This xkcd strip needs an extra panel

Via PZ’s blog Pharyngula I saw this xkcd strip.

I am a physicist and a physics teacher and so of course am well aware of this popular lecture demonstration where a bowling ball or some other heavy object is hung from the ceiling of a lecture hall. A person then stands some distance away from the lowest point at which the ball is resting and brings the ball up to their nose with the rope kept taut. The ball is then released and it swings away from the person and returns, just like a pendulum. The point is whether the person will flinch when they see the ball come back towards their face. The law of conservation of energy predicts the ball will not rise higher than the initial release point and so will never hit the face, and the panel looks at how different scientists might respond to being in that situation.

I disagree with the engineer’s reasoning. I cannot think of any way that you can accidentally incorrectly hang the ball so it rises higher at the end of the cycle than at the beginning. To achieve that would require the deliberate insertion of some device at the suspension point to impart some energy to the system, like the way that we do with a pendulum held in one’s hand, where a slight wiggling motion by the hand can increase the amplitude of the oscillation.

So we would be safe on that score. As a physicist, you might think that I would be confident and not flinch. Actually, I never do this experiment and it is not because I do not trust the law of conservation of energy or worry about incorrect hanging. I think that flinching, as the biologist says, is a perfectly valid reaction and says nothing about one’s faith in the law. Why I do not do it is because the person might subtly and involuntarily move their head forward after releasing the ball and so get hit. I would not risk myself or my students to such a possibility.

But the physicist’s argument about trusting the law of conservation of energy is also worth considering. What does it say to ‘trust’ a law of physics? Does it mean that we are 100% confident it will not fail? The history of science has shown us that all laws have to be considered fallible. Philosophers of science have concluded that we cannot be absolutely sure of any law and that however many times a law has been sustained, it is no guarantee that it will not fail the next time. This is known as the problem of induction.

But we have no choice to trust the laws of science all the time because our entire lives and societies are based on them. When I fly on a plane, I am putting my trust that the laws of mechanics and electrodynamics and others that enable it to fly will not fail, even though I am not 100% sure of that. So why am I willing to get on a plane, risking my life, while not putting my face in front of a pendulum, even though in both cases I am placing my trust in laws of science that are not guaranteed to be true? It is because of a simple cost-benefit analysis. The benefits of plane flying more than compensate for the very slight chance of the laws failing. There is really no benefit in risking getting hit in the face by a bowling ball other than as a demonstration of my ability to control my flinching reflex. It is mostly an act of bravado.

So xkcd should put in an extra panel (or replace the engineer one) with a panel with a philosopher of science expressing the reason why it is not a good idea.

As you might have guessed, this discussion of why we can and need to trust the scientific consensus even if we are not certain that it is true is the heart of my book The Great Paradox of Science: Why its conclusions can be relied upon even though they cannot be proven. You should read it if you have not done so already!


  1. brucegee1962 says

    I cannot think of any way that you can accidentally incorrectly hang the ball so it rises higher at the end of the cycle than at the beginning.

    It couldn’t rise higher, true, but if the experimenter tied the rope with a granny knot, the knot could come untied a few inches from your face and the ball would at least give your chin a good smacking.

  2. Jean says

    In case of failure, a bowling ball to the shin would also hurt and would have a lot more velocity than it would if it hits your chin,

  3. lochaber says

    It could also be poorly attached at the top, and bring something down on the participant’s head.

    (like if the rope were tied around a ‘decorative’ beam, or those metal slats to hold up dropped ceiling panels)

  4. says

    I suspect there will always be people who bias the experiment by their own actions, i.e. pushing and not releasing, thereby adding energy and causing it to hit the person.

    Released by a device, I’d trust it. Human hands, forget it.

  5. mikey says

    If it was hung by looping a rope over a horizontal frame member, such as a joist, oriented in the direction of the pendulum’s swing, it could slip in the direction of the ball’s travel.

  6. naturalcynic says

    Why should the prof do the demonstration? That’s what grad assistants are for. After all Penn never puts himself in peril, but Teller does.

  7. says

    Speaking as an engineer, yes, there are lots of ways this could be hung incorrectly, as others have already noted. One possibility is that whatever it is hung from might not be strong enough and breaks/bends, pushing the pivot point closer to the person. Imagine the rope is tied to part of angled beam which breaks free at the bottom and then bends from the added stress at the top. This places the ball forward and down, meaning it may well smack a certain sensitive part of the anatomy.

    Practically speaking, this not an exercise I would do in a lab because all it takes is the slightest inadvertent push upon release, and the ball will smack the participant on the way back. I am less concerned about the participant moving their face toward the ball accidentally.

