One of the big problems for statistical analysis of soccer is defensive pressure. Opta tracks many things, but it does not track the location of opposing defenders. A shot taken from the same position, under minimal defensive pressure, is much more likely to score than a shot taken with two big center backs denying space. Until we have the equivalent of the NBA’s incredible SportVU system, it is necessary to create proxies for defensive pressure.
Dribbling is a good proxy. To illustrate the value of dribbling, I want to take a look at three goals. Two occurred during the recent World Cup, and the third was scored a week ago in the English Premier League.
First, Clint Dempsey kicked off America’s tournament with one of the fastest goals in World Cup history against Ghana.
The key to Dempsey’s goal was his quick dribble past Ghana defender John Boye. The location of the shot was only good, not great, as Dempsey was wide of the danger zone, but because he was under minimal defensive pressure, having dribbled past the last man, he has time to pick out a shot past the outstretched leg of Adam Kwarasey. Dempsey shoots from a location where only maybe 15 to 20 percent of shots are scored, but he has probably doubled his chance to 30 to 40 percent. The dribble is extremely valuable.
Second, we have Lionel Messi’s brilliant individual goal for Argentina against Bosnia-Herzegovina.
Messi creates a pocket of space for himself at the top of the 18-yard-box by dribbling past two defenders. He still has a fullback filling part of his shooting lane and other defenders quickly tracking back, plus a goalkeeper to beat. It’s a perfect little Messi finish just off the inside of the post that makes the goal. The dribble probably added a few percentage points in expectation, but it did not double his chance of scoring like Dempsey’s dribble did.
I can also illustrate this point using a larger data set. I have taken all shots taken with feet, not assisted by crosses in the English Premier League in the last two season. I broke them into two groups: Those that were immediately preceded by a dribble, and those that were not. Within each group, I broke the shots down by adjusted distance from goal. As you can see, at distances in the range of 10 to 15 yards from goal, a dribble adds hugely to expectation. At 25 yards from goal, the difference is quite small, though the dribble seems to help a little.
Finally, there is another, even better kind of dribble. In stoppage time against West Ham United last weekend, Tottenham Hotspur found a breakthrough goal. Right back Eric Dier, in his English debut, made a run behind the last defender, and Harry Kane found him with a through-ball. Matched one-on-one with West Ham keeper Adrian, Dier calmly dribbled around him and passed the ball into the back of the net.
Beating a defender to get space to shoot is good. Beating a defender and going one-on-one with the keeper is better. But nothing compares to beating the keeper and passing into an open net.
In my data set, I have 33 shots following a successful dribble past the keeper. Twenty were scored. Shots that missed came almost exclusively from very difficult angles. Shots from within or near the danger zone score at even higher rates.
A chance like Dier’s, then, has something like a three-in-four chance of being converted, with only the possibility of a clumsy shot or a clearance off the line standing in his way. His shooting location is no better than Dempsey’s, but he is probably about twice as likely to score.
The individual take-on, in traffic, is a difficult move to pull off. But when it can be done, in areas not too far from goal, it creates excellent scoring chances.
All data provided by Opta unless otherwise noted.