Whether you play prediction games like Superbru, bet on sport or just like messing around with algorithms – prediction metrics are curious things.
These days, data science is becoming increasingly sophisticated and accurate. Of course, many factors which we simply cannot factor into machine learning can sometimes turn things around, but there are always anomalies. The robots would have taken over years ago already if there weren’t.
And while we might not constantly be sitting around with spreadsheets trying to calculate which month of the year is likely to see record wins in the English Premier League (October and November, by the way), data predictions can be lots of fun.
With the 2019 Rugby World Cup final between England and South Africa looming, what would a week be without a bit of this prediction fun?
Dr. David Scott is a New Zealand statistician who has developed an intriguing formula for predicting not just the outcome of rugby matches, but also the likely points difference.
Throughout the 2019 World Cup, it’s been pretty accurate. Out of all the matches so far (excluding the cancelled ones), just four predictions have been wrong.
But there’s a bit more intrigue yet. Out of the 39 correctly predicted matches, just one prediction (Australia vs Wales) was spot on.
In total, 22 out of the 39 correctly predicted games have gotten the score correct within a zero to ten-point margin.
The fixtures that were incorrectly predicted are far more interesting though. Fiji, for example, were predicted to smash Uruguay by 30 points, but ended up losing by three.
Japan surprised the algorithm twice, against Ireland and Scotland. In the Japan vs Scotland fixture, the predicted margin (Scotland by seven) ended up being the margin by which Japan won. England, meanwhile, not only showed up the algorithm’s predicted winner (New Zealand) for their semi-final match, they also scored more points than the algorithm predicted the All Blacks would.
Rugby World Cup predictions by algorithm overview
Embed code for this graphic
And that brings us to this weekend’s prediction. It is a monumental occasion, emotionally loaded and historically significant. No algorithm can take that into account.
Finals are notoriously tight affairs, so it’s no surprise that the predicted margin is by just three points in favour of England.
But, if you’re a South African fan, as algorithms and history have shown before – what’s the fun in sport if predictions can’t sometimes be wrong?
You may republish this article – provided you comply with our Terms and Conditions. Please credit Extra Time Media/the author and add a link back to this article. You are not permitted to republish the cover image.