It was a wild year for Chelsea: two managers, a European trophy, a huge league improvement to cruise into 3rd place and a great present in Mourinho's return once the season had concluded.
There were lots of interesting things going on in regards to Chelsea's on field Premier League performance, but two things stood out for me. One good, one bad.
The good: Chelsea's strong performance when the game was tied (Tied Game State) and I am going to look into the underlying driving forces behind this.
The Bad: Chelsea's weird performance when leading by a single goal (Plus 1 Game State).
Considering how good the good was it makes it all the more baffling that the bad was so bad. All will be explained.
The Good
And my it is good. Chelsea's Goal Difference at Tied Game State was +19. When the game was tied Chelsea recorded a +19 GD, it was worth repeating because it was mighty impressive.
| Tied GD +/- | |
| Man United | 13 |
| Man City | 11 |
| Chelsea | 19 |
| Arsenal | 13 |
It's the best number in the league and by a distance. The question is what was powering that mightily impressive GD?
Let's look at Chelsea's Total Shot Ratio and Shots On Target Ratio.
There really is nothing here that gives us a clue as to why Chelsea outscored the opposition to such an extent. Infact, Chelsea's SoTR dips below 50% at games 25-30, which means Chelsea at that point were being outshot by the opposition. Chelsea's TSR is good, but it is less important that SoTR.
The ratios don't help us, Chelsea were not out shooting the opposition to any extent, let alone an extent that would help us to figure out that excellent Tied GD.
How about the Scoring and Save percentages?
Now we are talking. Scoring% league average is ~30%, Save% league average is ~70% and PDO is 100. As you van see in the chart above not only was Chelsea's Tied Game State Sc% and Save% incredibly consistent as the season wore, those percentages are also more than a little high.
Scoring% is 10 percentage points above the league mean and save% is about 9 percentage points above the league mean. Now, these percentages may be failing to capture Benitez's tactical setup, which if these numbers are anything to go off, may have bordered on genius for recording a PDO of 119 is no mean feat.
Chelsea's Tied goal difference was the best in the league, Chelsea's shots ratio's were not nearly high enough to be classed as the cause of such a good goal difference. Instead, Chelsea's performance at Tied Game State was driven by the percentages in the chart above. There are 3 reasons, that I can think of at least, that are driving those percentages:
- Chelsea were employing a superior tactical setup than the rest of the league at Tied Game State?
- Chelsea had the best goalkeeper (probable) and the best players at converting shots on target into goals in the entire league when the game was tied?
- A dollop of luck?
Considering how consistent Chelsea's Scoring and save percentages were, we may have to think long and hard about assigning the excellent Tied GD to just luck. Whilst Chelsea have the best goalkeeper in the league it can be argued they don't possess the best finishers in the league. Could it have really been Benitez's tactics that were behind this stellar performance? I do not know, for I feel some facet of Chelsea's excellent performance is not being captured with the numbers I possess.
Section two of this article is where things get really interesting. Those high percentages at Tied, those guesses at Benitez's tactical setup are completely absent when we look at Chelsea's Plus 1 Game State performance.
The Bad
Chelsea were not a good team at Plus 1 Game State: A Plus 2 Goal Difference at this game state is in stark contrast to the rest of the top 4:
| Plus 1 GD | |
| Man United | 12 |
| Man City | 10 |
| Chelsea | 2 |
| Arsenal | 6 |
This number is sub par when placed against Chelsea's peers, this number also caused Chelsea to drop 18 points from Plus 1 winning positions in 2012/13. Remember that Chelsea were +19 at Tied are have dropped to +2 when leading by a single goal. So how did this Chelsea team, so excellent at Tied, drop so far when leading by a goal?
As with the ratios at Tied, these don't tell us a whole lot. Chelsea's Total Shot Ratio under RdM was, and let me be careful how I put this, shit. The TSR was hovering at about 41% when RdM was canned and the SoTR was dead on 50%. Benitez, or regression, improved both of these ratios as the season wore. But let me reiterate once again that these ratios don't explain why Chelsea dropped so many points and had such a low GD when leading by a goal. These ratios, especially the SoTR, were too far above terrible to explain the poor goal difference.
So what can explain Chelsea's relatively bad performance at Plus 1?
This chart looks mighty different to Chelsea's Tied Game State chart. At Tied Chelsea rode those percentages and here at Plus 1 that luck/tactics/player skill evaporate. Chelsea's PDO finishes at a shade under 100, but Plus 1 Game State usually sees a PDO average north of 100. So why were Chelsea so bad?
Chelsea cannot have employed the best tactical setup in the league at Tied and then just minutes later, when at +1, changed into a tactically weak team. Is it plausable that Benitez struggled to get this Chelsea team to defend a lead? Possible. Chelsea don't have an average goaltender, and they do not have subpar talent on this team.
Could it be bad luck? There is definitely more variance throughout the season in the chart above than there was at Tied Game State, but again I don't feel entirely comfortable, at least not yet, to say that those percentages are all luck, good or bad.
***
The point of this article is not to lecture or to highlight with any certainty why Chelsea were so good at Tied and relatively bad at Plus 1, but merely to put this information out there and try and get some ideas and feedback as what may have caused this Jekyll and Hyde style split in performance at these game states.
Maybe there are tactical reasons that I missed (I am not a Chelsea fan), maybe there are causes that the numbers don't capture. Either way, I am open to suggestions before any belief that the percentages are responsible further ingrains itself into my thinking.
Game States
