Dominoes - HARDCORE Mode - Smarter Every Day 182

Okay, let's just get this out there right now. I know this is weird. You probably watch this channel because you want to see slow motion phenomenon of like bullets hitting stuff, and fracture mechanics, and water drops bouncing, and animals squirting things, whatever you're into. But this is crazy. Most people think, "Oh look, the cute little nerds set up the dominoes, and they knock them down, and everyone claps and it's great." No. This is different than that.

I was naive enough to think that I was gonna set up a high-speed camera, point 'em at dominoes, I was gonna set 'em up, knock 'em down, vary the spacing and the type of floor it was on, and I was going to transcend the knowledge of dominoes instantly. You were gonna get smarter every day, and we were gonna go about our merry way. But no. Look me in the eyes. This has broken me. I do not understand dominoes.

The reason this's broken me is because I've been sitting here doing all this math to determine the gap necessary in order for this red domino's weight to always knock over the blue one. But when I set up with my calipers, it doesn't work. And if you think about it, the same people that taught me how to do this math also told me if I needed to, I could assume a cow is a sphere in a vacuum. Paper physics can only get you so far.

If there's anything I've learned from my job as an engineer, it's that there's no replacement for actual real-world testing. If we want to make observations of the exact times these dominoes hit down to the millisecond, it's really hard to do it from a camera like this. The reason is, look over here. You can see the inside of this domino, and then the inside of this one, and somewhere in the middle it changes. So if we want to get a shot where we get all this data, we have to have a moving camera, not only that, but a moving camera that is perfectly timed right here in the middle.

So we can see when this edge hits this edge, which means we're gonna have to time it perfectly with the wave of dominoes as it runs along the track. Do I want to keep making videos on smarter every day about dominoes? Yeah, if that's what it takes to find the answer. Yes, that's what I want to do. This is about not quitting. So now we're gonna give dominoes the respect they deserve, and we're going hardcore mode.

We're at fifteen millimeters. The second one over here. Welcome to hardcore mode. It's the same as before, only I try harder. So here we go, we're gonna build stuff and we're gonna do experiments, and we're not gonna quit, and there's a twist at the end of this video. Let's do it.

Okay, this is Trent. He is the contraption fabricator that I talk about on the Patreon page, and this is my cousin Hayden. Hey, Hayden. Okay, so this is a high-speed camera on a skid, and Trent made this. Good job, Trent! And we are about to try to pace these dominoes so that we can tell the difference between a hardwood floor and, you know, a surface that's grippy.

Okay ready? That looked awesome! Oh, it outran us! It moves a lot faster than the wood floor, and I'm so used to the wood floor. He's frustrated. I know well enough to know when he's frustrated. He's frustrated. You missed that one. Are you frustrated, Trent? No? All right, we're going to figure out exactly how to nail this.

We're gonna get all the dominoes in frame as they fall so we can gather the timing data and figure out what steady state velocity is. It took some practice because sometimes the dominoes would outrun us, and sometimes we outran them. So I keep messing up the high-speed, which means Hayden has to keep setting up dominoes.

Which means, up! If we're gonna... my half-court shot! We got it! It'll work, it'll work this data. This is the first good run we got on felt. Watch closely and you can see that the bottoms are trying to slide out from under the domino, but it digs into the felt and stops.

I think we got it! Do that one more time. I feel really good about that. Oh, we did pretty good! Hayden? Hayden! We got it! That's the one, right? This is the best run we got on hardwood. If you look at the bottom, you can see that it's slipping out from under the domino, making it rotate around its center of gravity. And also, if you look at the top, you can see that they're not hitting square on.

So which one is faster? We've got hardwood on the bottom that slips and we've got felt up top that doesn't slip. Trent and I both felt felt was faster, so how do we know? We have to quantify this. The answer is brute force science to figure this out. We did a ton of runs, mostly filmed at 5,000 frames per second.

The goal is to figure out exactly what millisecond each individual impact occurs, which made for over a hundred and twenty thousand frames of data to reduce. Your boy here has too many kids and too many jobs to reduce that much data, so I blasted out a spreadsheet template to everywhere that smarter every day touches the internet.

I compared you guys against each other—Twitter got it done fastest, people followed me on Instagram and Snapchat, got it done quickly. Facebook helped surprisingly. The physics subreddit wasn't interested in doing physics that day. The smarter every day subreddit was great and accurate, but Patreon was by far and away the most accurate source, which makes sense because these are the best people in the world.

There's a reason I gave each social media outlet its own spreadsheet: it's called the wisdom of the crowd. Here you can see exactly what frame each individual chose at the impact point and the average of those selections. These averages were then graphed for each individual video file, and there you have it. If you average all the runs together, it appears that the non-slipping felt is faster than the slipping hardwood.

Before I go into why I think it's faster, I'm gonna let Grant from the YouTube channel Three Blue One Brown take a crack at it because this dude is fantastic at reducing super complex mathematical systems down to quick little YouTube bytes. It's amazing! You should check his channel out. Anyway, go for it, Grant.