    BTW, this experiment was done on the old 1990s TV series “Beakman’s World”. Even as an adult I loved the show. It was sort of like a cross between Bill Nye and PeeWee’s Playhouse with a little Devo thrown in (the theme having been written by Mark Mothersbaugh).

  8. davex says

    Maybe not higher, but swinging further would be possible if a mounting bracket loosened and began to pivot, adding length to the top of the pendulum.

    I just tried this setup in a doorway, with my finger as the block:

    x <-solid mount
    |x <- if this block falls out at vertical, the pendulum swings back past release
    ……* <- bowling ball

  9. xohjoh2n says

    That above points about failure modes leading to the ball hitting you *somewhere* are true and what you appear to have missed, however assume that the fixture is secure and well made and none of that happens…

    You then say one could “inadvertently” move yourself info the path of the ball… however in that case you’re only going to be a small distance within the path -- maybe a centimeter or an inch or so. By the time the ball gets there it will have bled off almost all of its kinetic energy, so when it does contact you, it’s just going to gently puch you back a bit, not actually injure you. So I wouldn’t think thats a problem worth seriously worrying about.

    (The rope breaking before the peak and the ball falling on and breaking your foot is going to be much worse.)

  10. John Morales says

    xohjoh2n, so there’s an opportunity the Jackass team missed.


    So, apparently, v = sqrt [2gL(1−cosθ)] (where g is gravity, L is the length, θ is the angle from the horizontal).
    So, now the question becomes: how fast must a bowling ball be going to hurt it it hits your nose? Then you can determine just how much you can lean in, depending on the mass of the bowling ball and the length of the pendulum.

    I bet it wouldn’t be much.

    (It’s pretty massive, no? (the bowling ball, not the nose!))

  11. Daniel Holland says

    I am not a physicist or an engineer, but what would happen if the rope was even a tiny bit elastic and able to elongate and therefore store/ release energy? Wouldn’t that complicate the whole conservation of energy argument? I really wouldn’t want to be part of that experiment, at least not with my face in the way.

  12. brucegee1962 says


    I just tried this setup in a doorway, with my finger as the block

    Between them, Monroe and Singham send a generation of engineers to the hospital with what comes to be known as the “bowling ball challenge” to create a self-face-smacking setup.

  13. kenbakermn says

    There’s a video on YouTube you can find pretty easily showing a professor doing this demonstration except that he asks a student to be the subject. The student releases the ball but then leans forward a good 8 or 10 cm and of course gets bonked pretty hard in the face. If I were doing this experiment the one element I would trust the least would be my ability to stand still without some fixed reference for my head position.

  14. Rob Grigjanis says

    John @11: Where did you get that formula? If θ is the angel from the horizontal, the square of the speed is

    v² = 2gL(sinθ − sinθ₀)

    where θ₀ is the angle at which the ball is released (with initial speed 0).

    You’ll more commonly see the angle from the vertical, in which case replace sin with cos.

  15. Evil Dave says

    One other concern would be that the distribution of mass in a bowling ball’s core may be asymmetric. While the center of mass would not pass its starting point, if not properly mounted, a portion of its cover could end up further out.

  16. Rob Grigjanis says

    John @16: It doesn’t work if release is from the horizontal, and the angle is from the horizontal. It just looks like it might.

    In that case, sinθ (the correct dependence) does agree with (1 − cosθ) at the angles 0 and π/2. It just doesn’t agree between those angles.

  17. John Morales says

    I believe you, Rob. Been >4 decades since I studied trig.

    Point being, the velocity is a sinusoidal relationship and it’s only at the very end that the velocity reduces sharply, and that getting bonked with a bowling ball ain’t gonna be fun even if it’s only moving slowly. So not a lot of margin.

  18. Douglas Jones says

    I agree with the engineer. The error in attaching the string to the ceiling that could hurt is one where you make a mediocre attachment, then take the free end of the string and (as insurance against the mediocre attachment coming loose) make a second attachment that is, say, a foot closer to the point of bowling-ball release. On release, the ball swings back almost to its original level far from the releaser, then the mediocre attachment comes loose, the ball drops about a foot, and then begins its return swing. You, the releaser, won’t get hit in the nose, but you’ll get a solid thump in the chest as the ball tries to go about 2 feet through you.

    I got my BS in physics 48 years ago, but Newtonian mechanics still works. I do note that the risk is eliminated if the string is long enough that the earth’s rotation makes the bowling ball precess off to your side.

Leave a Reply

Your email address will not be published. Required fields are marked *