Slightly surprised this doesn't say "Stats Twats (virgins)" tbh. pic.twitter.com/K0MDA9SCHk
— Ted Knutson (@mixedknuts) August 28, 2017
So this stupid meme got me thinking about something: why do so many people think that “metrics” (analytics) and “positional play” (traditional tactics) form a natural dichotomy in football?
One reason may be because we know the two don’t always align. For example, a tactics-minded person may make a perfectly plausible argument for why a manager is responsible for a poor run of results, while an analytics-minded person might point to their Expected Goals model to argue those results are more down to bad luck than bad performances. In this instance, the analyst is probably correct: the xG model is likely to be a better predictor of future performance.
But this example–which is regularly trotted out time and time again by writers like me–probably doesn’t actually come up all that often at most clubs. Instead, you might have one key decision-maker who consults with various stakeholders–scouts, agents, players, first team coaches, analysts–before taking this or that action, whether that means utilizing a certain set of tactics or recruiting a certain player.
In cases where there is disagreement, some decisions may go in favour of the traditionalist, and some in favour of the analyst, depending on the circumstances. The success rate for both may vary and involve a lot of luck, making it difficult to tell whether the stats-minded or tactics-minded stakeholder is better at making informed decisions. And, of course, there may be many instances in which the two mostly agree (which makes the decision-maker’s job somewhat easier).
While this roundtable model is often touted as a decent ‘middle path’ between traditional club management and data analytics, I actually think there is a better, more consistent and less luck-prone way for clubs to accommodate both worldviews–the tactical and analytical–within a single club.
That model involves distinguishing between conditions and causes for winning. This distinction is brought up a lot in Buddhism, for example. Buddhists will sometimes point out that meditation, mindfulness and right speech/action/livelihood don’t cause enlightenment, but form its ideal conditions. The metaphor they use to illustrate this is sleep: you can’t cause yourself to fall asleep (except maybe by downing a bottle of sleeping pills); rather, through turning off the lights, tucking into a cool bed and turning off your phone, you set the conditions for sleep to occur.
In the same way, statistical analysis, which tends to be more predictive of long term rather than short term trends, can provide clubs with a blueprint for the basic conditions for winning– a decent xG ratio, taking deadlier shots more consistently, buying players who are more likely to peak at your club, etc. Yet analytics is more agnostic when it comes to the more proximate causes of winning, as in: what tactics are best suited to creating more dangerous chances with the squad at our disposal? How do we best negotiate for our preferred set of transfer targets within the limits of our wage/fee budget? Or: what scouting factors become important after you’ve narrowed your list based on excluding certain candidates based on statistically predictive factors (decline in performances, luck in finishing etc.)?
In this way, analytics can help build the conditions for winning, while traditional coaching, tactics or “positional play”–what it takes to win on any given day, the stuff we like to read about in the papers–forms the cause.
Rather than one side necessarily influencing the other, dialogue should go both ways. Key decisions wouldn’t rely on arbitrary consultation with various club staff, but would follow a clearly delineated process. This of course necessitates a certain level of trust and humility on the part of coaches and analysts, which may not be possible at a certain level in club football. And there will naturally be disagreements. But at least in this way, both sides would play more naturally to their strengths, and need not be in constant opposition to each other, or fight for influence over a single decision-maker.