I wrote previously about some of the top attacking players in League 2 from last season (here). In this post, I wanted to look into this in some more detail. Last time, we focused mainly on NPG/90 (non-penalty goals per 90 minutes played) and shot volume. However, this time, I want to focus more on the location of those shots.
Clearly, a shot from 2 yards out right in front of the goal is far more likely to result in a goal than a speculative effort from 30 yards. However, in the basic shot volume analysis last time, both of these attempts would have been counted equally. In League 2, precise shot location data is simply not available, or not publicly at least, so we have to work with the more basic data that we have. This data comes from the BBC Sport live text commentary, who appear to get their data from Opta.
In the live text commentary, there is a description of each shot, including the outcome, whether it was with the left foot, right foot or header, plus a shot location description. These descriptions essentially narrow a shot down to six locations - 'Very Close Range', 'Centre of Box', 'Left/Right Side of 6-Yard Box', 'Left/Right Side of Box', 'Difficult Angle' and 'Outside Box'.
These descriptions are far from perfect, but hopefully they are consistent across the season and we can work with them to an extent.
Over the course of the 2014/15 season, there were 10,789 shots in total in League 2 leading to 1,174 goals (excluding penalties and free-kicks). In other words, 10.9% of shots in League 2 last season resulted in a goal. However, obviously not all shots are equal. For example, header attempts from 'Very Close Range' were converted at a rate of 43.0%, while efforts with either foot from outside the penalty area had a conversion rate of just 3.5%. Therefore, we would like strikers to be taking the majority of their attempts from 'Very Close Range' with an overall conversion rate of 51.4%, 'Left/Right Side of 6-Yard Box' with a conversion rate of 23.6% or from 'Centre of Box' with a conversion rate of 16.1%.
We would rather that players not be regularly shooting from 'Left/Right Side of Box' with a 9.0% conversion rate, 'Difficult Angle' with just 4.8% conversion rate or 'Outside Box' with just 3.6% conversion rate.
Given the conversion rates for those six locations, we can construct a very basic expected goals metric for each player in the division, based on the number of shots taken from each zone, whether the shot was with the foot or the head, and the overall conversion rate for the relevant shots in those zones. Obviously, this is very basic, but the data is pretty basic and it can act as a starting point.
The table below shows the ten strikers from the final table in the previous article. However, it now has an extra column - expected goals.
We would expect that the top scorers in the division overachieve somewhat compared to their expected goals, but we can see that some have overachieved significantly more than others.
Last time, we concluded that Andy Mangan was one of the top strikers in the division, albeit with a slightly high conversion rate. However, if we look more closely into his shot location, we can see that he scored twice as many goals as we might have expected, driven significantly by scoring all eight goals from just seventeen shots inside the penalty area. This is a pretty high conversion rate, even given the shots are from good locations and it would seem unlikely that this conversion rate will continue over a larger sample of shots.
We noted last time that Marc Richards had an exceptionally high conversion rate and this is reflected in the expected goals of just 5.6. With just 47 shots, eleven of which were from outside the penalty area and none of which came from very close range, his tally of thirteen goals is really quite extraordinary. While the possibility stands that he is an excellent finisher - double figures in six of his last seven seasons at this level suggest that he is a good player (including penalties) - his goal tally in 2014/15 probably flattered slightly.
Let us now look at the top ten players in terms of expected goals from last season. The table below shows this data.
We see that Shrewsbury's James Collins had the highest expected goals of any player in League 2 last season and that his actual goals land almost spot on what we expected. Cult hero, Adebayo Akinfenwa, is second on this list, fractionally ahead of the veteran Jamie Cureton.
Looking in slightly more depth, we can see that Akinfenwa had no fewer than eight chances from 'Very Close Range', which is actually more than Plymouth Argyle created in the entire season as a team. However, seven of those eight chances were with the head, which reduces the expected goals slightly. In terms of the most chances from 'Very Close Range' with the foot, Matt Tubbs and Jake Hyde are tied with six chances. Jake Hyde actually scored five of those six chances, which means that his other three goals came from 72 attempts from other locations, which is really quite dreadful.
Oxford United's Danny Hylton is another underachiever here. While he finished the season with a respectable 14 goals, six of those came from the penalty spot, and had he lived up to his expected total of 11.2 NPG, he might have expected to get closer to 17 or 18 goals, which could have fired Oxford closer to a playoff spot.
As we did in the previous article, let us now build in the total minutes played to get an expected goals per 90 minutes table.
* minimum 1000 minutes played
We can see that Bury's Danny Rose is the top player in terms of expected goals per 90. It will be interesting to see whether he can make the step up to League 1 with Bury next season, but he was clearly one of the outstanding strikers in League 2 last year. Jake Hyde shows well here as well, but he will be hoping to start converting some of his other shots next season. If he can come close to repeating his expected goals tally next year, but convert those attempts at even just a league-average rate, York could have a very decent striker on their hands at this level. Ivan Toney is an interesting prospect. At just 19-years old, this was his first full season and the early signs are promising.
Looking at a couple of players that did not quite meet the 1,000 minute minimum requirement for the earlier table, Billy Kee had an ExpG/90 of 0.62 for Burton and Mansfield, but is currently available on a free transfer. While he only scored three times last season, he could be worth a punt for a League 2 team.
The other interesting name is Oxford United's James Roberts. Having only just turned 18, he made his first team debut this season and was predominantly used off the bench. Despite that, his ExpG/90 of 0.48 is very promising and averaging 2.6 shots per 90 puts him well up there in the division. He could certainly be one to watch next season in what the bookmakers are viewing as a potentially title-challenging team.
 |
| Danny Rose was one of the top strikers in League 2 last year |
Clearly, a shot from 2 yards out right in front of the goal is far more likely to result in a goal than a speculative effort from 30 yards. However, in the basic shot volume analysis last time, both of these attempts would have been counted equally. In League 2, precise shot location data is simply not available, or not publicly at least, so we have to work with the more basic data that we have. This data comes from the BBC Sport live text commentary, who appear to get their data from Opta.
In the live text commentary, there is a description of each shot, including the outcome, whether it was with the left foot, right foot or header, plus a shot location description. These descriptions essentially narrow a shot down to six locations - 'Very Close Range', 'Centre of Box', 'Left/Right Side of 6-Yard Box', 'Left/Right Side of Box', 'Difficult Angle' and 'Outside Box'.
These descriptions are far from perfect, but hopefully they are consistent across the season and we can work with them to an extent.
Over the course of the 2014/15 season, there were 10,789 shots in total in League 2 leading to 1,174 goals (excluding penalties and free-kicks). In other words, 10.9% of shots in League 2 last season resulted in a goal. However, obviously not all shots are equal. For example, header attempts from 'Very Close Range' were converted at a rate of 43.0%, while efforts with either foot from outside the penalty area had a conversion rate of just 3.5%. Therefore, we would like strikers to be taking the majority of their attempts from 'Very Close Range' with an overall conversion rate of 51.4%, 'Left/Right Side of 6-Yard Box' with a conversion rate of 23.6% or from 'Centre of Box' with a conversion rate of 16.1%.
We would rather that players not be regularly shooting from 'Left/Right Side of Box' with a 9.0% conversion rate, 'Difficult Angle' with just 4.8% conversion rate or 'Outside Box' with just 3.6% conversion rate.
Given the conversion rates for those six locations, we can construct a very basic expected goals metric for each player in the division, based on the number of shots taken from each zone, whether the shot was with the foot or the head, and the overall conversion rate for the relevant shots in those zones. Obviously, this is very basic, but the data is pretty basic and it can act as a starting point.
The table below shows the ten strikers from the final table in the previous article. However, it now has an extra column - expected goals.
| Player | Team | Actual NPG | NPG/90 | Shots/90 | Expected Goals |
|---|---|---|---|---|---|
| Andy Mangan | Shrewsbury Town | 8 | 0.71 | 2.7 | 3.9 |
| Danny Rose | Bury | 10 | 0.52 | 2.9 | 8.4 |
| Marc Richards | Northampton Town | 13 | 0.51 | 1.9 | 5.6 |
| Tom Nichols | Exeter City | 12 | 0.49 | 2.2 | 7.1 |
| Jamie Cureton | Dagenham and Redbridge | 19 | 0.48 | 2.4 | 14.0 |
| Daniel Nardiello | Bury | 9 | 0.45 | 2.2 | 6.0 |
| Matt Tubbs | AFC Wimbledon/Portsmouth | 19 | 0.44 | 1.9 | 13.1 |
| Ryan Bird | Cambridge United/Hartlepool United | 8 | 0.43 | 1.5 | 5.1 |
| Tyrone Barnett | Oxford United/Shrewsbury Town | 8 | 0.41 | 2.6 | 6.6 |
| Mark Cullen | Luton Town | 13 | 0.41 | 2.0 | 7.7 |
We would expect that the top scorers in the division overachieve somewhat compared to their expected goals, but we can see that some have overachieved significantly more than others.
Last time, we concluded that Andy Mangan was one of the top strikers in the division, albeit with a slightly high conversion rate. However, if we look more closely into his shot location, we can see that he scored twice as many goals as we might have expected, driven significantly by scoring all eight goals from just seventeen shots inside the penalty area. This is a pretty high conversion rate, even given the shots are from good locations and it would seem unlikely that this conversion rate will continue over a larger sample of shots.
