Football is a simple game...

The saying, in England, is that football is a simple game: 22 players run around for 90 minutes and, in the end, the Germans always win. In the rest of the British Isles, I imagine the phrase is only slightly different: 'in the end, the English always lose'.

If you ask a certain group of nerds, of course, they have a different view: '22 players run around for 90 minutes and, in the end, rational thinking always loses'. In the broadest sense, this is the pitched battle that number-heads have been trying to have with football, stretching all the way back to the late '00s, Freakonomics era, like 2009's Soccernomics.

The introduction is titled 'Driving with a dashboard'; one of the early chapters, 'Gentlemen prefer blonds: how to avoid silly mistakes in the transfer market'. In my country, the book itself had a different name: 'Why England lose'. I hope, for their sales' sake, they kept the name in Scotland too.

This represents one of the two types of presentations at analytics conferences.

And, yes, after a lot of consideration (the ten minutes to write the opening three paragraphs) there are only two: 'here is a sophisticated use of data' and 'here is how to make organisations smarter'.

Take this week's Field of Play conference:

• 'Personal analysis: How elite players are gaining a tactical edge': Data use (although here 'data' means 'video')
• 'Designing for the decision, not the dashboard': Smarter organisation
• 'Unveiling true talent: Separating player skill from team context in football analytics': Data use
• 'The missing half of performance: the power of player grades in men’s and women’s football': Sales [Data use]
• 'How to be Hollywood smart with statistics: explaining the value of data and analytics to a Doubting Thomas': Smarter organisation
• 'Asking the right questions in women’s football': Smarter organisation
• 'Bridging analytics and tactics: turning benchmarks into on-pitch impact': Smarter organisation
• 'A model is only as good as it is understood': 20-minute stand-up set [Smarter organisation]
• 'Data Science meets NFL: Developing a winning project pipeline': Both, I think

One of the interesting things about this dichotomy is that on one hand you get work like 'can we identify good defending by running tracking data through a set of counterfactuals', and on the other hand you get talks like 'make peoples' lives easier and be a good hang'.

There is, though, a second set of nerds, who would sum up football in a different way. 'Football is a simple game: dynamical systems try to achieve a goal and, in the end, random variance has an annoyingly large impact'. (If you want a paper that has citations pointing to both pass completion models and work on starlings, here you go).

This would be the most accurate definition, on and off the pitch, even if it wouldn't pass the 'be a good hang' test. The perception of success for a club and its 'backroom' processes can hinge on a single match, can be disrupted by freak events, can be derailed by rogue individual decisions or mistakes.

'Trust the process' became a meme in the NBA, but one of the defining differences between sporting organisations isn't so much trust in a process, but the true understanding of what 'a process' is. A process, as in an infrastructure, rather than an assembly line. There's a fine line between wanting to reduce controllable uncertainty and wanting to deny the existence of uncontrollable uncertainty.

But, I think, one of the unspoken questions hanging over the 'smarter organisation' type of analytics talks is 'how do you know how good your process is?'. The 'sophisticated use of data' talks can at least point to a bunch of validation tests. Did the process work, or are positive results an unsustainable overperformance?

Maybe modern football needs a different maxim, to work for both the on- and off-pitch stuff. Football is a simple game: 22 players run around for 90 minutes and, in the end, some nerd is always asking to see the xG.