Ok, so I didn't know where to put this so I will just put this here. Mods please move if you think neccessary.
Here it goes:
The Wii has a world voting system on its console. (population being those who have a Wii and are actually participating in voting) The questions asked are generally about people's perceptions or prefrences for whatever the proposed question is.
Questions include:
- Which do you prefer, rain or snow?
- Which would you prefer, a backyard or a swimming pool?
- Do you order Chinese or Pizza for take out?
- Do you listen to mp3s or CDs?
The point is to try and predict what other people will pick. You vote and then it is tallied with everyone elses and in a week, you get the results back to see if you were able to succesfully predict the answer to the question. However, here is the key. The answers are all dependent upon what
everyone else votes and depends on whether they are trying to answer based upon their personal preference or actually
predict what everyone else will answer.
So, my friend has voted about 125 times and has a prediction accuracy of 70% on the Wii. I told him that I predict that after 500 times of voting he will be at least less than 70% maybe even close to a 50 % prediction accuracy which he obviously did not like, so we placed a wager for at least the two months time we will be staying in Japan. Will his prediction accuracy go above or below 70% by the time we leave. (We leave in 2 months) If statistical theory is correct, it should be reasonable to presume that there is a highly probablistic chance that I am correct, but then again their are always outliers.
He claims that there is a right answer to all questions and that, as long as he successfully makes his predictions, he will be able to keep sustain an overall prediction accuracy that is above 70% the more times he votes. I disagree.
Here is how I see it. Given that all questions are mutually exclusive and that the questions themselves have a varying degree of deviation in which some questions will be 20/70, 70/20 or dead on 50/50. Based on statistical theory, as his sample size increases (or the more times he votes) no matter what, he will come closer and closer to the predicted probability of each question. There is a 50/50 chance of getting an answer right or wrong.
However, he argues that there
is a right answer and that it is reasonable for anyone to consider that there is a obvious answer to many of the questions. But I say (and this is everyone around the world with a Wii) that you have no way of knowing what peoples predications or preferences* are going to be, so you are left to the theory of probability. The probability of you getting an answer of couse is 50/50.
So, will I win this wager or will the theory of statistics fail me?
*This contradiction alone makes it impossible to predict whether or not subjects are actually trying to predict answers or giving preference but that is a debate all together.
