What statistics fail to tell you about a player
How do you judge a player? In most sports the answer is by impression and word of mouth.
There is a story of a former England player watching the young New Zealand left-hander Bert Sutcliffe in a pre-season net at Lord’s in 1949. After twenty minutes or so, he said, “2000 runs this summer, no doubt about it.” He was right. Sutcliffe made around 2500 on the tour and his scores in the four Tests were 32,82, 57, 9, 101, 88 and 54.
On the other hand, such impressions can be wrong. One of my best friends at school was a graceful batsman with a delightful cover-drive. Watch him in the nets and he looked by some way the best bat we had. He never made many runs in school matches and slipped down the order, retaining his place because he was our wicketkeeper. Of course, the old England player who watched Sutcliffe in the nets was a far better judge than I was, and I don’t know whether my friend failed because of temperament or some technical weakness.
Cricket selectors have long had far more statistical evidence on which to base their decisions than selectors in other sports. But bare statistics can be deceptive. They do not, for instance, differentiate between runs made against moderate bowling on a perfect batting wicket and runs made in more difficult circumstances. This was why in the 1930s you might ask of a batsman “what has he done against Yorkshire?”
Or take the case of Doug Wright who bowled wrist-spin at medium-pace for Kent and England wither side of Hitler’s war. Go by his Test figures – 108 wickets at 39 – and you would never call him a great bowler. Yet both Wally Hammond and Don Bradman thought he was. He had the ability to bowl the well-nigh unplayable ball to dismiss even the best batsman when well set.
It is different now. In all sport there are statistics galore, at the upper levels anyway, and these are supported by analytic evidence. Surprisingly, football seems to be leading the way in the use of such analysis; I say “surprisingly” only because until recently there was very little of such evidence available to managers and selectors. Now there is almost a glut.
Eddie Jones remarked on this in an interesting interview with The Guardian this week. Football, he said, is ahead of other sports “in performance-metrics and statistical-based analysis”. Then he went on to speak of the England prop Mako Vunipola. “He carries the ball 15 times a game, his longest carry is probably three seconds.” (In which, I suppose, he may run ten or twelve yards at most.) Jones continued, “he has the ball in his hand for a total of forty-five seconds. That means he is working off the ball for 79 minutes and 15 seconds – and we don’t really have the metrics to measure the effectiveness of his movement off the ball.”
Of course, Jones’s figures aren’t quite right. For one thing the ball isn’t in play for the 79 minutes and 15 seconds Vunipola doesn’t have it in his hands. Time is spent strolling to lineouts and in setting up, and-all too often- resetting scrums. The ball is usually in play for, at most, half a match. Moreover, Jones does have other statistical evidence to hand. He knows, for instance, how many tackles Vunipola makes and his analysts can tell him how many of these are what we now call “dominant tackles” in which the ball-carrier is knocked backwards, and how many merely stop him or bring him to ground.
There can be little doubt that this is the way sport is going and equally little doubt that in the past decisions were made on impressions and little solid information. Often the best way of evaluating a player’s performance was to forget about the match you were watching and concentrate on the individual you were assessing. This was unavoidably imperfect. A player might do all the right things, making very few mistakes, and yet have little influence on a match. Even statistics may be deceptive, for they don’t necessarily reflect the quality of decision-making. You may learn that a player made a certain number of passes, but not whether he chose the easy option rather than the risky but potentially effective one.
Statistical-based analysis tells you how hard-working a player is but will often reveal little about the quality of the work. They don’t always account for the contribution of a player who seems to do very little for long periods of a match before making a decisive intervention. In football there have always been strikers like that. I can recall games in which Denis Law seemed almost invisible yet was on hand to score a couple of goals. The same might be said of Michael Owen. There too the statistics might suggest he should have worked harder; yet his apparent idleness might have created the conditions which made his goal-scoring possible. The busiest players show up well on the metrics but may contribute less to a win that the goal-poacher lurking in the shadows.
I have no doubt Eddie Jones will make good use of all the information that new analytic tools have made available to him. Yet, I am equally certain that his decisions about selection will be informed as much, even primarily, by his experience, observation and intuition. I doubt the best coaches in any sport trust information compiled by others and supplied to them as much as they trust their own perceptions. For one thing, they base some of their judgement on character, something that can’t be measured. David Leslie, the Scottish backrow forward of the 70s and early eighties, was neither very big nor very fast, and was also a reluctant, even lazy trainer. But the greatest of Scotland coaches, Jim Telfer, always wanted him in his team because of his indomitable attitude and because, as he once said to me, “I always know he’ll do the right thing.” But I am not sure he would have got top marks on any statistical-based analysis.