Our View of the Game - The Underlying Principle


Netball is a sport with a unique structure. Alternating centre passes after each goal rather than possession going to the non-scoring team means both teams start the game with equal opportunity to score ahead of them – or equal ‘expected goals’. The game is in balance, but as it plays out somescoring opportunities will be realised while others will be lost, creating an imbalance in favour of one team.


In our definitions, possession for a team falls into one of two categories – a Centre Pass or a Gain.

A Centre Pass starts either at the beginning of a quarter or after a goal has been scored and finishes either with a goal being scored or a loss of possession. A ‘Gain’ is any gain in possession off the opposition, regardless of whose centre pass it is, and occurs during play. A Gain can be made through a defensive rebound, an intercept or tip picked up or an error by the opposition. We consider that once a Centre Pass possession is lost, then that Centre Pass is determined as unsuccessful, regardless of whether possession is regained and scored by that team – in this instance the new possession would be counted as a Gain. A Gain can therefore be made off an opposition Centre Pass or off an opposition Gain, and there could be multiple Gains by both teams in the course of one goal being scored. When using this definition, we can then see that the number of Gains each team makes relative to the other becomes the determining factor in winning the game. Given that both teams start with equal Centre Passes, the team that can make more Gains than the other team creates more scoring opportunities and so cannot lose.

Note that we say ‘cannot lose’ rather than ‘will win’ – this is because there are instances where teams can be in possession at the end of the quarter but not have time to score. In these instances a scoring opportunity is lost, but there is no corresponding Gain for the opposition team. There is also the possibility that one team will have an extra Centre Pass due to the timing of the
final whistle. These cases can result in one team having more Gains than the other but ending in a draw rather than a win, but note the team with more Gains cannot lose.

We do have a mathematical proof for this concept, but in simple terms, if a team can make more Gains than the opposition team, they cannot lose the game. It doesn’t matter how many gains they make as long as it’s more than the opposition, which brings the focus not only to ways to gain possession, but also ways to keep possession, or deny opposition gains. This way we view the game leads to an analysis framework built around understanding the effectiveness of a team’s ability to maximise gains while minimising losses. At an individual level we want to understand and measure the effectiveness of a player’s ability to carry out their positional role while contributing to this maximal gain/minimal loss target, also taking into account differing team structures and game plans.

Why we see things the way we do

When we collect our data we end up with a huge amount of information on teams and individual players, but the way we report that data can have a significant effect on the interpretation of the performance of those teams and players.

Raw Counts vs Proportions

When considering individual players and how to measure their performance, the common method is to simply provide a raw count of relevant events for that player. For example, the number of feeds a player makes, or the number of Centre Passes received. While this is a valid method for some measures, in others – especially those relating to work rate variables – just providing the raw count doesn’t consider the team structure or the relative strengths and success of the teams involved in the game.

To use feeding as an example, two players on opposing teams may finish the game with the same number of successful feeds made. Using just the raw count to evaluate this area of performance would suggest they performed at a similar level. However if we consider the total number of successful feeds by each team, we may find that Team A had significantly more possession and were able to feed the ball into their shooters much more often than Team B. In this situation if we look at the individual player counts as proportions of the team totals, the player on Team A would actually have carried out a much lower feeding workload within their team than the player in Team B. For this reason, we report several our individual player measures as proportions of team totals, to give a true indication of a player’s contribution to their team’s performance.

Raw Counts vs Success Rates

Similar to the issue above, in some measures reporting raw measures can skew perception of a player or team’s effectiveness in a particular area, because no indication is given on how many opportunities or attempts they actually had. Using the feeding example from above where our two players finished the game with the same number of successful feeds, just reporting the raw count may hide the situation where one player has made significantly more unsuccessful feeding attempts than the other.

Shooting statistics have always been reported as the number of successful goals along with the total number of attempts – we need to take this approach with other areas of the game where accuracy is important as well.

Rebounding is another classic example – is the player in a competition who secures the ‘most’ rebounds the best, or is it simply because they and their shooting partner miss significantly more goals than any other shooting combination and so have more opportunity to get rebounds? Would it not make more sense to consider the number of rebounds secured as a percentage of the total number of rebound opportunities (for both team and individuals)?

Making Comparisons

There are two issues with making straight comparisons between individual players – they may be playing different positions, and they may have had differing amounts of court time. Netball is a structured game – not only do different positions have different roles, they are also limited by where they can go on the court and where the ball can go. Unlike most other team sports where although there are positional roles, any position can perform most actions (scoring, rebounding, defending etc),
in netball playing positions strictly define what a player can and can’t do. A GK cannot feed, a GS cannot receive the Centre Pass, a WA cannot shoot, the list goes on. As a result, instead of throwing all player measures into the same data table, we report our data by position, so accurate comparisons can be made. If a player has played more than one position, they will have separate sets of measures for each position, to reflect the different roles they would be expected to fulfil on the court.

To deal with the issue of differing court time, we adjust all our measures to make sure they are given on the same scale.
Proportions and accuracy percentages are all fine to be compared as they have no unit, but all straight counts (e.g. penalties, gains, tips etc) are adjusted to be on a per-quarter basis, so that we can accurately compare Player A who has played 20 quarters over the season with Player B who has had twice as much time on the court. By doing this we are giving a comparison of a player’s effectiveness in the time they were on the court, rather than giving weight to player results simply because they have had more opportunity.

We also have a system to extrapolate out results for players who have only played partial quarters, or who have changed positions during a quarter (as our results are separated by position), however we do advise caution when considering this partial data, or data for players with very limited court time.