Who is going to win triple j’s Hottest 100 of 2019?

With voting closing tomorrow (20 January 2020) at 3pm AEDT, we are starting to get to the pointy end of our countdown prediction.

At the time of writing, we have collected 36,643 votes counted across 4,009 entries. Over the last few weeks we have seen the top 10 and top 20 jump around a fair bit, however, the top 3 have stayed consistent throughout.

This poses the question;

Who will win this year's countdown?

Even with the amount of data we have collected this year across many social networks it's nearly impossible to call a winner. Looking at our data, the top ranked songs are too close to pick.

This isn't the first time this has occurred in our prediction. This happened in 2016, when we predicted Amy Shark to take out first place, with Flume to take out second place. On the day of the countdown, we found that the actual result was the reverse.

In 2016 we were working with a sample of about 2.9%. It allowed us to get a general indication of the top 10 but was not enough data to predict the positions with absolute certainty. The following table shows this:

Song ABC Rank Our Prediction Error
Never Be Like You {Ft. Kai} - Flume 1 2 1
Adore - Amy Shark 2 1 1
Jungle - Tash Sultana 3 3 0
1955 {Ft Montaigne/Tom Thum} - Hilltop Hoods 4 6 2
Redbone - Childish Gambino 5 4 1
Believe {triple j List A Version 2016} - DMA's 6 5 1
Papercuts {Ft. Vera Blue} - Illy 7 7 0
Say It {Ft. Tove Lo} - Flume 8 12 4
Stranger {Ft. Elliphant} - Peking Duk 9 10 1
Starboy {Ft. Daft Punk} - The Weeknd 10 18 8

There was not enough data for us to predict the top two in correct order, but taking a more macro view, we successfully predicted 8/10 songs placed in the top 10, and 3/3 songs placed in the top 3.

Making exact predictions is difficult, especially when you only have a small fraction of the data. The data will generally give you a good indication, but will never allow you to predict, with certainty, that a particular outcome will occur.

So, who will win triple j’s hottest 100 of 2019? It could be Denzel Curry, it could be Billie Eilish, or perhaps even Tones and I. With the current sample of data and the fact there are only a few hundred votes separating the top two, it's impossible to say with absolute certainty.

What can be stated with much higher certainty is that 100 Warm Tunas should successfully predict a good portion of songs which will reach the top 10 and top 20.

See Also: