The Warmest 100 and the Tepid 100: How Statistical Predictions Reshaped Triple J’s Voting Landscape
In our earlier coverage of Triple J’s Hottest 100 predictions, we traced the rise of data-driven forecasting methods that briefly outwitted the station’s voting system. By 2012, Brisbane statistician Nick Drewe had demonstrated that public social media feeds could be scraped to reconstruct the countdown with startling accuracy, collecting over 35,000 ballots. Now, shifting focus to current realities, we examine not only how Triple J responded but also the broader legal and privacy implications that emerged as these prediction projects evolved into a pattern of unauthorised data collection.
Nick Drewe’s Social Media Scrape and Triple J’s 2013 Lockdown
Drewe’s original Warmest 100 exploited Triple J’s decision to allow voters to share their ballot on Facebook and Twitter. By harvesting those public posts, Drewe compiled a list that accurately placed 92 of the eventual 100 songs. “The sample of 35,081 ballots represented an extraordinary 2.7% of all votes cast,” Drewe noted at the time. Triple J immediately closed the social-sharing feature for the following year. The initial lockdown worked—Drewe declined to repeat the project. Yet Australian-born economist David Quach later encouraged Drewe to explore less obvious social media sources, prompting the 2013 Warmest 100 based on a smaller dataset of 1,779 ballots (0.11% of votes).
Ed Pitt’s 2015 Tepid 100 and the Instagram Workaround
After a one-year hiatus, Melbourne student Ed Pitt launched the Tepid 100 on Tumblr, pivoting to Instagram as his data source. By tracking the #hottest100 hashtag, Pitt collected 2,064 ballots (0.1% of votes) and produced a spreadsheet predicting the full 2015 countdown, including songs ranked 101–200. Although the accuracy dipped compared to Drewe’s efforts, the Tepid 100 demonstrated that motivated individuals could still reconstruct the list even after Triple J tightened its system.
| Project | Year | Data Source | Ballots Collected | % of Total Votes | Songs Correctly Predicted |
|---|---|---|---|---|---|
| Warmest 100 (Drewe) | 2012 | Facebook, Twitter | 35,081 | 2.7% | 92 |
| Warmest 100 (Drewe) | 2013 | Other social media | 1,779 | 0.11% | N/A |
| Tepid 100 (Pitt) | 2015 | 2,064 | 0.1% | Lower than 2012 |
“The 2012 Warmest 100 counted 35,081 ballots (2.7% of votes), while the 2013 Warmest 100 counted 1,779 ballots (or 0.11% of votes). The 2015 Tepid 100 managed to collect 2,064 ballots (0.1% of votes).”
— Source: Original Tepid 100 page (archived at archive.org).
Legal Options & MDL Status: The Unseen Fallout of Social Media Scraping
The data-harvesting methods used by Drewe and Pitt sit at the intersection of statistical curiosity and digital privacy law. Today, unauthorised scraping of user-generated content has spurred a wave of mass tort litigation. Multidistrict litigation (MDL) has been consolidated in several US federal courts for claims against social platforms that permitted third-party data extraction without informed consent. In Australia, the statute of limitations for privacy-related class action claims is generally six years, meaning individuals whose data was captured during the 2012–2015 prediction projects may still have standing to bring suit.
Each plaintiff in an MDL or class action typically seeks compensation for harm arising from the adverse event of data exposure—even if that exposure was “public” at the time. While no global settlement has been reached specifically for Triple J-related scrapes, parallel cases against Facebook and Instagram have resulted in multi-million-dollar payouts. The FDA, though primarily a drug regulator, has been cited in legal arguments as an analogy for the need to impose adverse-event reporting standards on data breaches. These developments underscore that what began as a hobbyist experiment now carries serious legal exposure for anyone replicating such methods.
- 2012–2015 Data Scrapes: Drewe and Pitt harvested votes via Facebook, Twitter, and Instagram without platform authorisation.
- Triple J’s Response: Disabled social-sharing in 2013; subsequent workarounds forced reliance on manual hashtag tracking.
- Legal Precedent: Ongoing MDLs for social media scraping highlight the risk of class action liability and statutory penalties.
- Practical Advice: Anyone considering similar prediction projects should obtain explicit API permissions or risk exposure to litigation and settlement demands.
Your Next Steps: Understanding Your Rights in the Wake of Data Exposure
If you voted in the Hottest 100 during the 2012–2015 period and shared your ballot on social media, your data may have been used without your knowledge. While Triple J itself was not the scraped party, the platforms involved now face scrutiny over their data-handling practices. To learn if you qualify for potential compensation as part of a mass tort or class action, you should:
- Review your social media activity from the relevant years.
- Check whether your posts were publicly accessible at the time.
- Consult a legal professional familiar with digital privacy litigation and statute of limitations deadlines.
- Document any evidence of unauthorised scraping or data reuse.
The 2012 Warmest 100 and the 2015 Tepid 100 remain case studies in statistical ingenuity—but they also serve as a cautionary tale about the legal boundaries of public data collection. As of 2026, the MDL landscape continues to evolve, and plaintiffs have won early settlement victories that could set precedent for future claims. The era of unchecked social media scraping is over, and accountability is now the order of the day.
Contact us today for a free case review and learn if you qualify for compensation.