Episode 152: Arrow Machine Gun

Air Date: November 3, 2010

The ancient Greeks designed and built a fully-automatic weapon that fired arrows.


Based on fragments of design information from historical records, Adam and Jamie began to put together a period-appropriate weapon platform, using sprockets, a chain, and a hand crank as the heart of the reloading mechanism. After some preliminary tests in the M5 workshop, they took the prototype outside for further trials. Three criteria were chosen to evaluate its performance: range of at least 200 yards (183 m), rate of fire at least 5 rounds per minute, and enough accuracy to hit a human-sized target.

In the first field tests the weapon achieved a 230 yd (210 m) range but the mechanism repeatedly broke down, requiring further shop work. Adam and Jamie added wheels for maneuverability, as well as an angle finder to adjust the trajectory, and fine-tuned the reloading mechanism. They set up 5 targets at 90 yards (82 m) and brought in a professional archer, Brady Ellison, to provide a benchmark for comparison. Brady hit all of the targets in 2 minutes, using 11 arrows. After a few attempts and more tinkering, Adam and Jamie accomplished the feat with 15 arrows in 1 minute and 50 seconds. They classified the myth as plausible since they could find no records of this weapon actually being built and used in combat.

Driving while suffering from sleep deprivation (“tired”) is more dangerous than driving under the influence of alcohol (“tipsy”).


Grant designed courses to test both stop-and-go city driving with various distractions, and highway driving with long, monotonous stretches of road. For the “tipsy” test, Kari and Tory each took enough drinks to give them a blood alcohol content just beneath 0.08%, the legal limit in California. Kari made 7 mistakes in the city test, then veered out of her lane 104 times (58 seconds altogether) during the highway test. Tory’s made 2 mistakes in the city test and crossed the lines 4 times (1 second altogether) during the highway test.

Once they had both sobered up, they had to remain awake for 30 hours in the M7 workshop and they were not allowed caffeine or other stimulants during this time. The following morning, they drove both courses again. Kari had 6 mistakes on the city course and went out of the lines 210 times (332 seconds) on the highway course. Tory had 5 mistakes on the city course and went out the lines 110 times (77 seconds) on the highway course. Based on the general worsening of their performance after going without sleep, the Build Team declared the myth confirmed.


  1. MikeK says:

    Yeah, Heron’s Ballista. I’ve made one. It’s scaled down to fire pencils but they pack a punch. I can send pictures if you want some.

  2. Whacko says:

    What struck me in the “Tipsy vs Tired” myth, is that they didn’t do a control “sober” test. I mean, Tory clearly drove much better when tipsy, while Kari was terrible.

    A sober test would prove whether that is close to their normal driving skills.

    • Steve says:

      Exactly my thought, as well. Without the “sober” control, the experiment is poorly designed.

      • No One says:

        The experiment is perfectly designed to test the specific myth. Read again, they’re comparing A vs. B. Not A vs. control

  3. mwood036 says:

    i agree with whacko they should have done a control to see there driving skills in the 1st place i was yelling at the t.v. where is your control test !

    • No One says:

      And the mythbusters would be yelling back: pay attention to the myth

  4. Jayvee says:

    yes of course “tipsy” can be more dangerous than tired because drinking alcohol can cause wild frustrations, rather than being tired because your not doing anything your just uhmm… Tired

    • Gariandos says:

      I have my doubts here, because while you’re drunk, you still have a source of self-awareness, but when you’re tired, all that awareness is close to zero.

  5. Right says:

    With the driving tipsy vs tired I think they should downgrade it to ‘plausible.’ Everyone has different tolerances for sleep and alcohol. Also as Whacko said driving skill is also different for everyone.

  6. this.self.thing says:

    A sober control in unnecessary, the test is to compare tipsy vs. tired. In other words, the relative performance of one vs the other. It doesn’t matter what their unimpaired performance would have been, as they’re only looking for a relative difference. Also, because each person is being compared to themselves, variables such as driving skill and individual tolerances, cancel out.

    • Confused.individual says:

      Except, the myth wasn’t about “do Kari and tori drive better drunk vs sober”.. So then by only comparing them to themselves is kind of pointless to even experiment.

  7. peter gallardi says:

    I was a bit put out they did not run the course while sober and not tired.To me it proved nothing.

  8. Knightwar00 says:

    I have not seen this episode becuase I am deployed. but speaking from a person who have done duty to where I have stayed up to straight 30 hours awake. 30 hours of none sleep is not equal to a .08 blood alcohol level in any shape, form, or fashion. It made be closer to 2.0 blood alcohol. Most people start to feel the effect of drowseness at about 18 hours of being awake every hour atfer that is like drink a can of beer,glass of wine, shot of liquor. What they should have done is test there driveing at every hourly intervel to learn when the body start to feel the full effects of drowsiness and determined when staying awake become more dangerous then driving drunk. This show sound like it really could have helped people see why it not good to push past the last hour of drive from one point to the next. But the overly rare case when someone is up for 30 straight hours seems a bit foolish to me. At least put it this way some student in college wakes up at 7:00 am to go to class he/she is in class till 4:00 pm spents a couple of hour completeing home work. Takes another hour to pack geting ready to leave on a 5 hour drive from college to his parents house. What is that student level of danger as apose to DUI?

  9. Jaime says:

    Anyone ever been in the Army, had to get up early in the morning to pull a detail, conduct PT (physical training) then pull a 24 hour staff duty detail without any sleep allowed, then brief your relief the next day, and THEN have to drive home afterwards? I have.

    You’re talking between 30-35 hours of being awake, then driving home. Americans, if you live near a military base, think about this before you drive. Then write to your state congressman or assemblyman. Thanks!

  10. Andrew says:

    re: control test. For every scientific test you need a control. It’s like making a line graph or a bar graph intentionally misleading. Without a control you do see that being tired leads to worse driving than being slightly drunk, but you don’t know how much being slightly drunk is worse than being tired and awake. It’s a point of reference. Without a control, the results aren’t as easily applied to what we know as viewers. Every experiment needs a control.

    • Norwegian Dude says:

      This is only partially true..
      As you wrote; we need a point of reference. That is true.
      So if we pretend they did nothing but the tipsy-driving, we wouldn’t know if it was actually worse than anything.
      However, and this is the clue many seem to have missed,we DO have a point of reference. For the tipsy-driving, the reference is the tired-driving and vice versa. We want to compare them to each other. If the test was about finding how much each of these conditions affect us, we’d need a reference, though, as the test was constructed to compare the conditions to each other, they did exactly that.

      • Tim says:

        The problem is that we have no frame of reference for the difference between the two- it may be relatively insignificant. For example, lets talk math for a moment- if I tell you that x is 9 greater than y, the relative difference between the two is huge if y is 2, but miniscule if y is 14,000.

        It is clear with the experiment that there is certainly a difference. But how big that difference truly is is difficult to establish.

  11. Alexander says:

    Just below the legal limit is only a little tipsy, but 30 hours of sleep is very very tired. Not a good comparison. That’s like comparing 3 kilos of gunpowder to 1 kilo of tnt and saying the gunpowder is stronger.
    A better test would be 2-4 hours of sleep the previous night, or 20-24 hours no sleep.

Leave a Reply