Performance Analytics: Where Performance Analysis and Soccer Analytics meet.
There’s a lot to be said for public soccer analytics work. A lot of it focuses on fairly basic things that (most) clubs and the media focus little on: the quality of shots taken, variance in performance and how shots are created.
Practically none of the work that I’ve seen recently marries together Performance Analysis and Soccer Analytics that well. All I’ve been able to find is that players are running distance X and sprinting at speed Y on the Performance Analysis side and the aforementioned shots and performance variation on the Soccer Analytics side. To me - that’s meaningless.
Performance Analytics should be a combination of the two of these: using the data from a players’ performance to drive decisions on the pitch.
I know that some of the Sport Science work in Soccer and the mass adoption of Catapult Sports has lead to “workloads” being found for individual players on a team. Teams can now analyse when a player should sit out from training, and when they could potentially be trying harder.
This is definitely useful to manage injury prevention and keep a team at peak fitness levels, but could these data and techniques be used to change what happens on the pitch?
For me there are various ways in which Performance Analytics - the combination of these two things - should work:
1. Optimising the time substitutions are made - and ensuring all three are used.
Teams should be squeezing every last drop of available energy out of a player - and substituting them as soon as their performance levels drop.
The timing of changes and the positions changed will undoubtedly vary team to team, therefore this would require an initial analysis by the team using match-by-match performance data. Previous research in this area by Colin Trainor can be found here and Jonathan Liew here can be used as a starting point.
It also does not make logical sense to finish a game having used only 12 out of 14 available players. Those two additional substitute players could be the difference between winning a game and drawing.
Now there are reasons that some changes need to be made, like a red card for a goalkeeper forcing a substitution (well, hopefully) or an injury, but in most games these situations will not arise - and mean that a team should have a pre-determined plan in place.
2. Teams should vary tactics to attack the changing weaknesses of the opposition.
Following on from the first example, if teams are able to look at tracking data for their opposition they will be able to see when their opponents begin to lose energy, in what positions and think about ways to exploit this.
This could be tracked by distance travelled across the game or something simple like pass completion split across the game (pass completion in 0–10 mins, 10–20 mins etc.). If a players passes become less accurate later in games, in a large enough sample - that’s a weakness.
For example, a left back that bombs on a lot in the first half may not have enough in the tank to perform at his maximum for the whole game. The team he plays for should look to either substitute him or accommodate him tactically so his position is not exposed - the opposition should try and exploit him.
To me this should be fairly easy to do - and would be a relative quick win compared to traditional means of opposition analysis using video. You could even look at adjusting an individual defensive actions per five minutes of the game to the number of attacks faced (so 0–5, 5–10, 10–15 etc.) to see whether a not a player makes less tackles or interceptions when the ball is near him or not.
3. Use the natural athletic ability of your players to benefit the situation within the match.
Picture this, you are 2–1 up in the 80th minute of a crucial play-off game and you are being bombarded by a team that are very exposed at the back, but playing very deep in your half. You have one substiution left. What do you do?
For me this is a time where I’d kill for a player like Aaron Kovar. Kovar is the successive winner of the Seattle Sounders squads pre-season bleep test. If anything says natural fitness, it’s being good at the bleep test.
Kovar would be great at closing down the opposition as they build up possession in our half. He’d also be great at moving the ball between defence and attack - leading to the potential to create a counter attack against the opposition.
Equally, going back to an earlier point about attacking the opponents weaknesses, if you have a very fast young attacker on the bench then it would be worth consider bringing them into the game to give them match time and exploit a backline that is disorganised and tired.
Finally - thinking forward to a recent example in the Premier League - can we move one of our larger, stronger defenders up front to take advantage of a tiring opposition backline by being a nuisance? Could the Impact replace Didier Drogba after 70 minutes with one of their centre-backs for the final 15/20 minutes to further tire out the opposition? Worth considering (and from a fan perspective, I’d love to see Laurent Ciman up top).
Soccer is a game of exploiting your oppositions weaknesses and I believe that the combination of analytics and performance analysis can help take teams a step further.
It’s also a sport of marginal gains, so if you’re not looking into ideas like the above to gain an edge over your opponent why aren’t you? It could be the difference between winning and losing.
Dollars and Decisions