We noted last time that Marc Richards had an exceptionally high conversion rate and this is reflected in the expected goals of just 5.6. With just 47 shots, eleven of which were from outside the penalty area and none of which came from very close range, his tally of thirteen goals is really quite extraordinary. While the possibility stands that he is an excellent finisher - double figures in six of his last seven seasons at this level suggest that he is a good player (including penalties) - his goal tally in 2014/15 probably flattered slightly.
Let us now look at the top ten players in terms of expected goals from last season. The table below shows this data.
| Player | Team | Actual NPG | NPG/90 | Shots/90 | Expected Goals |
|---|---|---|---|---|---|
| James Collins | Shrewsbury Town | 15 | 0.37 | 3.1 | 15.1 |
| Adebayo Akinfenwa | AFC Wimbledon | 13 | 0.32 | 2.6 | 14.1 |
| Jamie Cureton | Dagenham and Redbridge | 19 | 0.48 | 2.4 | 14.0 |
| Matt Tubbs | AFC Wimbledon/Portsmouth | 19 | 0.44 | 1.9 | 13.1 |
| Jake Hyde | York City | 8 | 0.25 | 2.5 | 12.4 |
| Reuben Reid | Plymouth Argyle | 12 | 0.31 | 2.4 | 11.6 |
| Danny Hylton | Oxford United | 8 | 0.20 | 2.0 | 11.2 |
| Barry Corr | Southend United | 11 | 0.33 | 2.2 | 11.2 |
| Ivan Toney | Northampton Town | 7 | 0.26 | 2.8 | 10.1 |
| Lewis Alessandra | Plymouth Argyle | 10 | 0.23 | 2.1 | 9.7 |
We see that Shrewsbury's James Collins had the highest expected goals of any player in League 2 last season and that his actual goals land almost spot on what we expected. Cult hero, Adebayo Akinfenwa, is second on this list, fractionally ahead of the veteran Jamie Cureton.
Looking in slightly more depth, we can see that Akinfenwa had no fewer than eight chances from 'Very Close Range', which is actually more than Plymouth Argyle created in the entire season as a team. However, seven of those eight chances were with the head, which reduces the expected goals slightly. In terms of the most chances from 'Very Close Range' with the foot, Matt Tubbs and Jake Hyde are tied with six chances. Jake Hyde actually scored five of those six chances, which means that his other three goals came from 72 attempts from other locations, which is really quite dreadful.
Oxford United's Danny Hylton is another underachiever here. While he finished the season with a respectable 14 goals, six of those came from the penalty spot, and had he lived up to his expected total of 11.2 NPG, he might have expected to get closer to 17 or 18 goals, which could have fired Oxford closer to a playoff spot.
As we did in the previous article, let us now build in the total minutes played to get an expected goals per 90 minutes table.
| Player | Team | ExpG/90 |
|---|---|---|
| Danny Rose | Bury | 0.44 |
| Jake Hyde | York City | 0.39 |
| Ivan Toney | Northampton Town | 0.38 |
| James Collins | Shrewsbury Town | 0.37 |
| Adebayo Azeez | AFC Wimbledon | 0.36 |
| Jamie Cureton | Dagenham and Redbridge | 0.35 |
| Vadaine Oliver | Mansfield Town | 0.35 |
| Graham Cummins | Exeter City | 0.35 |
| Rakish Bingham | Hartlepool United/Mansfield Town | 0.35 |
| Andy Mangan | Shrewsbury Town | 0.34 |
We can see that Bury's Danny Rose is the top player in terms of expected goals per 90. It will be interesting to see whether he can make the step up to League 1 with Bury next season, but he was clearly one of the outstanding strikers in League 2 last year. Jake Hyde shows well here as well, but he will be hoping to start converting some of his other shots next season. If he can come close to repeating his expected goals tally next year, but convert those attempts at even just a league-average rate, York could have a very decent striker on their hands at this level. Ivan Toney is an interesting prospect. At just 19-years old, this was his first full season and the early signs are promising.
Looking at a couple of players that did not quite meet the 1,000 minute minimum requirement for the earlier table, Billy Kee had an ExpG/90 of 0.62 for Burton and Mansfield, but is currently available on a free transfer. While he only scored three times last season, he could be worth a punt for a League 2 team.
The other interesting name is Oxford United's James Roberts. Having only just turned 18, he made his first team debut this season and was predominantly used off the bench. Despite that, his ExpG/90 of 0.48 is very promising and averaging 2.6 shots per 90 puts him well up there in the division. He could certainly be one to watch next season in what the bookmakers are viewing as a potentially title-challenging team.
This comment has been removed by a blog administrator.
ReplyDelete