The Missing Ingredient

Ages ago, Opta used to have an OptaPro blog (I wrote the first article).

I also wrote a blog about trying to utilise how close (or not) off target goal attempts came to requiring a save.

It was called something like "Don't be afraid to miss" and centred around data relating to Robin van Persie. (It was that long ago).

The site has long gone, but with access to more extensive data, such as shot placement (including off target attempts), I've revisited the idea to see if the intuition that "good finishers", when they miss, don't miss by much, is valid or not.

The idea is fairly basic. A shot that hits the post, is inches away from being a high quality post shot xG, whereas one that flies high and wide is going to need a fair bit of resighting to trouble the keeper.

The metric will also be related to the situation from which the attempt originated (open play, free kicks- etc), where on the field it came from (six yard box, outside the box) and whether the head or the boot was used.

So I took every off target non penalty attempt from the big five league for the last three completed seasons (over 53,000) and modelled by how far a typical big five player missed the target with their wayward efforts based around these pre-shot variables.

I then compared the "expected waywardness" to the actual waywardness of individual players for every play type scenario.

I expected Messi to come top. He didn't. He came second out of over 3200 players, although he did have three times as many errant efforts than the player who beat him (Matteo Politano). So Messi's number are more robust.

Here's the top ten players whose off target attempts are close enough to elicit an "Ohh" from the crowd, along with the ten players whose misses gave the goalframe the widest berth.

The lists seem to pass the eye test. Messi, de Bruyne and Son in one list and Maupay and Havertz in the other.

I took the 30 top post shavers and looked at their NPxG compared to their actual goals and it was a cumulative 489 NPxG compared to 556 actual goals, an over-performance of 13.7%.

The worst 30 hopelessly wayward had a cumulative NPxG of 434.9, but just 394 actual goals scored. An under-performance of 9.4%.

Roughly 40% of goal attempts are retrieved by the ball boy/girl. But rather than discarding that sizeable chunk of data, there might be good reason to at last try to gather some insight from these wayward efforts.

The Ageing of the Ageing Curve

A long, long, long, long time ago (2013) there wasn't that much granular data around and certainly hardly any based around metrics that eventually led to xG entering the mainstream of football/soccer analytics. Therefore, proxies abounded and share of playing time to evaluate a player's ageing rise and fall quickly became a go to method. Pull enough player data for minutes played, trawl through wiki for dob's and you came up with a pleasing curve that rose from the teenage years to the early twenties, peaked in the mid to late 20's and fell away as the ex pro went away to impart gnarled cliches on Sky. Plot the season by season change in playing time, the delta method & the linear trendline cut the axis where it was assumed that effortless push came to more laboured shove.

When a player's minutes finally fell instead of maintaining an ever shallower upward trend, the assumption was that he had become less effective on the field, performance levels had fallen, the manager had taken note and action had ensued. Physical decline, it was assumed, had begun to outstrip experience and smarts. We're now over a decade further down the road to analytical enlightenment, where on and off ball stuff gets routinely measured, even if there's still no consistency in naming metrics. Creativity, shooting execution & positional sense, ball progression via passes or carries & risk/reward has become less blurry & more transparent. We've also seen advances in sports science to prolong a the peak age of performance & witnessed anecdotal evidence for increased player longevity, even in demanding roles. So it's well overdue to usher "share of minutes played" into the lobby, treat it as just a fraction of what happens along the age curve & try to understand what might be going on in a player's career arc. Non shot expected goals added was one powerful metric that sought to measure by how much individual ball progression improved a team's likelihood of scoring. The delta approach to non shot xG added per 90 shows a gradual increase in performancein the early years of a player's career, but thereafter there's virtually no change, on average in the performance levels achieved per 90 minutes. It's not quite an expected trendline, more early improvement, but then flatlining.

The ball progression illustrated combines passes & carries and it's generally accepted that the latter is more physically demanding than the former. Perhaps players are replacing any shortfall of xG added via carries by upping their output from passes. To see what may be happening I looked at how NS xG added from just carries has changed over the last three completed seasons for all players as they age and again there's virtually no change on average in the rate of xG added from carries as a player ticks off their birthdays.

Persuasive speculation that as a player ages, their performance levels dip & they get left out of the side more frequently, kept "share of minutes played" a respected metric for nearly a decade. But how the new quantifiable metrics don't change that much well into a player's 30's may suggest that there is a slightly different dynamic at play for individuals, overall. Namely, player metrics, even more physically demanding ones which involve ball carrying, can with good managing of playing time enable players past what was considered their prime to maintain their own high standards. In short, you might get very similar levels of performance in a player's early 30's as you got in their mid 20's.....just not quite as often as you did previously.

Football Analytics' Big Own Goal

Just a small vent regarding what a poor job the early analytics community did and continue to do when naming metrics. I know it's been widely pointed out, but "expected goals" is an awful name for the premier metric and nothing screams elitism and jargon than almost always going to acronyms. xG, PSxG, xA, NSxG may be easily understood by anyone who has immersed themselves in the topic, but as someone who has tried to get these ideas to a wider, more general football obsessed audience, they are an immediate barrier. It's almost certainly too late to begin using titles that use everyday language *and* are self explanatory, chance quality, for example rather than expected goals or on target chance quality, rather than post shot expected goals, (which isn't even accurate!). If you need a glossary to write an article about a team or player, who've failed. If you need to write "expected goals, which is"........... The same. Numerical values and decimal places are usually enough to disengage otherwise passionate fans of a sport. Chuck in jargon and you're almost inviting a negative reaction, regardless of the points you are trying to highlight. Three rules of naming metrics. 1) Don't use acronyms. 2) Use familiar language, ideally associated with the sport. 3) DUA