written by – Rangers Report
Don’t tell me about heart, an advanced stats analyst for Hockey Graphs, has recently been at the forefront of applying Expected Goals models to the NHL. Having a hockey point of view on this ‘fancy stat’, that has become common place in the fancy stats world of football, adds a new twist on how we can interpret these statistics. I recently caught his Expected Goals table for the NHL on Twitter & really liked how he laid out Expected Goals alongside actual goals & PDO.
The twist that I appreciated the most was transferring Expected Goals for & against into a ratio, similar to Total Shots Ratio. For example, if a team is averaging 2.0 xG a game & allowing 1.0 xGA per match they are creating 67% of the Expected Goals per game. In other words, they are generating better scoring opportunities then the opposition & much more of them. Over the long haul, that team should win the majority of their matches. Don’t tell me about heart also used the same approach for actual goals scored & allowed & then listed all of this alongside PDO. If you combine those three stats you can get a real sense of which teams are maximizing their performances, which teams are trending upwards, & which teams are mired by poor play.
Previously when I have looked at Expected Goals in relation to PDO, I only looked at the offensive side. However, in hindsight that analysis was incomplete because it overlooked the defensive efforts of a team & their ability to limit quality chances. Expected Goals Percentage (xG%) provides much more insight to the overall performance of a team by looking at the ratio of xG & xGA.
Here’s the Scottish Championship stats so far, with teams ranked by Expected Goals Percentage.
|Team||xG per game||xGA per game||xG%||GF per game||GA per game||GF%||GF% – xG%||PDO|
|Queen of the South||0.91||1.23||0.43||1.00||1.50||0.40||-0.03||0.933|
- The Championship’s two best teams meet on Sunday. Rangers & Hibs dominate in nearly every advanced stats metric & if they can sustain their play this season it won’t be a surprise to see both teams earn promotion.
- Rangers produce 77% of the Expected Goals in every game & 85% of the actual goals. Yes, there is a discrepancy but some of that comes from teams having to change their tactics once Rangers take a lead which can lead to late goals in bunches. Plus, Rangers Expected Goals has been steadily declining this season as teams begin to put more & more players behind the ball. Either way, they are dominating the rest of the league, but Sunday will be their biggest test yet.
- Look out for Livingston. They have recently made a change in goal which has had a significant impact on their goals allowed & they have begun giving productive players like Jordan White & Jordyn Sheerin more playing time. White & Sheerin have only played about 40% of the minutes this season but have undoubtedly been their team’s most productive players (White has 0.92 points per 90 & Sheerin has 0.74. The next best on Livi has 0.44). Livingston are generating 53% of the Expected Goals in their matches but only 37% of the actual goals. That distinct gap suggests that there is real room for improvement. Their 4-1 victory over Queen of the South may be just the beginning for a team left for dead just a week or two ago.
- If there’s a team that may begin to regress on this list it may be Falkirk. They play an effective defensive game but tend to struggle to create regular scoring chances. The fact that they score 63% of the goals but only generate 52% of the Expected Goals highlights their relatively high PDO. Those underlying stats suggest that their PDO is in danger of regressing towards 1.000 – which would come from their shooting percentages &/or save percentages taking a dip. They have the highest shooting percentage in the league, so expect a decline there. Lucky for the Bairns they have built up a seven point lead over Queen of the South, who show no real signs of improving any time soon.
Speaking of PDO, the next logical step is to examine this stat, that ostensibly measures luck, & look at it in direct relation to Expected Goals Percentage. Again, the higher a team’s xG% the more likely they are to sustain their success, while a lower xG% suggests a team may struggle to get points.
xG% is along the horizontal axis & PDO is going up the vertical axis.
- The top four teams in the league table are Rangers, Hibs, Faklirk & Raith Rovers. The combined high xG% & PDO suggest that Rangers & Hibs are good enough to sustain their high PDOs – basically they are good enough to keep creating their own ‘good luck.’
- Raith Rovers & Falkirk should maintain their spots in the table but may be more susceptible to a decline in results, especially Falkirk whose PDO is dangerously high.
- Again, watch out for Livi. They have a very low PDO, but their xG% is above 50%. They are second from the bottom in the table, but part of that may be from ‘bad luck.’ Both their save percentage & shooting percentages are below the league averages. Their save percentage is particularly bad (0.608). However, they have recently replaced Darren Jamieson (who has the second worst Goals Minus Expected Goals in the league) with Mark McCallum. The former Scotland U19 goalie is on loan from Dundee United & has had much more success in goal then his predecessor. If he can continue his strong play & if Livi continue to play some players that they have overlooked for much of the season they are primed to shock some teams going forward.
- The fact that St Mirren, Morton, Queen of the South, & Dumbarton all have xG%s less than 50%, it is difficult seeing them improving anytime soon. St. Mirren in particular, despite a solid performance against Rangers, are poised to regress. Their actual goal differential percentage is 44% which is outperforming their dreadful xG% (39%). Their PDO isn’t too high, but the stats suggest that changes will likely have to be made before you see drastic improvements in their results.
You can follow Rangers Report on Twitter @TheGersReport