clock menu more-arrow no yes mobile

Filed under:

Sharks Gameday: Rise of the Machines

New, 184 comments
The human concept of "high-fiving" confuses Marleau-bot.
The human concept of "high-fiving" confuses Marleau-bot.

Of course you're all aware of the computer that is humiliating our greatest heroes on the popular television program Jeopardy, right?

The Jeopardy computer, Watson, is an amazing feat of technology and also a big jerk. Here's an example: at the end of the second episode on which it appeared, the Final Jeopardy category was "U.S. Cities" and Watson answered "TORONTO????" Everyone gasped because the machine had gotten the question wrong after absolutely destroying the human players all day. Then when Watson's wager was revealed everyone laughed: it had only wagered a few hundred dollars, still leaving it far far ahead of its fleshy competition. Alex Trebek called Watson a "little sneak," but my theory is that the mechanized demon actually did the whole thing on purpose, simply for the thrill of getting our human hopes up and then smashing them under its metallic foot*.


The whole Watson experiment has gotten me thinking: how long until the machines can beat us at hockey? They're already better than us at Jeopardy, chess, dancing, and playing the violin - it strikes me as foolish to assume that robots won't eventually be able to surpass humans at hockey as well.

Programing machines to play hockey will certainly be more difficult than designing computers that can answer trivia questions, but it's a reality that definitely seems to be within our grasp. The hockey intelligence required to play the game might actually be the easiest part - the kind of strategic thinking that goes on for the computer at the highest difficulty settings of NHL video games is already an impressive feat. Getting the robot to synchronize control of its various body parts along with visual feedback and split-second thought processing will be the real challenge.

At first thought, balancing on ice skates might seem to be a huge obstacle for a hockey-playing robot - but this is a bridge that can be crossed. After all, we already have military robots that can do this** so in another twenty years a robot that can ice skate is certainly possible.

The world of soccer is already ahead of the game in the area of mechanized-players, with the international organization known as the "RoboCup" having set a lofty goal for themselves: by the year 2050 they want to have a robotic team that can beat a human team at soccer. Why is soccer taking the lead in the robotic athlete arena? Probably because literally anything is more exciting than watching humans play soccer.

Boxing, too, is exploring the world of robot competitors, as detailed in the upcoming documentary film "Real Steel," starring Wolverine.


Ouch. It would have been a double-minor for high-sticking on the opponent, but the Clowe 5000 doesn't bleed.

If we are going to have hockey-playing robots, then we'd better hope that the designers follow Asimov's Three Laws of Robotics, or we could all be in a lot of trouble. But how would these three laws affect hockey as we know it?

Law #1 - A robot may not injure a human being, or, through inaction, allow a human being to come to harm.

Okay, so that's a positive. The robots are already better than Matt Cooke.

Law #2 - A robot must obey the orders given it by human beings except where such orders would conflict with the First Law.

So the robots will obey their coach absolutely, and will never argue against a penalty. That's kind of boring.

Law #3 - A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

If robots are programmed to protect themselves from harm then we can expect all of them to refuse trades to the New York Islanders.

Assuming hockey-playing robots do NOT enslave humanity, I'm pretty excited about it. I don't think I would be interested in an entire league of robot teams, but I would definitely watch a few matches a year between robot squads, possibly affiliated with different universities or corporations. I'd even lobby for an exhibition match between humans and robots - maybe replace the All Star game with a game featuring the best players in the NHL battling it out with the best robot team? This set-up would only work for a short time, though; at first the robots will barely be able to play and it will be sort of a joke, but within a few years they will be beating the humans so badly it will be too embarrassing to continue***.

Robot players will never be able to replace humans, of course. Just like very few people are interested in watching two computers play chess or Jeopardy by themselves, a robotic NHL would only be interesting from a technological perspective, without any of the human drama and intensity that is so important in the world of sports. Sorry, robot hockey team owners - without human players, your sport just won't capture our weak human hearts the way real hockey can.

Speaking of real hockey...

Next Game

Washington Capitals
@ San Jose Sharks

Thursday, Feb 17, 2011, 7:30 PM PST
HP Pavilion

Complete Coverage >

The Sharks play the Capitals again tonight. San Jose beat them a week ago, with Niemi posting a shutout.

The lack of scoring from the Sharks, especially from Joe Thornton and Dany Heatley, is a major concern. In the eight games the team has played so far in February, Heatley has four points and Thornton has three. For the two most highly-paid players on the team, the players that are supposed to be the standouts that make San Jose an offensive threat, this is abysmal. If Antti Niemi hadn't started playing out of his mind a couple of weeks back, the Sharks would most likely have been left in the dust by the other teams in the West.

Since opening the month scoring five goals against the Coyotes and four against the Ducks, San Jose has had a rough time offensively. They've scored two goals or less in five of their last six games (and only three in the other game during that stretch). The team is still generating a lot of shots, but, as has been the case all season long, the number of shots doesn't seem to matter. The team's winning percentage when they out-shoot the opponent is exactly 50%. They're first in the league in shots per game, but SIXTEENTH in goals scored per game (after finishing last season fourth in this category). This suggests that the shots the Sharks are taking are not quality shots, and that the current offensive strategy being pursued by the team isn't working. Todd McLellan has always been an offense-first coach, so if he can't find a way to get his talented players going soon I really don't know why the team would want to keep him around in the future.

Prediction: Human players play a good human game, and the Sharks win 3-1. Highly-paid humans Joe Thornton, Dany Heatley, and Patrick Marleau all score - or at least have the human decency to feel bad if they don't.

* = this is just a metaphor...for now.

** = just imagine the BigDog robot in ice skates - how adorable/horrifying would that be?

*** = major Battlestar Galactica spoiler alert: did it bother anyone else that Anders, who had been an awesome Pyramid player prior to the attack by the Cylons, turned out to be a Cylon, meaning that he had an unfair advantage throughout his athletic career? I know it was basically the apocalypse and no one cared, but think about all the matches that fact invalidates! The Caprica Buccaneers had a freaking robot on their team the whole time! If I was a member of the Sagittaron Archers I would have been pissed.