Tuesday, February 24, 2015

To Much Predicting?


            In today's modern society, we use technology to predict things like weather and traffic. But as technology as began to advance, we have started to be able to predict just more then those two. Some hospitals are now predicting when your next trip to the ER will be, we can predict when volcanoes will erupt, and when tsunamis will occur. These are all very useful things to be able to predict, seeing as most of them are involved with a great deal of danger. With socially media, search engines, and other technologies, companies have now began to predict outcomes or scores of the World Cup, English soccer results, NFL games, but most recently the 87th Academy Awards.
            Two nights ago, Microsoft’s Big successfully predicted best picture (Birdman), best director (Alejandro Gonzalez Inarritu: Birdman), best actor (Julianne Moore: Still Alice), best actress (Eddie Redmayne: The Theory of Everything), and the results for supporting actor (J.K. Simmons: Whiplash) and actress (Patricia Arquette: Boyhood) out of the top awards for the 2015 Oscars. Basically, Microsoft successfully predicted 84% of the 24 results with only four incorrect predictions.
            Before the Oscars aired, Microsoft shared its predictions noting that they were “relatively unpredictable,” yet the21 of 24 Oscar winners say otherwise. Last year Bing predicted 19 of 24 winners in 2013.  In comparison, the odds from the Wynn casino weren’t nearly as accurate predicting only 4 of 6 correctly for best picture, best actor, best actress, best supporting actor, best supporting actress, and best director. Microsoft predicted all six accurately.
            Bing’s correct predictions made me think back a few weeks ago to a different article that I read in regards to predictions. In this case, it was a videogame that did all the predicting. Each year for over a decade now, EA Sports and the Madden franchise has simulated the Superbowl. Madden has a pretty solid prediction rate of 8-3. For the 2015 Superbowl, they updated the rosters and simulated the Seahawks/Patriots showdown on Xbox One. (Here is the link for anyone interested https://www.youtube.com/watch?v=XAE6XPdLSxw)
            After a strong lead in the first half, the Seahawks get clobbered in the third and fourth quarters as Tom Brady leads the Pats to a 29-24 victory. The game then gives Tom Brady MVP thanks to his 335 yards and 4-touchdown performance. Julian Edelman grabbed 8 catches for 106 yards and Marshawn Lynch ran 118 yards scoring two touchdowns. Seems like a pretty exciting game doesn’t it, but lets see how it stacked up to the real game.  
            Just like in the Superbowl, the game did predict the Pats making a second half comeback for the win and Brady receiving MVP. In the real game, Brady threw for 328 yards and 4 touchdowns, which is only seven yards short of the prediction with 335 yards and 4 touchdowns. The game also predicted Edelman’s winning touchdown and was close in number to other key stats.
            With all of these predictions now coming out, there has been some backlash that has followed. Many people are upset that these predictions are ruining the experience of watching the program for them. What is the point of watching something if you already know what’s going to happen, if it ruins the surprise of it? It is the same concept of you missing your favorite TV show to only have your friend tell you what happens before you get to watch it.  So my question then is,  are we starting to predict to much or is there no such thing?

1 comment:

  1. Interesting post. Personally, I don't think that predicting things like sports and entertainment awards are a big deal. I actually think it's pretty cool that we have the ability to program computers that can accurately predict seemingly random things. This is especially useful for crime prevention. If we can create software that can accurately predict things like the Oscars and the super bowl (down to nearly correct stats), it doesn't seem inconceivable that we could use software to predict future crimes or potentially even acts of terror. However, on the other hand, it shows how powerful computers can be, and poses the possibility that maybe computers are not too far off from being able to actively learn or something. All in all, predicting things is cool, maybe it helps fight crime, but the future implications of computing power are sort of frightening.

    ReplyDelete