I was lucky enough to be given an article that talked about how Arsene Wenger, the manager of Arsenal in the English Premier League was a huge proponent of using statistics to measure player performance. A couple of years ago, before I was aware of what Arsene was doing and before I had read the article, I implemented a statistics based playing system for our boys varsity soccer team at Tigard High School in Tigard, OR. At Tigard, we didn’t have a team with a lot of size or speed in comparison to our competition, so it was important that we were able to compete in other ways. We did have players with decent skills and abilities at soccer, so we had to win games based on our skill on the field, our ability to hustle, and our fitness. One of the things we decided to implement was a system for tracking players statistically. We felt it would not only help us decide who should be on the field during a game based on facts rather than based on the gut feel for how a player was playing during a game. It also allowed us to share the statistics with the players so that we could have fact based discussions about individual play. This was especially useful for discussions about playing time.
This blog entry allows me to share the spreadsheet along with the calculations we used to measure players and why we used those calculations. There are obviously a lot of improvements that could be made to what is tracked, and there are other ways to measure and interpret the data. Remember, that our statistics were used based on the fact that we were a smaller, yet skilled team. Teams with different “player profiles” would probably choose to implement the statistics differently.
Attached is a copy of the spreadsheet. Please feel free to download it and adjust as you need. Also, feel free to contact me directly if you have questions about the calculations and why we tracked things the way we did.
There are several tabs in the spreadsheet as described below.
- Player Stats – BV – 09-04 – This tab is the most critical. Every game, we had a printout of this tab so that we could track all of the players in one place. It allowed to do in-game comparisons if we were trying to decide what players to play in what situations. It should be noted that I was the assistant coach, and my sole focus during the game was keeping statistics. There is no way the head coach could do this, and then also manage the game and team.
- Behavior – We tracked all red and yellow cards for all players. In high school, this was important because players could easily get a yellow or red card for something like foul language in addition to things like hard challenges. Also, in high school soccer in Oregon, when you get a yellow card, you are taken off the field until the next dead ball. We wanted to be able to track who was costing their team possession and loss of advantage by their behavior on the field
- Goalkeeper Stats – We kept goalkeeper statistics separate from field player statistics as the requirements were different.
- Individual Player tabs – For each player, we kept historical statistics for every game. This allowed us to look for players that were improving in the statistical categories we were measuring.
- Player Stats – Total – We kept a running total of player statistics across all of the games in the season. Additionally, we did some breakdown of the statistics by position so that we could see if we were strong or weak on any position on the field.
- Player Comparison – We developed an overall rating for each player for each game. Then, we tracked the overall rating for all players for each game on one spreadsheet and charted the results. It allowed us to visually be able to compare the players.
Each section of the document below provides details about each tab of the spreadsheet and how we used it.
Player Stats – BV – 09-04
Notes (bottom) – We had some general categories where we just wanted to keep track of what we were doing,
- Offsides – Trying to track how frequently we were offsides and who was doing it. This could indicate a timing problem between the forwards and players making passes to them. A limited number of them may be ok if you’re playing an attacking style. Also, a game where there are none may represent that you’re not attack minded enough.
- Gave up dangerous free kick – This may indicate players that are being careless in where they commit fouls if a player is accumulating a lot of them
- Corner Kicks that went through – This is an offensive stat. If we take corner kicks where no one is getting to the ball, it’s a problem.
- Total corner kicks – Tracking how many corner kicks we are taking.
- Successful cross – This is an indicator where we are trying to see if the ball is being played to a player of just into space and no one is getting to it. It may be an indicator for players making poor crosses, or an indicator for players not getting to balls that are crossed in. A successful cross does not have to end up as a goal. We are more focused on players getting to the ball.
- Cross – This is any cross at all. This is used to the last stat to indicate how often a cross is made that is good. There should be minimal difference between the number of crosses and the number of successful crosses.
- Quality shots – In addition to tracking shots, we also want to track quality shots, which we counted as a shot with power behind it at the net. In high school soccer, there are a lot of power shots that go far too high or wide of the net. Also, if a shot was weak or easily saved by the keeper, we did not count it as a quality shot.
- Missed PK – We are trying to keep track of how many and who to indicate if we are taking advantage of our chances, and who our best PK takers are.
- Defensive clear – This is an indicator of what we do with the ball when under pressure in our own end. This could indicate a player that always clears the ball without thinking, or midfielders that are not getting open to receive a pass from the defense.
- Sissy points – This was a comical topic that we tracked, which works especially well with boys. Any time a player ducked away from a header or didn’t go into a challenge hard, we gave them a point. The person with the most points at the end of a week of games had to wear a pink pinnie the entire next week of practice. We wanted to encourage players to always play hard and tough.
