It is no secret that training should mimic the game for which one is training. The specificity principle dictates that the training for tennis differs from the training for soccer, football, basketball, etc. Those of us who have attended any one of a number of coaching schools learned that training for soccer (or any sport for that matter) is the delicate interaction of fitness (physical and psychological), technique and tactics. Technique I understand. Repetitive application of new skills allows us to relegate the performance of those skills to a subconscious level; you just do the right skill for the right time when you have to - you don't really think about it. Tactics I also understand. These are all the strategic maneuvers that players are taught in order to score at least 1 more goal than the opponents. Of course, you may know what to do, but if you can't trap a ball, you can't execute the tactics. By the same token, you may know what to do and how to do it, but can't perform the skills and tactics the way you want if you are fatigued. Then you might be able to run all day, but can't trap the ball. Or you can run and run, control the ball, but can't decide what to do with the ball when you get it. You get the picture.

            An organized training program prepares the players for all aspects of the game. But how are practices organized to make sure the players are ready for what goes on once the whistle blows?  Let's see, I want my players to be good take-on artists, so we do lots of 1 v 1. But the game is 10 field players vs. 10 field players, so we should play full team, full field. But 10 v 10 prepares a team to play the same 10 v 10 they see in practice and when you meet a team who is more prepared, especially more physically prepared, your team is at a disadvantage.

            Let's move back one step. What is the point of the game? Obviously, to score 1 more goal than the opponent. Thus, shouldn't we know just what goes into a possession that leads to the chance of a goal? This is not American Football. Most all plays in football, if perfectly executed, should score a touchdown. We know that every possession in soccer doesn't lead to a shot or a goal. Therefore, to train players to attack, we should know what are the typical features of a possession that leads to a shot. When we know that, then we can follow the specificity principle and train our players to recognize the clues that suggest a shot is possible.

Characteristics of Shooting Possessions

            First, realize that the statistics I will describe are based on professional and national team level play, but the nature of the game is the same regardless of the level of play. Once players learn the game and we as coaches can challenge our young players to think the game, the principles of play are the same. Execution is the difference. What I will describe are things you probably know, but never had any numbers to justify your perceptions.

The landmark work of Reep and Benjamin (1968) showed that the overwhelming majority of possessions in soccer involved 3 passes or less regardless of the level of play. Jinshan and co-workers (1993) reported on the characteristics of goals from the 1990 World Cup and showed that nearly 70% of all goals were scored in the 2^nd half as well as about 1/3 of all goals came off set-plays. From the same championship, Yamanaka et. al. (1993) documented what might be described as 'cultural differences' in patterns of play between British, European, South American and African teams. Lanham (1993) pointed out the frequency of possession change (overall about 180 possessions per goal) and Bate (1988) discussed the use of direct play and the potential for shooting and scoring. Differences between collegiate soccer and the World Cup show greater turnover of possessions in the lesser skilled college players with possessions of fewer consecutive passes (Partridge, 1993). Overall tactical patterns were reduced to 2 factors by Pollard (1988) that were termed 'degree of elaboration' and 'use of centers.' They distinguished teams according to their use of direct play vs. an elaborate style and their use of centering passes. For example, teams that were high on long passes were Sheffield Wednesday and Watford while teams low on long forward passes were France and Brazil. The opposite was the case for multipassing movements. On all their factors, France was low in long forward passes, long goal clearances and centering passes and high on regaining possession in attack, possession in defense and multi-passing movements. More recently, Garganta (1997) studied 5 top-level European professional teams (Porto, Barcelona, Bayern Munich, Inter Milan and PSG). Well over 50% of all goal-scoring possessions were under 10 seconds in duration, involved 3 players and 3 passes or less with possession being obtained in the offensive 1/3 of the field.

The general goal of most all possessions in soccer is an attempt at scoring a goal. How a team manages to develop a scoring chance is sometimes thought to be somewhat cultural (Yamanaka, 1993). Coaches need to prepare players to be ready to attack when the opportunity presents itself. To do so, characteristics of possessions need to be described so players might be better able to recognize these opportunities. Further, the emergence of the women's game requires that their game be studied to see if there are differences with the men's game which would then mean there might be some tactical considerations unique for each gender.

