Knowing when the ball is out of play (“ball-out” events) in football is important in order to calculate effective playing time and the possession metrics of the two teams. But what do you do when you don’t have ball-out data? In this post I present a calibration tool to estimate match time lost due to […] Read more ›
Last night saw the start of England’s inevitable qualification for the finals of Euro 2016. Victory over Switzerland was a welcome opening result, achieved in fairly comfortable manner once the players took the advice of ITV’s summarizing commentator, … Read more ›
August sees the NFL gearing up for an early September start, so this month’s articles are split between American football and the association variety.
As well as partially sharing a name for their sport, the NFL and the EPL also return to centre stage at roughly the same time and share many of the often ignored traits which can influence the result of a sporting contest.
Seasonal results are often taken to wholly represent team ability, with all vestiges of luck erased, but this is rarely the case.
Does luck even itself out over a season? Not really.
Even a 38 game EPL season can see teams raised or lowered to finishing positions that do not fully reflect their true abilities, especially among the also-rans where teams of similar talent tend to cluster.
The NFL runs to just 16 games for those teams which fail to make the playoffs.
Being splendid one season in a statistical category that is strongly related to winning, in reality can also be a case of simply being both merely good and lucky. So there is often a need to separate the persistent from the transitory when using last year’s form to predict this years.
There is also an eagerness to draw definitive conclusions from early season results. So shot location and goal expectation models, which use increased sample data, can be used to simulate individual games, as well as the season to date, in an attempt to see if a team’s current league position is consistent with the shots they have created and faced so far.
Are Swansea performing better than Southampton just because they are above them in the table?
And Manchester City may gain some slight comfort from shot simulations that imply that, despite their lacklustre showing at home to Stoke, they should still have been beaten just once in every 10 trials.
On to the articles, click the links for the whole article.
How to predict the CFL Grey Cup winner
Canada plays football as well and despite the differences, it also shares many of the uncertainties with a close cousin, the NFL.
What are the HFA factors that may affect NFL betting?
Gaining an edge over the bookmaker using NFL HFA data
Ubiquitous in sport, but the NFL has a few additional wrinkles of its own. Would you really like to start you game day preparation at 6 am, whilst your opponents have a lie in or play in the noisiest stadium in sport after a long haul flight?
Which factors are likely to persist from last season and the myth of being handed a tough schedule.
July’s articles are available here.