Segments most liked by viewers: gel torso used for testing a weapon
Spike TV's The Deadliest Warrior, now in its 3rd season, is a television program that creates hypothetical battles between famed historical or current fighters (or 'warriors', as they put it) to see which one would be the 'deadliest'. In the past, this has included such awesome (and occasionally bizarre) matchups like the American Green Berets versus the Soviet Spetsnaz, the Nazi SS versus the Viet Cong, Al Capone and his cronies versus Jesse James and his gang (my wife's favorite), and a pirate versus a medieval knight. (The knight lost.) Later this season, we'll see a blood thirsty battle between Vampires and Zombies. Yes, Team Edward will go up against the shambling undead. I'm rooting for the Zombies.
This probably sounds a lot more like a drunken argument after a few too many rounds of Halo, but to give Spike credit, the winner of any give matchup is determined by what folks in my line of work call simulation analysis. What that means, basically, is that every potential feature each 'warrior' would bring to the fight is examined, then turned into numerical percentages and number-crunched via a computer program that then simulates a thousand different battles. The Deadliest Warrior is of course the side who wins most often based on the comparison of the data.
In the first and second seasons this really just amounted to a bunch of gleefully fake blood-soaked weapons testing, which was awesome. In the third and current season, however,
Now if they would just cut out the irritating trash talk between the guest combat experts, the show would be just about perfect.
As I mentioned above, the key de'nouement of the shows consists of simulating a series of battles between the two opponents (using the computer program aka a black box), and we get to see who wins via a fictional battle between the two warriors (played by actors and stunt people). I love how obvious it is that all the final battle segments are filmed in California (even though the warriors featured in the episode lived say in Asia or the Arabian Peninsula).
In order to find out Who Is Deadliest? (The catch phrase of the show and definitely said in ALL CAPS), battles are simulated 5,000 times (4,000 more times than in Seasons 1 & 2). Then, the results are tabulated and the warrior who wins the most battles is declared the ultimate victor. So far, this
George Washington vs. Napoleon Bonaparte
2,530 (50.6%) 2,470 (49.4%)
Joan of Arc vs. William the Conqueror
2,587 (51.74%) 2,413 (48.26%)
U.S. Army Rangers vs. North Korean Special Operation Force
2,504 (50.08%) 2,496 (49.92%)
Season 3 Episode 1 - George Washington trying to motivate his troups
But if you look carefully at the stats, the pairing results are very, very close. In fact, the greatest difference is about 3.5-percent. As someone who uses statistics all the time in my research, I can tell you that the difference in percentages between all but the Joan of Arc versus William the Conqueror is actually statistically insignificant. In other words, most of the time both warriors are deadliest.
The reason for this has to do with basic stats, but when I tried to explain it to my darling wife her brain exploded. If you don't know much statistics, I'm afraid you'll have to take my word for it. For the more statistically inclined, I've laid out Spike's simulation problem below:
It all comes down to an extremely important item that's missing from the sims: the uncertainty associated with the simulation output.