Hey, Destin! It's Grant. So I wouldn't... I haven't looked at some of the data that you sent, and here I'm just gonna pull up two examples: one of them with dominoes on felt and the other on hardwood. These are just the plots of the average velocity of each individual domino as they fall, and obviously, the first thing that stands out is just how chaotic this is, right? They're super jaggedy.

And I think some of that might just be that the distances between each domino were probably a little bit variable, right? That's probably not perfect. But the way that we were computing those velocities kind of assumed that that distance was a constant. So one pretty simple thing I did to just kind of smooth out this data is rather than graphing each individual velocity point, take a look at the previous 20 different velocities and then average all of those.

That way, if we look at the moving average, it should hopefully kind of smooth out any of the small variations and things like distances between dominoes. So when you do that, it does look like that in general, felt is a little bit faster than hardwood, right? But that's not always consistently true.

Like here, if we look at the last few falls, it seems like the hardwood sort of catches up with it. And this is just two examples, but I saw this in a lot of the other ones, so if you just look at the overall average velocity, felt definitely does look faster, but I wouldn't be comfortable saying that it's consistently faster. I think there's probably a lot of other variables at play here that sometimes might pull hardwood ahead.

Grant's channel is amazing! You should really go check it out, but you gotta skim over the chaotic part of the graph, and for me, that's the most fun part. When you track a rocket with a radar, you get a general idea of the acceleration profile by looking at the distance versus time plot. But the magic is over here in the noise.

When a rocket makes a correction, usually that shows up as a perturbation in the raw data. And I'm seeing perturbations in this data, and there seems to be an actual pattern here: a natural undulation of the curve. Initially, I didn't have enough dominoes set up to determine a steady state velocity, but now I'm wondering if there even is a steady state velocity because if you look at the data, it goes faster and slower over and over again.

I think this might be happening for two reasons. Number one, you can see little twists. You can see that the dominoes aren't hitting each other square, and in fact, the twist gets worse and worse and slower and slower until it gets so slow that the pushers behind it fall down and straighten them back out. This slam down is that velocity spike that we're seeing.

Watch this twist. Things are stacking up and slam! See that spike? It's a natural cycle in the graph. There's a twist, a twist, twist, twist! They're starting to stack up and slam. The twist is energy lost in the system, and the slam is the positive correction.

The second thing I'm seeing becomes more clear when we spread out the dominoes. Where each domino hits the next one seems to determine how much it rotates versus how much it's pushed. Watch this! When a domino is hit down close to the center of gravity, the impact will tend to translate the domino forward instead of rotating it because this translation calls it to still be upright.

When it hits the next one, it hits it up high, which makes the next one rotate faster. The next one gets hit down low, and you see that a back-and-forth pattern emerges. I'm not really sure how this plays in, but it's clear to me that where a domino's hit on its back determines how it falls forward.

When you spread the dominoes out farther, the effects of these phenomena are amplified even more. For example, watch this case: one domino in the entire chain doesn't fall. All right, so this is curious! So we see how that happens. Check this case out where the chain stops altogether due to the twist error building up past the point of recovery.

Okay, here's the twist to all this: this video is not about dominoes. This video is about creating an experiment to understand what seems like a simple system at a very, very complex level. Now that you know what to look for, you're able to see very intricate interactions in this domino chain, right now, right in front of you.

There's no real direct application, so why do we do this? This is what we call basic research, and it's very, very important. It's not like applied science where you're inventing things using some technology. It's about being able to understand and predict natural phenomena. Humans weren't able to go to the moon because they suddenly felt like it. It started by wanting to understand how birds fly. Someone had to first answer that question before we could ever start on our way.

Cancer won't be cured because we walk into a lab with a lot of money and want it to go away. It's gonna involve basic research like how cells work. These complex building blocks to understand the world are right there in front of us all the time, but we have to make that first little push into the unknown to start the chain reaction.

So here's the point: ask the simple questions! Keep an eye out for those little things that end up being building blocks that lead to the bigger stuff. At work, my favorite people are the ones that ask questions in meetings that might sound silly, but they're not because for every ten they ask, there's one that they'll ask that will start this chain of discovery and pull us all towards the profound.

So that's the takeaway: ask the simple questions and dare to try to answer them. I'm Destin, you're getting smarter every day. Have a good one!

All right, I'm gonna travel for my day job, so sorry for the weird transition here. This is the hotel room that I'm staying in. I want to say thank you to everyone who worked on the data. That was a big deal. It would've taken dozens of hours.

So thank you, everyone that did that, especially the patrons who helped me fund things like Trent making the little high-speed camera cart. Thank you to everyone who follows me on all the social media: Reddit, Twitter, Instagram. Obviously, I'm not like an Instagram model, but I try to do intelligent stuff or at least things I'm thinking about, which is normally kind of abnormal stuff. So thank you for following me on that and also thank you for answering the call when I have that data problem.

If you're new here and you like the idea of crowd science, feel free to consider subscribing to Smarter Every Day to help me do stuff like this when I blast that out on social media or whatever. I really appreciate the consideration for subscriptions and follows on all that stuff. I'm just a dad that works for a living, so whatever! Thank you anyway. I'm Destin. Thank you again. Have a good one!

Okay, you can do it! Trigger!