Imagine you have two 55 gallon drums, both of them closed. One has more water than the other, but you don't know which one. All you know is that they both have at least 20 gallons in them.. The only tool you have is small pump built into each drum. The pump can output one gallon every five minutes. You run the pumps for an hour.
Question: Which barrel held more water?
Answer: We don't know. Pumping out water for an hour doesn't tell us anything. After an hour, they both will pump out about 12 gallons. Maybe a hair more, maybe a hair less. Pretty close probably, but not exact. We can expect that one pump will have pumped out slightly more than the other, but this has nothing do do with the amount of water in the barrel. It has more to do with the design of the pump. So measuring the output for one hour doesn't tell us anything meaningful about the amount that could have been pumped out.
Lets flash back to the 2003 Idol finals. 24 million votes were cast. 12 million for Clay, 12 million for Ruben. Ruben ended up winning by about a hundred thousand votes. Statistically a dead heat. But there were at least 200 million voting attempts. Verizon reported a spike of over 116 million calls for that day. SBC reported an increase of over 115 million more calls than normal.
It it possible that 150 million of those votes would have been for Clay and "only" 50 million were for Ruben. Or maybe the other way around. We'll never know. All that we do know is that the phone lines were only able to handle 12 million calls each. So that's what they got. One had slightly more than the other, but that had nothing to do with "America's Choice." It was just a random <1% fluctuation in the answering system.
So lets assume that Fox quadruples their calling capacity. Each line is now able to take in 50 million calls. We still would't know who "should have won" unless there were less than 50 million possible votes for either singer.
Is there a solution?
You would have to make sure that the number of potential votes is lower than your vote counting capacity. This means either massively increasing the size of the phone pools, or lowering the number of voters.
My Solution: Limit the number of voters.
Let Coke offer American Idol Voting tokens in packages of Coke products. Each token would allow for one internet vote for one contestant. They could be on in the bottle lids or in the 12 packs, just like the current giveaways. You would get a big enough sample that you could have an accurate cross-section of the country. And Coke would make a fortune!
Posted by Aeroth (Member # 6269) on :
That would only piss off the American Idol viewers that buy Pepsi.
Remember, American Idol is not a talent search to find the greatest undiscovered singer in America. It is a popularity contest made to draw people toward the FOX Network...
...which just so happened to be a success.
Posted by Armoth (Member # 4752) on :
What some ppl love about the show is that its just so EASY to vote. It never really bothered me that Ruben won over Clay, even though I was a die hard Clay fan. So what? Clay is the most successful out of all the contestants on the show. He got just as much air time as ruben, and gets even more air time off AI. My problem is Latoya and Jennifer. I dont want Jasmine,John, or George singing on stage while Latoya and Jennifer are sitting at home. MY solution would be for the following year, to vote people OFF. Vote for the contestant you liked the least, and this way, they kick em off. The bottom 3 will be a TRUE bottom 3, and the loser will be a TRUE loser. No sympathy votes. Its the simple process of elimination.
Posted by pwiscombe (Member # 181) on :
But even with a "negative" vote you would still run into the same problem of maxing out the phone lines. It might work for the first couple of rounds, but once you get down to a final four or five, the lines get maxed and it comes down to a random fluctuations in the phone lines.