If we subscribe to the specificity principle, then we better know the characteristics of shooting possessions so we can train our players to recognize these situations. To do this, we chose to compare an equal number of games at the same level of the elimination rounds from the men's World Cup (1998) and the Women's World Cup (1999) and describe some characteristics of shooting possessions.

So here is what we did. We recorded international television network feed of all games from the quarterfinals, semifinals and finals from both the men's World Cup in 1998 and the Women's World Cup in 1999. This represented 7 games for each competition (table 1). In the women's Cup, the first 40 minutes of one quarterfinal game (Nigeria:Brazil) was not broadcast because the previously scheduled sporting event on the television station (an automobile race) went beyond its expected length. To make up for this lost game, a game from the round robin portion of the tournament (Germany:Brazil) that involved 2 eventual quarterfinalists was used. In the men's competition, the tape quality of 2 quarterfinal games (Italy:France and Brazil:Denmark) was poor and could not be analyzed. Two games from the round of 16 were chosen randomly as substitutes (Romania:Croatia and England:Argentina).  Obviously, these data are weighted due to multiple games by each set of finalists (men: France-Brazil; women: USA-China). Each game was then reduced visually and attacking possessions that lead to a shot were then categorized according to the following characteristics:

Where the possession began (offensive, middle or defensive third; right, middle, or left third. This divided the field into 9 sections)

Time of shot

How possession started (throw, tackle, restart, penalty kick, interception, header, goal kick, free ball, corner, goalkeeper)

Number of players involved in the possession

Number of passes in the possession

Location of the last pass (to the near post, to the far post, to the penalty area, to the goal area, to the penalty spot, midfield area (outside of the 18) )

Shot from where (in goal box, in penalty area, outside of penalty area, far right, far left)

Shot with: head, right foot, left foot

Touches: 1 touch, 2 touches, more than 2 touches

Outcome: block, save, hit the post, hit the bar, missed near post, missed far post, shot high, mis-kick, goal

What did we find out?

Goals:Shots

            The men scored 21 goals on 231 shots while the women scored 23 goals on 176 shots. For the men, this represented a 11:1 shots to goal ratio. For the women, this represented a 7.6:1 shots to goal ratio. Overall, there were 44 goals and 407 shots for a shots-to- goal ratio of 9.25 to 1. There were 4 penalty kicks each for men and women, all of which were successful. While the long accepted 10 shots per goal, a decades old ratio,  is rarely accurate for any individual game or player, for multiple games, as in both Cups, the general 10:1 ratio seems to still be appropriate. 

Time a goal was scored

            The time of each shot and goal was recorded from the running clock that was constantly shown throughout the game. Times were then reduced to 5-minute segments. Figure 1 shows when (non-penalty kick) goals were scored for these matches. There were some noticeable differences between the men's and the women's game. For example, 36% of all goals in the women's game were scored in the first 20 minutes while only 17% of the goals in the men's game were scored in the first 20 minutes. The first half had 47% of all goals scored by the women while the men scored only 23% of all their goals in the first half. For the last 20 minutes of the game, the men scored 41% of all their goals, but the women scored only 16% of their goals. An interesting finding was that the men scored over 1/3 of all goals (35%) in the first 5 minutes of the second half. The data show that goals in the women's game are concentrated in the first 20-25 minutes of each half while the goals in the men's game were most common at the start of the second half and the end of the game.

Fig1

Time of goal by sex

Free Play vs. Restarts

            It has been reported that anywhere from 40-65% of goals in the men's game come from restarts; either from fouls, corner kicks or throw-ins. The 8 penalty kicks were excluded in this analysis leaving 17 goals for the men and 19 for the women. Men scored 11 of the 17 goals (65%) from free play and the remaining from restarts. The women scored 15 of their 19 goals (79%) from free play. Goals from a restart that had over 2 passes were considered to have been from free play. Thus, from these two competitions, the majority of goals came from creative free play and not from planned attacks from a restart.

