David Scott retires

12 June 2017

Associate Professor David Scott
Associate Professor David Scott

Associate Professor David Scott is retiring at the end of the semester – but don’t think he’ll be putting his feet up. Oh no. Associate Professor Scott and his wife Bronwyn Holcombe are keen orienteers, and an orienteering trip to Portugal is in the diary. He has plans to do some voluntary work. And he’ll still be doing consulting, as well as tweaking the R packages he either wrote or supports and posting his statistics-driven rugby union and rugby league predictions on the department’s popular blog, Stats Chat.

Associate Professor Scott’s abiding memories of the department will be “the people I have worked with and the students I have supervised – supervising student projects has been one of the most enjoyable parts of my work,” he says. “The department has been a really enjoyable environment. I have said this often, but the excellence of the undergraduate teaching here and the resulting numbers of students taking those courses has been fundamental to the success of the department.”

Melbourne-born Associate Professor Scott got his BA and PhD at the Australian National University. He started his university teaching career at La Trobe University, also in Australia, in 1972, later becoming head of department there. He has also worked at the University of Sheffield in the UK, Bond University in Australia and Colorado State University in the US. He came to New Zealand in mid-1995.

Associate Professor Scott, a past President of the New Zealand Statistical Association, has a particular interest in heavy tailed distributions, which are typically used to model risk and survival, so they have applications in insurance, finance, engineering and health. One of his favourite consulting projects was a risk analysis of flights into Queenstown airport for an airline, and another involved using spectral analysis to compare different makes of continuous positive airway pressure (CPAP) devices used to assist those with breathing difficulties.

Outside the department, Associate Professor Scott is well known for his Stats Chat posts predicting the outcomes of games in Super Rugby, NRL, the ITM Cup, and South Africa’s Currie Cup – and he’s right most of the time. “I give precise statistical weightings to various factors such as a team’s long-term performance, the margin between teams, whether a team is at home or away, and tweak team rankings week by week,” he explains. “And then I can indicate the likelihood of a team winning and the size of the margin.”

Associate Professor Scott started doing the predictions as an exercise to show people the relevance of statistics to everyday life. He’s been pleasantly surprised in the interest his techniques have garnered, but he says that more data doesn’t necessarily mean you’ll clean up at the TAB.

“There seems to be a belief that if you have more data about teams and conditions, you will be able to greatly improve your predictions,” he says. “I don't believe that. More data and improved methods will make a bit of difference but there is definitely a limit, which varies from competition to competition.”

But don’t worry, retirement won’t mean the end of the posts – Associate Professor Scott is hoping to automate his predictions. “That means getting web-scraping working – but I haven’t had much success with that yet.”