Can Social Media Predict the Grand National Winner?
It’s Grand National time again – The one day of the year when millions of us cast aside our common sense instincts and instead fritter away our hard-earned money on an arbitrarily selected horse in the most gruelling and unpredictable race in the world.
I have backed the right horse on two occasions in what I have (somewhat depressingly), discovered is a 34 year old history of betting on the Grand National.
The first of these results was an each-way bet on Spartan Missile in 1981. Just to put the date into context, this was number one, Brixton was in riot and the computer mouse was about to be launched. Although I was 6 years old, this wasn’t a sign of precocious tipster talent as I waited a further 33 years for my next win: another each way bet on the eventual 2014 winner, Pineau De Rey.
So, how did I do it? Since I’m not 6 years-old anymore and am no longer choosing horses based on cool names (yes, I chose Spartan Missle because I loved the name), I had to find a more analytical way.
I chose Pineau De Rey because I used Social Intelligence to determine consumer sentiment and compared that with the horse’s odds to discover that it was a decent each-way bet. This decent bet turned out to be a winner.
At Synthesio, we are always interested in learning how people make decisions. There are many ways for us novices of choosing which horse to back, so we had a quick look at 2014 social content to see which were the favourites. Among the unskilled punter, having a cool name is the most important factor, certainly far more important than its odds, which goes some way to explaining why the bookies look forward to the National so much. It also explains why Shakalakaboomboom had the biggest share of voice in 2012.
Perhaps unsurprisingly, nobody is talking about combined Social Reputation Score/Odds analysis. So we will instead…
We harvested all mentions and analysed the sentiment of social media posts from the general public about each Grand National runner. We then compared the Social Reputation Score (SRS) each horse generated with its odds (at time of analysis) to generate a shortlist of options.
What we want to see is if social sentiment can be used as an extra in order to more accurately identify those horses most likely to make our Saturday joyful (and lucrative).
We are essentially looking for runners with high SRS and higher odds. These potentially offer the best return on your investment.
The chart below plots all runners according to their SRS and odds. Those in the top left hand side of the chart offer the most realistic return as “on the nose” and moving towards the right but staying in the top quadrant could be good options as each way bets.
So there we have it. A new way of identifying the winner? Maybe, but please don’t take my word for it. I don’t want you to have to wait another 33 years.