About Ranking Software (longer version)

Ranking Software started as a blog in 1994, back before blog was a word. There was a dedicated following of over a hundred direct email subscribers to the weekly tennis rankings for professional men & women, plus a wider audience in the rec.sport.tennis newsgroup. I was also writing about football and predicting AFL results, and had some success betting in both sports.

My name is Darren O’Shaughnessy, and back in 1994 I was pursuing a PhD in Theoretical Physics, specifically statistical analysis of quantum chromodynamics models on a lattice. In the early part of 1995 I probably spent more time analysing tennis than quarks. Even though I’d become a physicist to try and understand how the universe works at a fundamental level, it felt like my sports research was more connected to the real world than my lonely fully-funded research. The dirty secret of pure science in Australian universities is that most of it is not going to amount to anything, just help push the barrow for a staff member to continue in his/her position. I was quickly sick of toy models that were demonstrably nothing like the real noisy quantum universe. I was co-author of two papers in lattice gauge theory, neither of which I’d put on my resumé.

In sport you can get a better grasp of the truth. You don’t have virtual particles spawned in the billions blowing up your calculations, just some randomness that follows understandable mathematical trails. You can also enjoy the sheer athleticism of it, the determination to keep hitting that ball or chasing that opponent. The freak shot that landed on the line when Federer had no right to even try, and would only succeed 1 in 100 attempts if he tried it again.

In 1995 there was no such thing as sports analytics, and statistics in sport was just counting things and marvelling that Dominique Wilkins had 20,000 career points. Milestones are bland facts but they are achievements, and you can even hand them to a graphic designer and make a beautiful poster. As Ted Hopkins would say, to this day the main use of statistics is as propaganda.

I was very fortunate to work at the (now defunct) Australian Artificial Intelligence Institute from 1995-1999, learning the discipline of Analyst/Programmer and getting my hands dirty with agent-oriented architecture, parallel programming, neural networks, load balancing, and anything else that a business with 20 talented computer scientists could dream up.

By 1999 my tennis rankings were getting some attention from Professor Steve Clarke at Swinburne University and his research students, and I got to meet and talk with the legendary Ted Hopkins. Hopkins was best known as the 21-year-old who kicked four goals to help overturn a 44-point deficit for Carlton in the 1970 Grand Final. You can read his remarkable story in The Stats Revolution, an apt title. Through ten years at Champion Data, we were passionate about telling stories about sport through better recording of data, better analysis, and dynamic visualisation — what we now call informatics.

In 2009 I left Champion Data and Ranking Software became a consultancy business working 100% in sports analytics & informatics. Ted also left soon after, and we still collaborate on innovative sports offerings under his TedSport banner. I’ve been lucky enough to work with a lot of brilliant minds since branching out, with clients/partners including Hawthorn Football Club, SportsWizard, Tennis Australia, the AFL itself, ALBS.tv, the International Table Tennis Federation, Victorian Major Events, GLORY World Series (kickboxing), Racing Victoria, ASP World Tour (surfing), infoplum, North Melbourne FC, NRL, IMG, and Microsoft.

Research drives the business. Whether it’s a new way of delivering data visualisations for TV, adapting devices to record new types of data, a predictive model for a new sport, or simulating how to improve the player draft, we get to the physics first (causes) and think about how that gets reflected in who is winning (outcomes).

The extraordinary teaching culture at Hawthorn FC has also overhauled my thinking in the five seasons I’ve been working there. Presenting numbers to a mass audience is easy, because you can just highlight the outliers without having to delve into their causes. Working in a club environment with experts in the game, you cannot just point to an outcome and say “change that”. The team has to identify a pathway to that change. We are able to track every player on the field now, but working bottom-up from that detailed data will not have a meaningful impact on success. Someone who understands both the theory of the sport as it is described by experts, and the data, has to construct a semantic layer that bridges the gap between generic reports and actionable intelligence. Even more importantly, the club itself has to have an evidence-based and collaborative culture, which has been rare in sport.

Now it’s 2016, and my 9-year-old son has more WordPress experience than me. So it’s time to take the leap and start writing about some of these ideas, theories, and ways of looking at sport. Plus numbers and visualisations. I’m not the first — I’m probably one of the last loping behind some very talented people. I’ll write about them soon, too.