- Player – simply list each player on your team
- Shots taken – this includes any shot no matter what the quality is
- Shots missed – this includes shots saved by the keeper (or other player) along with shots that went to the side or over the goal
- Goals – Any goal by a player
- Assists – Any assists on a goal
- Def Stops – We used this to track any player (not just a defender) that challenged an opposing player and won the ball
- Got Beat on the Dribble – Any player that challenged an opposing player and got beat.
- Beat Player on the Dribble – Any player dribbling that got past an opposing player that closed down on them
- Lost the Ball on the Dribble – Any player dribbling that lost the ball to an opposing player when the opposing player closed down on them
- Recovered to win the ball back after getting beat – This was actually used to track times a player lost the ball on the dribble or got beat by an opposing player on the dribble, and hustled to win the ball back. We were looking for how players reacted to when they got beat.
- Gave up on a Play – Any player that lost the ball and stopped playing. We were promoting a hustle style of play, and we did not want players giving up when they made mistakes.
- Part of a Scoring Combination – Each player involved in a scoring play would get a tick mark, including players that may not be credited with an assist. We wanted to track if there was a pattern to how sets of players moved the ball to score.
- Successful Pass to Feet – This was critical for us. We were a small team and didn’t win the ball in the air much. So, we wanted to play a controlled passing game. It was critical for players to make passes that put their teammates in a good position to make a play.
- Unsuccessful Pass to Feet – Any pass that could have been made to a teammates feet, but missed the target, or was made in a way to put their teammate into a difficult situation.
- Long ball that found the target – Given our size, our long balls had to be accurate an on target. Generally, anything high in the air, we were going to lose. However, there were long balls that we could run onto or receive.
- Long ball without purpose – This is different than the Defensive clear category. We were looking to track players playing long balls that did not make it. At the high school age, players are often most interested in how long they can send a ball. A high number may indicate a player not thinking about what they are doing, or maybe a player with lack of confidence in their skill level.
- Win the ball in the air – Even though we were a smaller team, we wanted to win as many balls in the air as we could.
- Lost the ball in the air – Again, even though we were a smaller team, we wanted to lose as few balls in the air as we could.
- Opposition uncontested in the air – This was an important stat because even though we may not win many balls in the air, we wanted to challenge for as many as we could. If we waited for the other team to win the ball and play it, we would always be on the defensive.
Percentages and Totals (column T through Z) – These are columns that we used to compile the detailed statistics into specific categories. You can change them to have different definitions, but this is how we tracked things.
- Goal Percentage (=IF(B2=0,0,D2/B2))– A very obvious statistic. Number of goals divided by number of shots. The higher the percentage, the more likely a player would score when they shot. For our team, this was a difficult statistic because we didn’t shoot enough for this number to give us a lot of information.
- Def % (=IF(F2+G2=0,0,F2/(F2+G2))) – We added the number of times a player got beat with the number of times they stopped the opposing player. Then, we divided the number of defensive stops by that number. We wanted to get the % of time a player was successful at stopping the opposition.
- 1v1% (=IF(H2+I2=0,0,H2/(H2+I2))) – We added the number of times a player beat a player on the dribble with the number of times they lost the ball to the opposing player. Then, we divided the number of times they beat a player on the dribble by that number. We wanted to get the % of time a player was successful at beating the opposition.
- Hustle Points (J2-K2) – We simplified this to basically say we wanted players to hustle back to win a ball rather than give up. The number should always be positive if we’re hustling. This could be defined in a lot of different ways.
- Pass % (=IF(M2+N2=0,0,M2/(M2+N2))) – We added the number of times that a player made a successful pass with the number of times they made an unsuccessful pass. Then, we took the number of successful passes and divided it by that total. We wanted to gauge how frequently a player was going to make a good pass.
- Long Ball % (=IF(O2+P2=0,0,O2/(O2+P2)) - We added the number of times that a player hit a long ball that hit the target with the number of times they hit a long ball that didn’t make the target. Then, we took the number of long balls that hit the target and divided it by that total. We wanted to gauge how frequently a player was playing a long ball that made it to it’s destination.
- Air % (=IF(Q2+R2=0,0,Q2/(Q2+R2)))- We added the number of times that a player won a ball in the air with the number of times they lost a ball in the air. Then, we took the number of times they won a ball in the air and divided it by that total. We wanted to gauge how frequently a player was winning the ball in the air.
Behavior
We had a couple of measures for behavior to track the overall discipline of a player and the team.
