Showing posts with label methodology. Show all posts
Showing posts with label methodology. Show all posts
Saturday, 18 June 2022
Saturday, 26 June 2021
Saturday, 25 April 2020
Monday, 16 March 2015
Data is not the enemy
Ed Smith in Cricinfo
Taking the stats at your disposal into account does not mean your players cannot play a fearless and instinctive brand of cricket
The information is available to the players. It's their choice whether to use it or not © AFP
Poor Peter Moores couldn't have chosen a sentence more likely to turn him into a human dartboard. He needed to "look at the match data," Moores said after England's disastrous defeat to Bangladesh. To a press corps increasingly convinced that the England team has become formulaic and nerdy, this was the worst answer Moores could have reeled off.
It did, however, open up a whole new range of possibilities for the post-match interview. One wondered how historic sports interviews might have been different in the age of referencing match data.
Interviewer to George Foreman after he lost to Muhammad Ali in 1974: "What did you make of the fight, George?"
Foreman: "Haven't seen the fight data yet."
Interviewer: "Well, two men were standing up at the start of the eighth. Then there was one. You were on the canvas. In an algorithm: 2; 8; 1."
Foreman: "Right, got it."
And how would a data-inspired interview have run after the 1978 Oxford-Cambridge boat race, when Cambridge sank into the Thames, live on television?
Interviewer to Cambridge captain: "Disappointing race out there, I imagine?"
Captain: "Impossible to say before seeing the race data."
Interviewer: "Glug, glug, glug - ring any bells?"
All this mischief, however, does not explain very much why England crashed out of the World Cup. Are we really to believe that the central figure in the catastrophe was Nathan Leamon, the mathematician and former schoolteacher who is now England's stats analyst? That is ridiculous. It is Leamon's job to supply evidence that may help the management to make better decisions. It is the job of the coaching staff to use that information appropriately. So even if you believe, as I do, that England need to play a more fearless, instinctive brand of cricket, it does not follow that having access to potentially useful data prevents you from doing so.
The real problem is not maths, which by definition is flinty, pitiless, robust and unsentimental. No, the problem is management-speak, learned jargon and corporate-style snake oil. The unfortunate thing is that coaches can now use the phrase "match data" as just another thing to say when they are avoiding the subject. It slots into the lexicon of cliché, alongside "taking the positives", "skill sets" and "plan execution".
The irony is that real maths, in fact, is at the opposite end of the spectrum from jargon. Maths is exclusively content; jargon is content-free.
In his 1946 essay "Politics and the English Language", George Orwell despaired of political jargon: "As soon as certain topics are raised, the concrete melts into the abstract and no one seems able to think of turns of words that are not hackneyed: prose consists less and less of words chosen for the sake of meaning, and more and more of phrases tacked together like the sections of a prefabricated hen-house." Sport soon surpassed politics for meaningless waffle. Orwell deserves a new epitaph: "The man who foresaw the evolution of the post-match interview."
The former England captain Mike Atherton made several good points about the rush to blame data for all England's woes. First, analytics now has less influence over the team than it did in the more successful Flower-Strauss era. Secondly, Leamon's work is not pushed down players' throats. Stats for the particular ground, videos of an opposition bowler's range of slower balls, these things are available if players want to see them. If they don't, fine.
Nathan Leamon has a gift for numbers but he doesn't believe they provide all the answers © PA Photos
In 2006, while I was writing my book What Sport Tells Us About Life, I dedicated a whole chapter to the remarkable success of a school rugby team, unbeaten for three seasons, a sequence of 33 games. From the author's perspective, drawing lessons from a school coach was an unusual and risky approach. After all, other subjects of my book included Zinedine Zidane, Billy Beane and Michael Jordan. What was a school coach doing in that company?
The answer is that I thought his methodology was worth bringing to a wider audience. The coach was a plain-speaking, no-nonsense Lancastrian who pared down his comments to players. Rather than talking for talking's sake - as most coaches do - he researched what really happened in the matches and fed back small chunks of highly useful information. The quest was to find insight, concision and meaning; and to avoid noise, chatter and cliché.
Who was this progressive but unheard-of coach? Nathan Leamon. His approach, then and now, is thoughtful, flexible and open. His character is modest without being deferential, self-contained without being standoffish. Now, nine years on, he must find the way he is portrayed in the media as unrecognisable. Far from being a credulous geek, Leamon is an understated sceptic, a sensible and balanced man who happens to have a gift for numbers. Leamon is the last person to argue that data can provide all the answers - he's much too smart.
While England were exiting the World Cup, the retired NBA player Shane Battier was addressing a sold-out audience at the MIT Sloan Sports Analytics Conference in Boston. If you want to understand how data can help a sportsman perform better, read Michael Lewis on Battier's playing days.