Players involved more than once in a possession

            We looked to see whether a player was involved more than once in any play that led to a shot. Overall, a player was involved multiple times in any play 21% of the time. For males, the percentage was 25% and for females, this was 16%. For possessions that resulted in a goal, the percentage was 11.8% and 15.8% for males and females respectively.

Number of touches by shooting player

            A challenge for coaches is to design activities that mimic the game. Therefore, we noted if a shot at goal required 1, 2 or more than 2 touches on the ball. Overall, 56.9% of all shots required only 1 touch. Obviously, shots with the head made up a large component of this analysis in these games and 14.7% of non-penalty kick shots used the head. When shots with the head are removed, 49.4% of all shots were taken with just 1 touch, 23.5% required 2 touches and 27.1% were set up by dribbling. If we look just at goals scored, 58% of the goals scored by women were from 1 touch vs. 47% by the men; for 2 touches, women scored 26.3% vs. 35.3% for men. Goals from dribbling were about the same: 15.8% for women and 17.6% for men. Shots taken within the penalty area and goal box were overwhelmingly 1 touch. Nearly 70% of all shots in the penalty area were 1 touch and over 85% of shots in the goal box were 1 touch. Outside the penalty area 45% were 1 touch, 22% were 2 touch and 33% were over 2 touches.

Origin of the attacking possession

            The field was arbitrarily divided into thirds as offensive, middle and defensive thirds of the field. For all shots, 49.6% of the possessions that led to a shot began in the offensive third of the field, 34.9% began in the middle third and 15.8% began in the defensive third of the field. There were minor differences between men and women with the women having more possessions begin in the offensive third (54.1% vs. 46%) and fewer from the defensive third (11.6% vs. 18.9%).

If we focus just on goals, we see that possessions that resulted in a goal for men were equally split between the offensive and middle third of the field (41.2% each). However, in the women's game, an overwhelming majority of goal began in the offensive third of the field (73.7%) with 16% from the middle third of the field.

How the possession began

            We recorded how a team obtained possession of the ball. For men, the most common method of obtaining possession was controlling a free ball (29%) followed by a restart (19%), tackle (11%), throw-in (11%) and an interception (10%). For women, the most common method of obtaining possession was controlling a free ball (38%), restart (17%), tackle (10%), throw-in (8%) and an interception (3.5%).  If we focus on possessions that lead to a goal, the percentages are 38%, 18%, 18%, 0% and 18% respectively. For women, the percentages were 62%, 16%, 5%, 0%, and 5% respectively.

Number of players and number of passes in shooting possessions.

            Figure 2 shows the number of players involved in each non-penalty kick shooting possession for men and women. Figure 3 illustrates the number of passes in each shooting possession for men and women. From these two figures, it can be seen that over 60% of the possessions that lead to shooting opportunities for women involved only 2 players and 1 pass and that 85% of shooting possessions were combinations between 4 players using 3 passes. For the men, just over 2/3 of the shooting possessions were among 3 players and 2 passes and over 80% of the shooting possessions involved 4 players and 3 passes.

fig2

Number of players per shooting possession by sex

fig3

No of passes per shooting possession by sex.

Probability of scoring

            We divided the offensive end of the field into 5 areas. A diagonal line was drawn from the corner of the goal box through the adjacent corner of the penalty area (figure 4). This gave us 5 areas for shooting: in the goal box, in the penalty area in front of the goal, outside the penalty area, and finally the far right and left shooting angles.  Figure 4 shows the probability of scoring for men and women. The probability of scoring from the sharp angles is very poor as only 1 goal was scored during a women's cup game. The probability of scoring inside the goal box was about double for women than the men. In the penalty area, the probability of scoring was about the same for men and women. For long range shots (men and women both took about 25-30% more shots from outside the penalty area as inside the area), the women were just over twice as likely to score as the men. The better scoring probabilities by women for shots in the goal box and outside the penalty area is likely more of a statement about the difference in goalkeeping than it is about the accuracy of shooting.

  fig 4

Probability of scoring by sex

So what do we learn from such information?

