Photos: Did this SAP big data software really help Germany win the World Cup?
Image 1 of 4
\N
Andre Schu00fcrrle’s cross and Mario Gu00f6tze’s chest and volley may have won the World Cup for Germany in extra time but enterprise software giant SAP is claiming a hand in the victory.
Its Match Insights software, which employs SAP HANA in-memory computing, was originally taken up by Bundesliga side 1899 TSG Hoffenheim, which is about 20km from SAP’s Walldorf headquarters near Heidelberg, south-west Germany.
In October last year, SAP and the German football association – or Deutsche Fuu00dfball-Bund – started collaborating on adapting the software for the German national team for final preparations for the World Cup and during the tournament itself. A prototype was delivered in March.
The image shown here is taken from the software running analytics on the November warm-up game against Italy. The white lines across the German backline and forwards show the operator in this case is studying playing formations and the relative position of players in these units.
In practice games, players can wear wireless sensors from partner company Amisco to relay geospatial and performance data but in real matches those digital tags are not permitted for safety reasons and because coaches are barred from using in-game technological aids.
According to SAP head of head of discrete manufacturing industries Nils Herzberg, the technology is not just for coaches. It can also be used by players to analyse their own performances.
“It is more for training but also for post mortems. Well, in the case of Germany there wasn’t one. Everyone else could have done with one,” Herzberg said.
The technology could also be used one day in other fields, such as in military exercises, for example.
“You could actually use it for the training of soldiers for special missions, if you want to make sure certain instructions and behaviours are followed,” Herzberg said.
\N
Although the software can perform a running analysis of the performance of both teams, a coach or player can also highlight an individual and zoom in on his match data.
Here you can see the operator picking out midfielder Bastian Schweinsteiger.
For every second of the game, some 5,000 data points are captured, yielding information on measures such as average possession time, speed travelled including bursts, distance travelled, player position and number of touches.
For real games, where digital tagging is not permitted, the data is collected using pitch-side cameras and facial-recognition software to identify individual players and map their position on a grid using geospatial data.
The user interface itself is a combination of Java and HTML5 and is plugged in to HANA by a JDBC-ODBC connection.
\N
This image shows that the player picked out for special study in this case is left wingback Marcell Jansen of Hamburger SV. His image appears at the top left of the screen, and he’s highlighted on the pitch by a small blue circle.
Across the top of the screen you can read Jansen’s passing accuracy, shooting, movement and dynamics, which also breaks down into fast accelerations and hard decelerations.
At the top right is a graphic showing the player’s ball-handling statistics, showing data such as ball contacts, possession and an overall rating.
The bar at the bottom of the screen is the match timeline, showing important events in the game, including goals as well as the individual’s notable contributions, such as shots and important defensive interventions.
\N
The grey areas fanning out behind the German players show the defensive zones they are potentially covering and into which they could run. This feature can be selected so that coaches can study the cohesion of, say, the backline or the defensive midfield.
Because the whole team is being studied in this case the top bar has reverted to team statistics, rather than those of an individual player but a similar visualisation for ball-handling appears along with an overall team rating.