- Bad language – any time a player used bad language even if it didn’t involve the referee giving them a card
- Talked back to coaches – any time a player had a negative reaction to a coach. Discussion was ok, but yelling at the coach was not tolerated
- Talked back to a referee – same as talking back to the coaches, but more during the game. We wanted to see if a player was letting the game affect them mentally by how it was being called
- Red cards – any red card given
- Deserved red cards – we tracked when players got a red card and whether we thought it was warranted.
- Yellow cards – any yellow card given
- Deserved yellow cards – we tracked when players got a yellow card and whether we thought it was warranted.
Goalkeeper Stats
The goalkeeper stats that we kept were a bit less involved than the player stats, In addition to basic stats, we were also trying to keep stats on our judgement of how our goalkeeper reacted in certain situations.
- Date – Date of the game
- GK – name of the goalkeeper
- Shots – The number of shots (on target) that the opposing team took
- Save – The number of saves the goalkeeper made
- Goals – The number of goals the goalkeeper gave up
- Win – Whether the game was won or lost
- Loss – Whether the game was lost
- Shutout – Was the game a shutout
- Goals per game – running average of the number of goals per game the goalkeeper allowed
- 1v1 attempts – the number of times the goalkeeper faced a 1v1 situation
- 1v1 attempts disrupted – The number of 1v1 attempts that the goalkeeper came off his line and won
- 1v1 attempts that got beat – The number of 1v1 attempts that the goalkeeper came off his line and got beat
- Should have played the ball in the air – The number of times the goalkeeper held his line when he should have played the ball in the air
- Should not have played the ball in the air – The number of times the goalkeeper came off his line to play the ball in the air, but should not have.
- Came off the line too soon – The number of times the goalkeeper came off his line when a defender was in a position to make a play
- Held the line too long – The number of times the goalkeeper held his line when he should have come off it.
- Average goal kick distance – the average distance of the goal kicks the goalkeeper has taken
- Average punt distance – The average distance of punts a goalkeeper has taken in a game
- Successful clears – the number of times the ball is dropped from a defender to the goalkeeper and they clear it out of danger
- Unsuccessful clears – the number of times the ball is dropped from a defender to the goalkeeper, and the clear still puts them in danger
- PK’s – number of PK’s the goalkeeper faced
- PK’s saved – number of times the goalkeeper saved a PK
- Bobbles – Number of times the goalkeeper bobbled the ball when playing it with his hands
- Crosses caught – The number of times the ball is crossed in front of the net, and the goalkeeper catches the ball cleanly.
Individual Player Tabs
The individual player tabs are the same as the Player Stats tab, but represent the history of the player statistics for each game. So, each column on the player tab is the same as the column on the player statistics tab. We did create a unique player scoring system that is underneath the game statistics. We had the following ratings that we used for each player
- Shooting – Simple formula taking number of goals minus the number of shots. We were trying to emphasize shooting efficiency, not just shooting for the sake of shooting.
- Def Stops – We simply wanted to keep track of the number of times a player stopped and opposing player. Since we handled percentages on the main spreadsheet, we wanted to reward players that were making a lot of stops in our points system
- 1v1 – This was simply the number of times a player beat an opposing player on the dribble versus how often they got beat on the dribble.
- Hustle – This was the number of times a player recovered to win the ball back after they had lost the ball. If they were losing the ball, but able to win it back, we were ok with that.
- Pass Eff - The number of times they made a successful pass to feet minus an unsuccessful pass to feet
- Long Ball Eff – The number of times a player played a long ball that found the target minus the number of times the long ball did not make it to the target.
- Air Eff – The number of times a player won the ball in the air minus the number of times a player lost the ball in the air
- Overall Rating – The overall rating for the game was simply adding up the other categories.
Player Stats - Total
This tab added up the totals for the player across all the games. This could also be called the running totals for the season. In addition to the players, we also gave them a general position that they played. This was a bit tough because we had some players in multiple positions. So, we could have done a better job of tracking stats by position. Also, the section below the stats for each player, there was a breakout by position. We were interested in not only tracking the stats by player, but by positions.
Player Comparison
This tab added up the player ratings (ratings at the bottom of the individual player tabs as described above) across the different games. It allowed us to chart how players were performing from game to game, and also compare each player’s performance over time. We were looking for smooth lines in the player performance to show consistency. In addition to that, we wanted to be able to see consistent lines from player to player.
Conclusion
No statistical measurement system is perfect, especially for a game like soccer where every situation is different, and the game is constantly moving. There are no breaks between plays to make adjustments, etc. Players are playing the game “on the fly”, so the goal of your statistics should be to help players understand what their tendencies are, and help them make adjustments from game to game.