mailbag_onestat

This is a really good question and one that I find myself thinking about quite a bit as we at StatsBomb start interacting more with media and potentially with football clubs. For the most part, I think the answer depends on the audience.

In the last week, I have seen NPG90 (non-penalty goals per 90) used all over the damned place, including in a Paolo Bandini piece in The Guardian and Gabriel Marcotti either on ESPN or in the Wall Street Journal. This was definitely not happening a year ago. Smart journos know that this type of adjusted rate stat tells you a lot more about “useful” goal scoring than simply looking at who has the most goals in a league, and it’s gaining traction in the mainstream.

The owner of Squawka tweeted a leading goalscorer chart for the Premier League yesterday and a wave of derision followed after it. Not because people don’t like stats, but because it gave the same weight to penalties as to other goals. The world is changing fairly quickly on this at least.

I also like the key pass stat, since it adds a lot more information about which players are dangerous passers that are setting teammates up for shots versus those that are merely involved in the game. However, I think that particular stat is just a touch too abstract for popular acceptance right now. Europeans are barely aware of assists – moving one step beyond that into “passes that create shots” is probably a step too far.

At the team level, I think we’ll see ExpG (Expected Goals) take off like a rocket as teams become more analytically aware. The reason for this is because it creates clear points of actionable data for managers, players, and Directors of Football to talk about. Everyone involved in football knows there is a lot of luck in the game, but having metrics that try to cut through the luck and look at what actually happened in a match is extremely useful.

This is already happening with some teams in hockey. There was a story from Mirtle I read last week about coaches coming to the analysts immediately after the match to check on their PGS (probable goals scored) stats so that they could then have an objective point of reference to talk to the team and media after the game. Everyone involved in football should be evaluating via process more than via outcome, and I think you will see that happen more and more in the medium term.

None of the ExpG models are perfect yet, but they will become more precise in time, and as more data becomes available. In the meantime, what we have now is far better than nothing, or even what the world had access to a year ago.

With regard to conversion of fans, a lot of it comes down to comfort levels with numbers, and some of that involves simply seeing numbers used to compare players and teams on a regular basis. Fantasy football certainly seems to help with this (and I have been told by many FF fanatics that our site is SUPER helpful with picking their teams, despite the fact that none of us play), but so too does seeing useful stats appear regularly on Sky coverage and on BBC. (It helps that some of the same guys helping in the background on those shows are also fans of our site.)

People who do analytics in other sports seem to win the vast majority of their battles, and the media pieces written about basketball and hockey are completely unlike anything you would have seen 10-20 years ago. Football is tiptoeing in this direction, but this is probably the first year that is the case. At some point, literacy with stats will be a requirement for most intelligent sports writing, in America and in Europe as well.

Fast forward a decade from now, and baby steps that we’re taking in football analytics will probably seem perfectly normal stuff to talk about with a broader audience. In the meantime, tell your friends when you see stuff you like, and keep clicking and interacting with the writers that use stats to cheer them on. Almost none of these guys get paid, and they tend to operate on caffeine and kudos alone.

  • toshack

    So then, if I may: Kudos to you Ted, to Ben, Colin and the rest of the StatsBomb team.
    Keep up the good work! 🙂
    //Peter

    • http://blogs.columbian.com/portland-timbers/ Chris Gluck

      Interesting article Ted… and retweeted it to increase viewership. What I have found is that the combined view of a number of stats (and their overall relatiionship together) creates the best value for those wanting to see more relevance in statistical analysis… that has value to the end-state (winning games) — expected goals on their own is not representative and it misses a significant part of the game… defense…

  • Alex

    Owners/coaches might be starting to accept stats but I think most casual fans are still very, very hostile. At least for EPL fans I’ve heard they see advanced stats as Americans trying to come and ruin their beautiful game. I’m sure people will come around, but until then they’ll be a lot of arguments about players based on form and who “wanted it more”

    • tknutso

      Sure, but this is year 1. Hockey’s in like year 5 at least and waging major battles in the media about analytics with the Maple Leafs (and winning). Like I said though, change is happening already, perhaps in some ways even faster than many thought it would.

  • Ben

    I’d love to see less talk about possession; it seems like far too often we’ll get someone losing with 50%+ possession along with the comment that ‘numbers can’t tell the whole story’. So I’d be inclined to say that possession would be one of the most important metrics to consider when talking about statistical naysayers.

    • http://blogs.columbian.com/portland-timbers/ Chris Gluck

      Agreed Ben, possession in my research shows a few things of interest… a team like Toronto has much lower numbers in possession and quantity in other indicators as well and that speaks more to their systematic approach in attack – so while possession itself is not an indicator on who might win or lose it is a valuable indicator on what style a team plays too… what most get enamored with is the “next bright shiny object”… while expected goals has value it’s a bright shiny object for many in the soccer community because there is simply a ‘ton of work’ that needs to get done on a pitch before that expected goal (opportunity) can occur…
      even now sportscasters covering MLS are aware that possession is simply an indicator and that it doesn’t tell the story without considering other indicators in the game…

  • Errorr

    I’m mildly dubious that expected goals will be accepted by a wider public unless there is some marketing. The reason is that it is somewhat abstract and the actual words can lend itself to 2 different possible stats. ExpG could either be a proxy for luck or shot quality but I think the name would lend people to assume it is something like ‘accuracy of shooter’ and a proxy for some fuzzy idea.of clinicality. There may be something there but I think the ability to exceed some baseline level of accuracy is rare and not very important.

  • Geraint Morgan

    I will echo the fantasy thing, statsbomb has been very useful to me when picking my starting 11, especially in a league where it goes the more American style with each player being owned only once. The stuff on peak age of player performance that you were doing last year especially.
    If you lot did a piece next summer on players likely to make the jump for the next year (due to them ageing on, getting more minutes, change of team etc) or which imports to look out for, I promise I would read, reread and not tell any of my friends about it and instead keep all the goodness for me 🙂

  • Matt D

    Europeans are barely aware of assists, wut?

    I’m a stats convert but the condescending tone of this article is an example of what puts people off the stats community.

    • http://blogs.columbian.com/portland-timbers/ Chris Gluck

      in writing my blogs in america about the MLS, and the statistical processes i use to assess team performance, over half my readership, at times, comes from europe and the far east/africa – in considering their retweets and other comments; especially by a number of folks in the netherlands, you are right europeans are very much aware of assists, and indeed, recognize the value of statistics in assessing player and team performance… outside of the general information offered on expected shots and expected goals…