and look at the image as a whole, so by approaching it this way it ends up being easier for the artist.

At first this felt like it made a ton of sense. I was successfully using the technique to draw easy things. But then once I moved on to a harder image, I ran into a weird problem. I was going square-by-square, and I felt like each individual square I drew was pretty close to the one in the reference image. But then I zoomed out and looked at what I had... and it looked terrible!

Huh? How could that be? So I went through each of the squares I drew and compared them to the ones in the reference image. Maybe I screwed one of them up or something.

Nope. Each one seemed pretty good. And yet... the zoomed-out perspective still looked like garbage!

I was confused. Reductionism is a thing, right?

And then I realized: small errors in each individual square can compound and make the zoomed-out version look terrible.

And furthermore, I was paying too much attention to the individual squares and not enough attention to the zoomed-out version. To the trees instead of the forest. Let me explain what I mean by that.

Say that in cell E21 the line in the reference image starts about 25% down from the top and goes at an angle of about 45°. I was just trying to imitate that. But if my version of the adjacent square, E20, is screwed up, I should factor that in to my E21 square. E20 should "flow smoothly" into E21. But I lost sight of that. I just kinda moved from square-to-square without factoring in the adjacent squares too much.

And that's when it hit me that this is exactly what it means to lose sight of the forest for the trees.

It honestly felt like a very powerful moment, and I'd recommend that you go through the grid drawing exercise and try to reproduce that moment yourself. I've always had an intellectual understanding of "losing the forest for the trees", but something about last night really hit home and helped me internalize it. It's like the difference between reading The Lean Startup and spending two years of your life failing at a startup because you weren't agile enough. Internalizing is different from intellectual understanding.

1 comments

Comments sorted by top scores.