Or this, exploring why it is so hard, with the naked eye, to understand the way that Mesut Özil, the Arsenal footballer, creates space on a football pitch.
Like it or not, as professional sport evolves it will provide greater scope for academic rigour. A lot of clever people like sport and they are constantly developing ideas - some good, others less good - that may eventually become absorbed into the mainstream. Bill James' understanding of data changed baseball forever. Eventually no coach could afford to ignore James' ideas because it would cost them games.
That is why the status of the sports analyst is going up. Nate Silver, who has become the most important analyst of American presidential elections, cut his teeth modelling sports matches. Ideas that originate in sport are finding wider application in the outside world.
And yet I am equally confident that a central task for sports coaches - now more than ever - is to liberate players, to free them from stifling anxiety and fearfulness.
Those two truths exist in parallel, not in conflict. Coaches will inevitably want to use every tool at their disposal, including relevant data. Then they must have the psychological nous, the feel and the common touch, to allow players to express themselves.
In the future of limited-overs cricket, can batsmen build dominant positions early enough to reduce the risk of getting out so much as to take it out of the equation? size: 900 © Getty Images
In the end, the discussion of data in sport tends to reveal more about prejudices than the underlying reality. It's all too easy to blame other people for using either too much or too little evidence en route to their decisions. I am intuitive, you are strangely convinced, he is delusional. Or, if you prefer, I am rational, you are a reductionist, he is a slave to numbers.
There is another story to emerge from this World Cup. The central innovation, which has now transferred from T20 to ODIs, is that talented batsmen are lethal - perhaps unstoppable - when they play without any fear of getting out. In 2003 I played in the first ever T20 league. I wrote at the time that it allowed players to play as they do in the nets, when they are totally uninhibited.
This powerful freedom, however, is partly earned by the match situation, especially in the middle overs. If the batting team is behind in the match, and there are few wickets in hand, it is far harder to bat as though another wicket wouldn't matter. The challenge now, in all white-ball cricket, is to build a position so dominant that there is no risk attached to getting out. If you are batting at 360 for 2 with seven overs left - with, say, Glenn Maxwell padded up and waiting - there is literally no risk in trying to hit a six and getting out. Paradoxically, of course, that makes you more likely to hit the ball for six!
Perhaps someone can show me the data on how quickly this underlying dominance turns into an impregnable lead. If I was coach, I'd certainly want to know.
Taking the stats at your disposal into account does not mean your players cannot play a fearless and instinctive brand of cricket
The information is available to the players. It's their choice whether to use it or not © AFPPoor Peter Moores couldn't have chosen a sentence more likely to turn him into a human dartboard. He needed to "look at the match data," Moores said after England's disastrous defeat to Bangladesh. To a press corps increasingly convinced that the England team has become formulaic and nerdy, this was the worst answer Moores could have reeled off.
It did, however, open up a whole new range of possibilities for the post-match interview. One wondered how historic sports interviews might have been different in the age of referencing match data.
Interviewer to George Foreman after he lost to Muhammad Ali in 1974: "What did you make of the fight, George?"
Foreman: "Haven't seen the fight data yet."
Interviewer: "Well, two men were standing up at the start of the eighth. Then there was one. You were on the canvas. In an algorithm: 2; 8; 1."
Foreman: "Right, got it."
And how would a data-inspired interview have run after the 1978 Oxford-Cambridge boat race, when Cambridge sank into the Thames, live on television?
Interviewer to Cambridge captain: "Disappointing race out there, I imagine?"
Captain: "Impossible to say before seeing the race data."
Interviewer: "Glug, glug, glug - ring any bells?"
All this mischief, however, does not explain very much why England crashed out of the World Cup. Are we really to believe that the central figure in the catastrophe was Nathan Leamon, the mathematician and former schoolteacher who is now England's stats analyst? That is ridiculous. It is Leamon's job to supply evidence that may help the management to make better decisions. It is the job of the coaching staff to use that information appropriately. So even if you believe, as I do, that England need to play a more fearless, instinctive brand of cricket, it does not follow that having access to potentially useful data prevents you from doing so.
The real problem is not maths, which by definition is flinty, pitiless, robust and unsentimental. No, the problem is management-speak, learned jargon and corporate-style snake oil. The unfortunate thing is that coaches can now use the phrase "match data" as just another thing to say when they are avoiding the subject. It slots into the lexicon of cliché, alongside "taking the positives", "skill sets" and "plan execution".
The irony is that real maths, in fact, is at the opposite end of the spectrum from jargon. Maths is exclusively content; jargon is content-free.
