Published: Mar 18, 2015, 12:07 AM (IST)
Edited: Mar 18, 2015, 12:07 AM (IST)
Peter Moores has been lambasted for the early exit of England and his harping on data and statistics. However, while increasingly people find flaws with the data-based approach to coaching, Arunabha Sengupta says that the problem did not lie in concentrating on the numbers, but arose from poor inference and implementation.
Speaking in English and a poor heart
And then there was Peter Moores, the man who was supposed to coach the English team cricket. Last seen, he was drowning in a sea of indecipherable numbers.
To cap it all, when asked what went wrong he replied that he had to look at the data. That made him the target of wrath of a disappointed nation and its cricketing fraternity. After all, the data driven approach did not win the team too many fans and far less matches, especially as they floundered on the basic principles of modern day One-Day cricket. The coach could not have chosen a worse answer.
But then, we know all about the poor sod, don’t we? Let us divert ourselves with a fresh joke.
Have you heard about the American guy who was scared stiff of the increased incidence of heart attack in his land? So paranoid was he that to know about the risks of heart failure, he delved into statistics. He read, he took notes and he computed, although his friends told him that the key to good health was not research but a healthy life.
He discovered that heart attack rates were much lower in Japan. In the far eastern land, they seemed to eat less fat and consume less alcohol. And just before tuning his food habits, he looked at the data from France. The French also had far less reports of heart attack. But here was the caveat. They ate more fatty food and drank even more alcohol than the Americans.
The man was at a loss. He thought long and hard and inferred that there could be only one cause of heart attacks — speaking in English. He enrolled for German classes. His poor heart could not take this strain on top of the tension, and he died of severe myocardial infarction (Herzinfarkt, said his teacher).
His friends sagely nodded their heads. We told him so, they said. The secret of long life was healthy living, not analysis of data.
Stupid old fool.
He tried to number-crunch his way out of heart disease. Perhaps he also hired a stats analyst like Nathan Leamon, to accumulate, categorise and present the data. A fat lot of good it did him, didn’t it? After all, everyone knows that the key to staying clear of heart disease was a healthy life.
Well, let us look again.
There is no denying that what the man did was stupid. But what exactly was stupid in what he did?
Was it stupid to look at the data of heart attacks, to link it with food habits? I don’t think so. Without the data, one could not really decide which way of living was healthy. One could always proclaim one way was better than the other. And then one could just hope that one was right.
The stupidity lay in the interpretation and inference. The obvious root causes that he looked for did not seem correct, and hence he came up with a ridiculous suggestion. Neither did the data have anything to do with it, nor did Nathan Leamon. It just underlined one basic fact of life. Analysing and inferring from data are not very easy tasks.
Was there any other example of stupidity in the story? I would say so — if not stupidity, at least of blatant ignorance. It was in the sagacious nods of heads of the friends, who decided that data was not the solution.
Data — analysis and inference
Without data, most of the medical remedies we have today would not have existed. It is data that drives clinical trials. And for that, there are trained statisticians who carry out the inference.
It is the same with plenty of other industries and endeavours in human civilisation, including a lot of sports. Cricket, increasingly employing data in its decision making, is very much on the correct track. It is actually high time that cricket caught up with other sports in this regard. The game has depended too long on gut-feel and hunches.
However, the task of inferring from data is not easy. Take a sample population of 100 random cricket fans and ask them individually to explain the concept of net run rate. You will get what I mean after the eleventh attempt. Facility with numbers is not very common. And therefore, if one hears numbers cited when asked about the game one loves, it can be frustrating.
After all, mathematics is perhaps the only subject in the world people proudly proclaim of ‘not’ knowing. And hence there is this tearing hurry to shoot down the very concept of data-based planning and coaching.
That is also the very reason most people who grapple with the concept of net-run rate conclude that the Duckworth-Lewis (D/L) method is nonsense. Exponential decay function involving three variables is difficult to decipher. The truth is that D/L still happens to be the fairest result we can get in foul weather. The limits are most often in the problems of understanding and interpreting data that plagues the common man.
A coach, and that includes Peter Moores, is often not ideally skilled to infer correct solutions from the numbers. There needs to be a specialist eye. And if the inferences are wrong, one cannot blame the data.
Ed Smith put it so succinctly in a recent article in ESPNCricinfo: “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.”
If England played horribly, made quixotic decisions and fell flat on their collective faces, it had to do with poor execution. The data was there to be used, to take verifiable decisions, Nathan Leamon was there to present it to the team. However, it was up to Moores and the think tank to interpret and implement. And they made a hash of it. If Moores now harps on data it is bound to infuriate fans. It has nothing to do with the concept of using numbers for analysis. It has everything to do with mismanagement of numbers and then using it as an excuse.
That Moores is rather uninitiated to the world of data-based decision making is readily understandable from his comments: “We thought 275 was chaseable. We have to look at the data.”
Data can be used to help one analyse, help one strategise, help one take excellent decisions, identify strong and weak points, thresh out the most optimal approach. Data cannot be used to predict one single outcome.
Every statistical estimate comes with a degree of uncertainty. One can say that this is how the outcomes will be distributed over a number of trials. However, one cannot be certain that this is what an outcome will be.
Else, Don Bradman would never have scored a duck in his final innings.
Jet lag………Data tells me it should only last a couple of days……
— Michael Vaughan (@MichaelVaughan) March 9, 2015
(Incidentally, that holds true for Michael Vaughan’s jet lag as well. Data cannot say how long it will last. It can just tell us how long it normally lasts).
Moores’ statements demonstrate that he struggles with numbers. He misinterprets, looks at absurd inferences. If he looked at the heart attack charts of Japan and France, he would have probably enrolled in Berlitz for Deutschkurs.
If Moores had concentrated on cricket, with a more mathematically oriented mind performing the tasks of inferring and suggesting alternatives; if he had restricted his answers in the press conferences to areas of the game he was familiar with, perhaps the English debacle could have been less disastrous, and the stick that he is getting would have been less painful.
That does not mean the data is useless. It is anything but. It obviously cannot be a replacement for basic cricket skills, but it can be used to power those skills into optimal performance.
If one decides that the story of England’s pathetic journey is proof of the irrelevance of data, it is akin to turning one’s back towards science by clinging on to age old beliefs. If coaches are denied the use of the most powerful analysing tools in the game, it will take cricket backwards by several decades. Bill James has already put baseball far ahead in this regard. It is high time for cricket to catch up.
It is far more prudent to give the coach sufficient grounding, or helping heads, to make true sense of the numbers and perform correct analysis and inference.
Equally helpful perhaps will be some solid guidelines about answering questions of the pressmen after an unmitigated disaster.
(Arunabha Sengupta is a cricket historian and Chief Cricket Writer at CricketCountry. He writes about the history and the romance of the game, punctuated often by opinions about modern day cricket, while his post-graduate degree in statistics peeps through in occasional analytical pieces. The author of three novels, he can be followed on Twitter at http://twitter.com/senantix)
