Speaking of book releases from important figures in world football, Dennis Bergkamp has been making the rounds this past week doing interviews for his memoir, Stillness and Speed.

Ahead of the launch, Bergkamp conducted a series of interviews, including a short illuminating talk with the Guardian’s Amy Lawrence. In it, he discusses his views on best practices in player development:

The glory, for him, is all about control and touch.

Can that be taught? “The basics for me is the first touch,” he says, as if a perfect first touch is some kind of alchemy. “First touch in football is so important. If you talk about Mesut Özil people say he is not marked properly, he always has a lot of space but he has got that space because he can create space by his vision and his first touch. With that you create your own time.”

It is quite an arresting concept, creating time with a moving ball. “Teach that to children,” he says. “Do something with the ball, let it bounce, back, back, back against the wall, left, right, that’s the main thing.”

Bergkamp then goes on to criticize today’s emphasis on elite, technical coaching at all levels as the preferred means to produce world class footballers. He argues that it stifles the development of football intelligence. “If [young footballers] get a new situation, they look to someone as if to say, ‘What do I have to do now?’ I believe that is over-coaching. It’s too much. Let them have their freedom. You have to create the environment where they can be unique and not a clone.”

What’s interesting to me at least is how well this jibes with what we know for example about early childhood development (my wife is an early childhood educator). Leading research shows that play-based learning—in which children are given situations to self-direct their play as opposed to being constrained by systematic, rote exercises—is far more effective at fostering insight and creativity as children grow older.

This is part of the reason why I am skeptical of the utility of using individual player metrics in football to evaluate prospects unless some very strict criteria are met first (I wrote about this a few weeks ago). After all, as a scout, how do you measure a player’s capacity for improvement, or decision making off the ball, or confidence in individual technique? In this situation, surely the intuition of the scout to see things like football intelligence and technique trumps any stat you can come up with, right?

Well, not exactly. For example, as Nobel-prize winning economist Daniel Kahneman wrote in probably the most important book of the 21st century Thinking, Fast and Slow (exaggerating but not by much), in the 1950s psychologist Paul Meehl demonstrated fairly ably that in low validity environments, where uncertainty and unpredictability characterize the outcomes, “the accuracy of experts was matched or exceeded by a simple algorithm.”

Kahneman however notes that American psychologist Robyn Dawes later discovered that the algorithm(s) need not be particularly complex in order to perform better than intuition.

“However, Dawes observed that the complex statistical algorithm adds little or no value. One can do just as well by selecting a set of scores that have some validity for predicting the outcome and adjusting the values to make them comparable (by using standard scores or ranks). A formula that combines these predictors with equal weights is likely to be just as accurate in predicting new cases as the multiple-regression formula that was optimal in the original sample. More recent research went further: formiulas that assign equal weights to all the predictors are often superior, because they are not affected by accidents of sampling.

The surprising success of equal-weighting schemes has an important practical implication: it is possible to develop useful algorithms without any prior statistical research. Simple equally weighted formulas based on existing statistics or on common sense are often very good predictors of significant outcomes.”

The example Kahneman uses here is the Apgar score, which uses five simple, separate variables and an easy scoring system for obstetricians to assess the health of newborn babies.

This is remarkable for a few reasons. It emphasizes that practical analytics need not necessarily involve complex formulas or data points to be effective. I would be fairly shocked for example if most football scouts did not use something similar, say, a common sense score sheet with separate, distinct categories like pace, intelligence, off the ball movement, etc. Some of these may not need to be particularly complex, and they could be improved with research into probability for success for players with X per90 metrics, for example. It doesn’t preclude an in-depth knowledge of statistics, and could be viewed as a simple way to organize some of the characteristics scouts already look for in a player.

Would such a method be fail safe? Of course not. Would it provide more reliable judgments than a competitor who has convinced themselves their personal footballing expertise is good enough to decide who to recommend the manager sign? According to Meehl, absolutely. Would it be so specific in its requirements that it would miss ineffable qualities as those described by Dennis Bergkamp? Probably not. A scout may need only to come up with a category—say, Ball Retention—and a three point score.

A decent scout shouldn’t work for a club with a manager who demands a 100% success rate in picking players anyway. Any smart club knows the game is risk management, which precludes there will be some failure. Any smart club will recognize their loss aversion for what it is and instead focus on a streamlined process in player recruitment with an accepted level of risk. Like Bergkamp, a good scout needs the trust and freedom from their club to flourish.

Rolling Total Shots Ratio/PDO Table for the Premier League

Total shots ratio is shots for/(shots for + shots against), and it calculates shot dominance. It correlates very well to points finish in the table…for a better explanation, read this. In a few weeks’ time, we will include an expected points total in the table.

PDO meanwhile is just (sh% + sv%)*1000, and it is a rough measurement for luck, as these statistics tend to regress to the mean over time (except at the extreme positive end). A team with a low PDO should be expected to improve as the season progresses.

Finally, final third touches are a good predictor of future goals scored. Grayson has a post on the matter here, which breaks down the work of Mark Taylor.

Team P Pts TSR PDO Sh% Sv% Final 3rd Touches
Arsenal 8 19 0.543 1090 34.0 75.0 1682
Chelsea 8 17 0.644 1074 29.2 78.3 1561
Liverpool 8 17 0.494 1097 27.7 82.1 1598
Manchester City 8 16 0.635 1025 38.5 64.0 1708
Tottenham 8 16 0.652 978 17.0 80.8 1720
Southampton 8 15 0.550 1095 21.1 88.5 1411
Everton 8 15 0.562 888 27.3 61.5 1593
Manchester United 8 11 0.537 953 25.6 69.7 1726
Hull 8 11 0.390 1068 31.8 75.0 1025
Newcastle United 8 11 0.561 886 27.5 61.1 1272
Swansea 8 10 0.539 900 25.5 64.5 1485
West Bromwich Albion 8 10 0.476 1086 28.0 80.6 1284
Aston Villa 8 10 0.490 1034 34.6 68.8 1068
Fulham 8 10 0.305 1129 34.6 78.3 1059
West Ham 8 8 0.463 1131 38.1 75.0 1236
Stoke 8 8 0.477 954 13.8 81.6 1323
Cardiff 8 8 0.370 1145 40.0 74.5 1075
Norwich 8 7 0.407 930 24.0 69.0 1356
Crystal Palace 8 3 0.428 881 28.6 59.5 1154
Sunderland 8 1 0.476 787 26.3 52.4 1226