In his 1946 essay "Politics and the English Language", George Orwell despaired of political jargon: "As soon as certain topics are raised, the concrete melts into the abstract and no one seems able to think of turns of words that are not hackneyed: prose consists less and less of words chosen for the sake of meaning, and more and more of phrases tacked together like the sections of a prefabricated hen-house." Sport soon surpassed politics for meaningless waffle. Orwell deserves a new epitaph: "The man who foresaw the evolution of the post-match interview."
The former England captain Mike Atherton made several good points about the rush to blame data for all England's woes. First, analytics now has less influence over the team than it did in the more successful Flower-Strauss era. Secondly, Leamon's work is not pushed down players' throats. Stats for the particular ground, videos of an opposition bowler's range of slower balls, these things are available if players want to see them. If they don't, fine.
Nathan Leamon has a gift for numbers but he doesn't believe they provide all the answers © PA PhotosIn 2006, while I was writing my book What Sport Tells Us About Life, I dedicated a whole chapter to the remarkable success of a school rugby team, unbeaten for three seasons, a sequence of 33 games. From the author's perspective, drawing lessons from a school coach was an unusual and risky approach. After all, other subjects of my book included Zinedine Zidane, Billy Beane and Michael Jordan. What was a school coach doing in that company?
The answer is that I thought his methodology was worth bringing to a wider audience. The coach was a plain-speaking, no-nonsense Lancastrian who pared down his comments to players. Rather than talking for talking's sake - as most coaches do - he researched what really happened in the matches and fed back small chunks of highly useful information. The quest was to find insight, concision and meaning; and to avoid noise, chatter and cliché.
Who was this progressive but unheard-of coach? Nathan Leamon. His approach, then and now, is thoughtful, flexible and open. His character is modest without being deferential, self-contained without being standoffish. Now, nine years on, he must find the way he is portrayed in the media as unrecognisable. Far from being a credulous geek, Leamon is an understated sceptic, a sensible and balanced man who happens to have a gift for numbers. Leamon is the last person to argue that data can provide all the answers - he's much too smart.
While England were exiting the World Cup, the retired NBA player Shane Battier was addressing a sold-out audience at the MIT Sloan Sports Analytics Conference in Boston. If you want to understand how data can help a sportsman perform better, read Michael Lewis on Battier's playing days.
Or this, exploring why it is so hard, with the naked eye, to understand the way that Mesut Özil, the Arsenal footballer, creates space on a football pitch.
Like it or not, as professional sport evolves it will provide greater scope for academic rigour. A lot of clever people like sport and they are constantly developing ideas - some good, others less good - that may eventually become absorbed into the mainstream. Bill James' understanding of data changed baseball forever. Eventually no coach could afford to ignore James' ideas because it would cost them games.
That is why the status of the sports analyst is going up. Nate Silver, who has become the most important analyst of American presidential elections, cut his teeth modelling sports matches. Ideas that originate in sport are finding wider application in the outside world.
And yet I am equally confident that a central task for sports coaches - now more than ever - is to liberate players, to free them from stifling anxiety and fearfulness.
Those two truths exist in parallel, not in conflict. Coaches will inevitably want to use every tool at their disposal, including relevant data. Then they must have the psychological nous, the feel and the common touch, to allow players to express themselves.
In the future of limited-overs cricket, can batsmen build dominant positions early enough to reduce the risk of getting out so much as to take it out of the equation? size: 900 © Getty ImagesIn the end, the discussion of data in sport tends to reveal more about prejudices than the underlying reality. It's all too easy to blame other people for using either too much or too little evidence en route to their decisions. I am intuitive, you are strangely convinced, he is delusional. Or, if you prefer, I am rational, you are a reductionist, he is a slave to numbers.
There is another story to emerge from this World Cup. The central innovation, which has now transferred from T20 to ODIs, is that talented batsmen are lethal - perhaps unstoppable - when they play without any fear of getting out. In 2003 I played in the first ever T20 league. I wrote at the time that it allowed players to play as they do in the nets, when they are totally uninhibited.
This powerful freedom, however, is partly earned by the match situation, especially in the middle overs. If the batting team is behind in the match, and there are few wickets in hand, it is far harder to bat as though another wicket wouldn't matter. The challenge now, in all white-ball cricket, is to build a position so dominant that there is no risk attached to getting out. If you are batting at 360 for 2 with seven overs left - with, say, Glenn Maxwell padded up and waiting - there is literally no risk in trying to hit a six and getting out. Paradoxically, of course, that makes you more likely to hit the ball for six!
Perhaps someone can show me the data on how quickly this underlying dominance turns into an impregnable lead. If I was coach, I'd certainly want to know.
Subscribe to:
Posts (Atom)