I have been trying to learn what I can learn during these last few days of the Google Idol contest, in hopes of being able to maximize Mica’s chances. The first potentially relevant fact is that the website hosting the contest is registered in Brisbane, Australia. I had often found it difficult to guess what time the server would be ticking over into the next voting day, allowing all the IP addresses that had already voted to do so again.
This round ends on June 24th, but nowhere does the website specify at what time. As such, the earliest it could possibly end (00:01 Brisbane time) would be 2:01pm Oxford time on the 23rd. The latest it could possibly end (23:59 Brisbane time) would be 1:59pm on the 24th. If someone has figured out at what time of day their server ticks over, it would be very useful information.
Why?
Because the lead has been cyclical:
Chart based on data between 22:00GMT on the 18th and 22:00GMT on the 20th.
As you can see, the distance between the number of the votes for each video rises and falls according to an orderly pattern. I would guess that with ‘Twan, Sjoerd, Manuel en Iwin’ living in Western Europe and Mica coming from the West Coast of North America, there is about an eight hour lag between time equivalencies in the areas where most of their respective voters will be living. Those of Mica’s competitors rise eight hours earlier, vote, and go to sleep eight hours earlier.
The fact that the slope of Mica’s line is more constant may be the product of how I have been cajoling people on the east coast of Canada and the United States – as well as in the UK and elsewhere – to vote for him as much as possible. Alternatively, I may have nothing to do with it and people voting for him just vote at times more distributed across the day for some other reason or collection of reasons.
As such, it would be helpful to work out what time it will be in each place when the contest ends. Ideally, we would probably want it to end around midnight Vancouver time, when it will be about 8:00am in Europe. I think that would be about 6:00pm in Brisbane.
[Update: 22 June 2006] I have created a chart that shows the amount by which Mica has been winning or losing at various times when I have checked on it.
The contest ends at 7:00am Vancouver time, 3:00pm your time.
On the 24th.
What is the time scale on that first chart? If you had more data points, you might be able to do a Fourier transform.
B,
I considered doing that, but I don’t really think it would tell us anything useful and it would be a fair bit of work.