The brainchild of Brisbane resident and online marketer Nick Drewe, The Warmest 100 has used statistics gleaned from votes posted to Facebook, Twitter and other social media hubs to build what Drewe thinks is an accurate prediction of the final countdown.

“I was on Facebook on Sunday night two weeks ago,” Drewe explained to TheVine. “A few of my friends had posted their votes on the Hottest 100. I was just browsing through, having a look at them, and I noticed the URL structure of each of those pages had the same pattern.”

Drewe established that each vote carried with it a unique identifier, and began peppering Twitter with a clutch of different search queries. Within an hour of logging onto Facebook he had accumulated 6000 votes and applied them to a spreadsheet. “From there, it's a pretty simple process to calculate a prediction or estimate of what the Hottest 100 is going to be,” he said.

By the end of last week, and two days before voting closed, Drewe had collected 20,000 more submissions, which when compared to the 1.3 million votes tallied in last year's poll gave the Warmest 100 an impressive sample size of 2%. Joining forces with ex-We Are Hunted staffer Tom Knox, together the two called in a bunch of favours to create Warmest100.com.au.

“I spoke to Tom, and he suggested I put up a microsite where we could make all the songs playable,” Drewe told TheVine. “Just something that people could come to and see a prediction of the Hottest 100 and learn a bit about how we pulled it together. Make it a fun project.”

Drewe was careful to describe The Warmest 100 as a prediction, rather than a leak -- he says it's not his intention to ruin Triple J listeners' Australia Day Celebrations. The site is carefully laid out with a number of safeguards protecting those who might be curious about the team's methods, but want to remain in the dark about the final countdown.

“If they choose to they can scroll down and start seeing the list from 100. They can listen to a few songs if they want and keep scrolling down towards number one. And as they keep getting to the top of the list we'll continue with the warnings that they might ruin their Australia Day,” Drewe said. “We think it'll be quite indicative of the final list.”

To investigate Drewe's methods, TheVine contacted University of Queensland Chair of Applied Statistics, Professor You-Gan Wang. “Assuming everything's independent of each other,” Professor Wang said, “and if the 26,000 is a random sample among the total potential of 1.3 million … I think with the top 100, you might get 90 right. You might get 92 or you might get 95. It's unlikely you'll get all 100 correct.”

But Professor Wang also explained that the closer the sample list gets to number one, where a greater percentage of overall votes resides, the more accurate it becomes. It's an opinion that matches Drewe's own observations: as the sample size has grown, the top songs tend to lock into place or move just one position up or down the list.

“Towards the bottom of the list, there might be only a few votes difference between each song, so they're a little more likely to change positions,” he told TheVine. “Whereas the difference between numbers two and three on the list is 52 votes. If we've got 2% of the vote, translating that into the overall number of votes that will be submitted would equate to 2600 votes between numbers two and three. The difference between the number of votes in the top 10 is a lot more significant than the difference between positions 80 and 90, for example.”

Even if the predictions are off the mark—something neither Drewe nor Professor Wang anticipate—The Warmest 100 exposes the pitfalls of using social networking to virally market the poll, particularly when, at the time of writing, betting sites such as Sportsbet.com.au, Luxbet and Sportingbet are still offering odds on the countdown.

But gleaning votes from Facebook, Twitter and the like casts a final query over Drewe's sample. “Are these people, who are posting their votes to Twitter and Facebook, representative of the average Triple J listener?” he wonders. “I think the demographics line up pretty well – it is a youth radio station and they would be attracting a fairly tech savvy and socially engaged audience.”

Professor Wang agrees: “Maybe certain types of people use Facebook, and what is the nature of that relationship to music? I don't know – unless you have historic data on whether Facebook's accurate and test if it's actually a good source of data. But common sense tells me that it's probably valid. I don't see why it wouldn't be valid. I think it's a good way of collecting samples … It's a good way to forecast or predict the top 100, especially. I think they'll probably get most of them right.”

Late on Sunday night, just two hours before the close of voting, Drewe conducted his last collection of data, pushing his sample up to a massive 35,081 votes – or roughly 2.7% of the anticipated final vote – further refining his countdown's final tally.

“It's going to be an exciting thing for us, sitting there with our own list, ticking it off as each song comes on. We'll only have one shot at this – Triple J won't do the same thing twice, letting people freely publish their votes online in such a public manner.”

Voting for the Triple J Hottest 100 is now closed.

The Vine