            There are some interesting findings from this project, both in confirmation of prior work as well as similarities and differences between the men's and the women's game. For example, our work is in agreement with prior work showing that the vast majority of shooting possessions are very brief and involve few players and passes. Once possession had been obtained and the opportunity presented itself, teams were focused on the possibility of a shot. It would appear that 2 decisions had to be made once possession had been obtained: what should the first player do with the ball and what run should be made by a teammate. Square and back passing (we had very few of these kinds of passes lead to a shot) will slow down the play while a run and a pass that both penetrate the defense are most likely to lead to shots. This seems to be common to the game independent of gender. Consider McBride's goal in the September 3^rd USA:Guatemala Cup qualifier. A defender 2 touches the ball from just outside the penalty area to Kirovski in the center circle who controls and passes to Jones down the right flank. Jones settles, looks and finds McBride on the far post for a 1 touch shot. Four players, 3 passes (2 of which were penetrating passes), 1 touch shot covering 80+ yards under control all the way. Ten seconds from possession to goal would be generous.

            Differences in the men's and women's game can be seen in a few categories. For example, a large fraction of the goals scored in the men's game occur late in the second half while goals in the women's game are a little more evenly spread through the game. The men also scored 35% in the beginning of the second half. Maybe this is a result of teams being prepared to play immediately in the second half instead of using the early minutes to warm-up for the remainder of the 2nd half (Bangsbo, 1994). On the other hand, this might represent a fundamental difference in philosophy between the men's and the women's game where the men try to probe for weaknesses and the women attack from the start of the match. Goals from free play, number of players and passes were similar, but scoring possessions began in the offensive third for women more than men (74% vs 41%). This might be a combination of more aggressive play by female attackers and lesser skilled female defenders. Most of these possessions by the women began with a player controlling a free ball, possibly more data to implicate poor skill by female defenders. The current female defender is likely not as skilled or comfortable with the ball under pressure as are the men leading to poor pass selection and execution. In addition, parity in the men's game is much greater than the women's game. The women's game is still emerging on the world stage and stronger countries can pressure the defenders of their weaker opponents and carry the attack to the goal more than the more evenly matched men's game. A final difference is the probability of scoring. The men had a low probability of scoring in the goal box and outside of the penalty area vs. the women with the women having about 2x the probability of scoring from these two areas. It is likely that, at this stage of the developing women's game that the female goalkeeper is not as commanding of the immediate goal area and not as skilled at defending long range shots.

            A concern must be voiced about data like this. Much has been discussed about the concept of direct play as discussed by Hughes (1990). These data might be considered as supportive of direct play; high pressure defense by attackers and quick counter attacking passes through or over the defense. Some teams have the player skill and mentality to play this way (Norway men and women, Irish men for example). However, direct play might be better viewed as one aspect of the overall attacking pattern where possession soccer is used to force the defense to run and chase leading to fatigue that will impair the defense's ability to cover a direct attack when the offense finds the ideal opportunity.

How do we apply this information to training?

            Training implications are all about counter attacking tactics - quick runs off the ball, quick forward movement and numbers of supporting players from behind the point of attack - things that the experienced coach knows and practices. How the coach uses data like this and other data cited in designing training programs may require some creativity in practice organization. For example, shooting drills probably should focus on 1-touch shooting. Dribbling around inside the penalty area prior to shooting is pretty rare.

The majority of shooting possessions involves 4 or fewer players and 3 or fewer passes. If we play 4 v 4, we really do get all aspects of the game in a small-sided setting. One might even say that soccer is not an 11 v 11 game, but many 4 v 4 (or less) games all over the field. 4 v 4 lets you get various ways to obtain possessions, choices for support and off the ball running, penetration, and the opportunity to view the larger field to make tactical decisions. Because of this, 4 v 4 just might be the most ideal training exercise when game-like conditions is the focus of training.

