Analysis

Changes to the Scoring System

The change to the scoring system is one rule enhancement that has the potential to reshape how the game is played. The most obvious comparison to make is with Basketball and their introduction of the 3 point line. The 3 point-line was thought to be a gimmick when first introduced into Basketball and was eventually adopted in the NBA in 1979, but it didn’t permeate all levels of the game until 1986. 

Perhaps more interestingly it wasn’t really until the sports analytics evolution beginning around the start of the millennium that the true value of the 3 pointer was understood, and it still took many years before that knowledge was common practice among the majority of teams. 

This is best illustrated by the amazing work of Kirk Goldsberry. He has chartered the 200 most common shot locations and in the chart below demonstrates how an entire genre of shots has disappeared from the game – The Mid Range Shots! 

This chart is the 200 most frequently used shot zones and doesn’t mean to imply no shots happen in the mid-range but it is telling that the analysis says having a slightly worse efficiency for a 3 pointer and the ‘easy’ close shots are the ones to take – and the mid-range shots are simply not worth the effort. 

Before the introduction of this year’s rule enhancement the GAA analysts and managers  have also had their say on shot efficiency. Distance is the number one factor in predicting shot accuracy. 

In the chart below you can see that as you move further from goal the likelihood of scoring decreases [the anomaly is the shot at goal inside 13m – but after that it follows a very logical downward trend]. 

Just like in any evolution, some people can be ahead of the curve and I think Dublin is a great example of this. They have known for a long time that there comes a certain point where taking a pop at goal just simply isn’t worth it. 

Take a look at the chart below; it shows the % of a team’s shots (from open play) that are taken at 40m or further. It compares Dublin to every other county since 2011. Interesting to see the massive drop from 2011 to 2012 when Jim Gavin took charge.

In the vast majority of seasons teams are taking between 20-25% of shots from 40m or further out, while Dublin never get above 20% and even had as low as 4% in 2022.* 

The reason the mid-range shots have ‘disappeared’ in basketball is simple maths. Given the shooting efficiency of those shots, it’s more advantageous to step back outside the arc and attempt a more difficult 3-pointer. The increase in scoring return is worth the decrease in shot efficiency. 

The big question is will we see the same phenomenon in Gaelic Football and where is the mid-range likely to be? For a team like Dublin attempting shots from greater than 40m has never been worth the ‘risk’ so they try to work the ball closer to goal. But now it is worth something extra, will that change behaviour and at what point is it going to be worth retreating backwards to take the lower % shot?

Returning to the very simplistic line chart above we can pick a point along that line and see the shot efficiency from that point. Let’s take a concrete example. At about 45m from goal,  players score an average of 29% of the time. This shot will now be worth 2 points so the scoring return from this shot will be .58 or 5.8 points for every 10 shots taken. 

At what distance does a team need to be inside the scoring arc to get a similar return? It’s pretty much 25m from the goal. At that point the shot efficiency is .57 or 5.7 points for every 10 shots taken. 30m shots are scored 50% of the time meaning 5 scores per 10 shots. In theory 10 shots from 40-45m are more valuable than 10 shots from the closer range of 30-35m. 

All of this data is obviously calculated before the scoring arc existed and teams and players may change strategies to alter these numbers but this is the best knowledge we have today of current behaviour. Given this, it’s plausible to conclude at this stage that purely from a numbers perspective shots taken between 25m and 40m are going to be the GAA’s equivalent of the mid-range.  They are of less value than the risker and less efficient shots from further out. 

In Gaelic football, the limited number of shots per game—just 50 on average—means that the true shooting abilities of players and teams may not fully reveal themselves within such a small sample size. This stands in stark contrast to basketball, where each team takes nearly 100 shots per game across an 82-game season, providing a far larger dataset to analyze and allowing shooting percentages to stabilize over time.

The results therefore are heavily influenced by individual performances or single events, which makes the numbers in Gaelic football feel a bit unpredictable at times. It’s a reminder that with fewer chances, one great shot—or one bad miss—can change an entire season.

