Statistical Analysis and The Media

There is no doubt statistical analysis is beginning to creep into the mainstream media but progress is quite slow. Over the weekend I got the chance to watch a few different sports and the examples of and understanding of statistics varied greatly.

1. Michael Atherton v David Gower

England had been bowled out and while we were between innings David Gower and Michael Atherton had a conversation about the current and likely level of spin in the match. David Gower said he had evidence that the amount of spin the Indian bowlers were getting had increased since Pietersen and Cook were out. He was probably right, but Atherton was trying to make the point that the level of spin is not only dictated by the pitch. Bowlers confidence and Batting performance also have an affect. Atherton claimed he had evidence to suggest that the level of spin against inform batsmen was different than against tail-enders – regardless of pitch condition. Both were right, one was trying to add slightly more context but it ended up sounding more like two men squabbling rather than any type of intelligent debate. In fact I flicked back on at the close of play (India lost 7 quick wickets) David Gower was trying to tell Atherton he was right all along and the pitch was spinning more now.

2. Liverpool v Swansea

The game is in the last 5 minutes. The game stands at 0 – 0. Sky Sports bring up the attempts on goal stats. It says 14 – 14. Cue commentator to say something like ‘and that statistic alone proves that the game is finely balanced and a draw is a fair result’. Within seconds of him finishing that sentence up pops – Shots on Target. Swansea 4 – 7 Liverpool. I thought he might change his mind here or even qualify his last statement but instead he says; ‘and nothing in that statistic either’. Last time I checked 7 was quite a lot bigger than 4??? I know they say these things in the heat of the moment – but really… this is just poor in my opinion.

3. Formula 1

As Martin Brundle did his customary pit walk the question on everyone’s mind was whether it was going to rain or not. During commentary – and the first rain spell – Martin Brundle referred to the boffins saying there was only a 40% chance of rain and then he said ‘they clearly got that wrong didn’t they!’. Now I think Martin is brilliant at what he does but I think there is a clear lack of understanding of what a 40% chance of rain means. I would take that to mean, given the same conditions 10 times, the meteorologists would expect it to rain 4 times and not rain 6 times. 40% chance of no rain does not mean there is no chance of rain.

I obviously love the use of data in all sports, and I believe it can really enhance the broadcast quality but perhaps we still have a long way to go.

  • Fraser

    I find more often than not it’s misleading.

    I cringe when i’m watching football and they bring up possession stats and the commentators see 70% v 30% and say one team’s dominating based purely on that. You’ve just watched them pass it about the back four for the first 10 minutes, the other team could be heavily pressing when it gets to midfield and forcing them back.

    Then of course comes territory on the pitch, which funnily enough will show the opposite if this is happening and they spout something about not having the penetrating passes into the final third. It wouldn’t take much to start showing time spent in opponents third etc. for a change and i’m sure the commentators would like some more telling statistics.

  • Neil Melville

    There was some interesting analysis of ‘Fergie time’ on the BBC’s More or Less stats program. There was an excerpt on the World Service that Robert might be able to access from Ireland. Otherwise here is the link to the full Radio 4 version:

  • Chris Carling

    Spot on on your comment about the L’pool Swansea game, a 50% on target rate is a bit better than 23.5%!!!