One way to insert a more competitive session is to keep score, record which team (players) wins and post results. Rowing has an exercise called a seat race. They put 4 rowers in each of two boats and race them. At the end of the course, they swap one rower and race again. Each rower in a winning boat gets a "W" and the losers get an "L" (sort of like the plus/minus system in ice hockey). Seat races are then repeated during other training sessions by changing the mix of rowers in each boat. The best rowers should show up in the winning boats most often. In our 4 v 4 games, the player who "sees" and executes successful attacks should show up on the winning teams most often. This will tell you who has, and is successful at, this attacking vision. Add competitiveness for the keepers by recording their results separately. Play such games maybe twice a week (if competitive, it should be pretty high intensity so it will be a good fitness session). The length of each game should be short (e.g. 10 minutes), record the results, change a player or change opponents and repeat. Only about 1/3 of the active training session need be devoted to games like this.

Let the players know that this game attempts to duplicate where and how goals come from. Stress defense by attackers. While the attacker may not control the ball by tackling, they may well force errant passes that end up as free balls and can be easily controlled and now the forward must quickly change their mindset from defending to counterattacking. Increasing the number of players decreases this ebb and flow of the game leaving the player with fewer opportunities to make important decisions. You can play 4 v 4 on a penalty area size field, or increase the size (and add some restrictions to force more running) for fitness or decrease the size of the field to force quick decision making. Technique, tactics and fitness will all be worked very efficiently in such settings.

From a positional standpoint, female defenders probably need to become for comfortable with the ball when under pressure (which they won't get as often in larger sided games). The early portion of a match is critical for defenders as the forwards of better teams will apply high pressure defense in hopes of obtaining the ball near the goal and then strike at goal quickly. Goalkeepers need to become more commanding of their goal area and improve their defending of longer range shooting.

            In conclusion, in the modern game of soccer, the counter attack is one of the key tactical elements, but not likely to be the sole tactical approach to the game. There is a place for possession soccer as well as direct play. Data such as that presented above helps define the nature of quick attacks and give clues to specific tactical issues that the competitive soccer player and team needs to carry out to be successful.

 Acknowledgements

W. Wesley Dowd from the K Lab at Duke University Medical Center helped me collect the data and Tony DiCicco offered insights into the application of the results. This project was funded by a grant from Nike, Inc.

Bangsbo J. Fitness Training in Football-A Scientific Approach. HO+Storm. Bagsvaerd, Denmark. 1994.

Bate R. Football Chance: Tactics and Strategy. In Science and Football (T Reilly, A Lees, K Davids, WJ Murphy eds.). E&FN SPON. London, 1988. P293-301.

Garganta J, Maia J, Basto F. Analysis of goal scoring patterns in European top level soccer teams. In Science and Football III (T Reilly, J Bangsbo, M Hughes eds.). E&FN SPON. London, 1997. P246-250.

Hughes C. The Winning Formula. Collins. London. 1990.

Lanham N. Figures do not cease to exist because they are not counted. In Science and Football II (T Reilly, J Clarys, A Stibbe eds.). E&FN SPON. London, 1993. P180-185.

Partridge D, Mosher RE, Franks IM. A computer assisted analysis of technical performance - A comparison of the 1990 World Cup and intercollegiate soccer. In Science and Football II (T Reilly, J Clarys, A Stibbe eds.). E&FN SPON. London, 1993. P221-231.

Pollard R, Reep C, Hartley S. The quantitative comparison of playing styles in soccer. In Science and Football (T Reilly, A Lees, K Davids, WJ Murphy eds.). E&FN SPON. London, 1988. P309-315.

Reep C, Benjamin B. Skill and chance in Association Football. J Royal Stat Soc, Series A 131:581-585, 1968.

Table 1- Games recorded and analyzed

¹Pool play game to make up for incomplete quarterfinal taping of Nigeria:Brazil (see text)

²Video quality too poor for analysis. These games from the round of 16 were randomly chosen (see text)