*Before you ask I did check all other teams records and none come even close to Dublin across this time-frame. Kerry are probably the closest and in 2022 only took 3% of shots analysed from 40m or further, but that is an outlier. It seems while most teams and managers might logically conclude that distance is the biggest indicator of shot success, Dublin are the team who most often put it into practice.

Comparing players from different era’s is always a thankless task. The game changes, the demands required and the externalities make it a fools errand. But how about comparing free takers over the last decade or so. Sure the game has changed even in that short time frame but free taking, as a closed skill, has remained untouched by rule changes and could make for an interesting analysis of player improvement.

A quick note on the data. It’s all collected by my company Gaelic Stats. We have collected data on senior intercounty games since 2011. For each shot, the player, the foot used, the location (XY) of the shot on the pitch as well as a few other characteristics are collected. The following charts are all for football, championship games, from the seasons 2011-2023. I have only included frees (no sidelines/45’s or penalties). It covers over 6,000 shots in total.

The chart below shows the season on season shot accuracy % of frees at senior level.

Shot Success has (dramatically) improved

In 2011 the shot success (point or goal) of a free was 68% by 2023 that figure is at a whooping 82%. There is some deviation year to year but as you can see from the trend-line this has been largely increasing year over year.

How difficult was the shot?

The above graph is a simple average, all shots grouped together and takes no account of how difficult the shot was. If Player A attempts a free from 50 meters out v’s 13 meters we are hardly comparing like with like. There are more sophisticated ways to measure shot difficulty (think expected goals), but for the purposes of this analysis let’s keep it simple. Distance is the biggest driver of shot success. The further out, the harder the shot. To illustrate this point take a look at the graph below.

The blue line tracks the average shot success rate at each 1m interval of distance. I’ve highlighted in red where the shot % is still in the 90%’s, after this it starts to drop and falls pretty quickly. So we can see that distance matters, probably not so much up until that 22-26m range, but beyond that it certainly does.

Better or Easier shots?

Given that distance plays an important role we can isolate that when comparing players across the previous 13 seasons. The chart below shows the average distance of a shot since 2011.

In 2011 the average distance of a free was 36.3, in 2023 players are on average are shooting from 31.9 meters (over 4 meters closer to goal on average). This is further visible when we look at the shot maps. Take the example below (for simplicity) I’m comparing 2011 to 2023.

Although this is more of an eye test, the spread of shots is much wider in 2011 than in 2023. You see a lot more shots from near the sidelines and out around the 45. From these 2 charts and the average shot distance chart above it looks like players in 2023 avail of ‘easier’ chances on average than they did in previous years.

Controlling for Distance

The final piece is to compare era’s and shots at a simalar distance. To achieve this I’ve grouped the distances into buckets; 10-19, 20-29 etc…. I also want to compare co-horts. So I’ve split the sample into two groups. Shots in the seasons 2011-2017 (7 seasons) and then 2018-2023 (6 seasons). There is no exact science to this but given the relatively small sample size and variance within a season I went with this method.

The chart below shows for each given bucket how much better (more accurate) players are in 2018-2023 compared to 2011-2017.

It seems clear to me that two things are happening simultaneously to drive the increase in free taking from 68% in 2011 to 82% in 2023.

Firstly and probably more importantly teams are getting smarter about which frees are worth taking on, the average distance of a free is falling, teams likely (or at least should) know the range of their free takes and have adjusted their strategy to take on the correct opportunities.

Secondly, when controlling for distance we see a slight increase in performance. At the 10-19m not so much, but to be honest there is little or no room for improvement here. But as we get into the 30-39m range we see a 3% increase in accuracy. This is not massive, and I was probably expecting this to be bigger, but in a game of 1%’s maybe a 3% increase is more than a good enough in a relatively short time frame.

Are free takers better now than in 2011? Probably yes on both counts really, better at choosing the right opportunities and better at taking them when they come along.

PS – Because I know I will be asked. At the 20-40m range, free takers who kick out of their hands tend to be more accurate and only when you get past 50m does kicking from the ground seem to carry an edge.