Have you looked at the Games Performance Assessment Instrument (GPAI)? This covers both technical and tactical aspects of team sports. Also, look at Daniel Memmert's article in the Journal of Teaching in Physical Education, 2008, 27, 220-240.

Thanks for the response dear Ajit Korgaokar. Nevertheless, my goal is not evaluate through a qualitative observational way, but yes through technological devices based on players' position at each instant. Do you know any automatic device such as centroid position, stretch index or surface area?

Dartfish Team proposes such a device to track players and to have a lot of interesting information for tactical purposes. see at:http://www.buydartfish.com/

I don't know if it possible to have all the index that you need, but a lot of these seems available. Good luck

If you can gain access to ProZone then you can get all of the information you need. The problem is that ProZone is very expensive and, typically, professional clubs are the ones with this equipment. Dartfish, as Guillaume suggested, is one of the more cost effective software programmes and is fairly easy to use. I would also look and see what Polar has, because I know they are developing tracking equipment.

Interesting question and the answer is not easy to formulate. Research work of Frencken in Groningen is interesting in this area.

Key is the playing strategy of the team. In this strategy the team, line and individual players have certain tasks within the four team functions. To give specific feedback to the coach you should analyze the data, e.g. from Amisco, in the context of the coach tactical plan. and key perfomance indicators.

Most systems don't have the option to analyze the data according to the 4 team functions and you can't create the info needed by the coach. General data, e.g.number of successful passes, number of challenges, are not very valuable.

Dear professor Guillaume Laffaye...thanks for your answer...this system suggested is interesting and I will go contact them to further information...Thanks again

Dear Professor Jos Dijk

Indeed, systems such as Prozone or Amisco do not provide tactical information, but yes notational results. Despite of their importance, the results is only a outcome of the process. Thus, my research interests relates to analyze the process (i.e., collective behaviour).

Some authors (e.g., Frencken et al., 2011; Lames, et al., 2008; Bartlett, et al., 2012) provide new qualitative metrics, nevertheless their discussion is about systemic analysis and the practical applications do not correspond to coaches expectations. Considering that my PhD project is about tactical metrics applied to soccer game, my main goal is develop some metrics that allow to coaches a practical and easy information about collective tactical behaviour.

Thank you for your interesting and esteemed participation

Dear Filipe,

You have to define and formulate the most important coach questions.

e.g the coach tactical philosophy is; the team organization should be compact in ball possession opponent (BPO)

A consequence is that the defense line will follow the direction of the ball. When the ball is played backwards the defensive line also has to step forward. The distances between the defensive line and the midfield should be less than 15 meters and so on.

So based on the action theory, you can define "best" action of a player or cluster of players within the team function BPO. A metric could be distance between defensive payers 3 / 4 and defensive midfield players 6 / 8 over game time and relate this to the moments opponent players 9 or 10 are able to choose position and get the ball between 3/4 and 9/10. This metric over time could be related to chances / goals of the opponent.

This is just an example and many more metrics could be constructed.

I think a sampling of what coaches 'teach' their teams in collective play in the 4 main moments in a game - In Possession, Defending, Transition to Attacking and Transition to Defending. Can it be measured and standardized for coach application? I would love hear more!

Dear Tony Kees Jr

These four main moments can be an support to develop new tactical metrics. The main issue to improve the tactical match analysis should be done listening all possible coaches, researchers and analysts. In Europe we have automatic tracking systems that provide a notational analysis. Nevertheless, these information is not enough to understand the collective behaviour. We knows that in USA the match analysis is very developed, but as far I know do not exist tactical metrics based on tracking systems. Such as suggested in previous discussions, these tactical metrics should be defined considering the main soccer moments as you said. But the real question is how many ideas we can collect to work together in order to answer to this issue for coaches and their staff.

Dear Filipe,

There are certainly a lot of solutions, and I guess you are innovative engough to find these ideas. In my mind, what is important for you is to distinguish (1) your “metric”, that is the indicator that captures the displacement (or a part of the overall movement) of a team, from (2) the “collective variable” that measures the relation between the metrics.

According to the first point, you may be creative, and, as an example, propose a “weighting” of each player (according to their position to the ball, or to their position to the goal… and so on). According to the second point, the relative phase has been often chosen to account for this relation, but certainly others variables may be instructive.

In all these cases, it would be interesting to compare the new description you propose with the usual indicators (centroid positions, stretch index…) and/or with significant events occurring in the game. I mean.. when you do a description of the game using indicators, we need to know what is the “reality” you describe.. so the indicator in itself is often not instructive.

I also think that it could be interesting to look at some subparts of the team (only 2 or 3 players.. or more), and investigate in which way the team cannot be reduced to a superorganism…

Yours..

Jérôme

Interestingly the response of Jerome. However, while there is already a number of interesting investigations that used the relative phase, I think that is yet to demonstrate the real practical application of this method to be used as "collective variable". I think this is more like a "binary variable"... and sport games, in the case football, are a bit more complex

I agree with Jerome.