written by – Rangers Report
Three minutes & twenty-four seconds.
That’s how much time Rangers spent in the final third in the Scottish Cup final after reaching there with a controlled entry. That represents 3% of the 90 minutes from the loss to Hibs. In those three minutes, Rangers got both of their goals & that’s really about it. They generated no other shots, zero free kicks, & just two corners. In those three minutes they gave the ball up more than anything — with six turnovers, five failed crosses, & they were dispossessed of the ball 12 times.
All season long I’ve been curious to see what works when Rangers have success in the final third & consequently what doesn’t work when the attack feels stale & empty. The eyeball test (aka the subjective opinion) has filled the Twitter feeds & message boards for a while now – not enough tempo, too many passes, not enough balls into the box….JUST SHOOT THE BALL!!!!!!
We have screamed all of the above & more at the television (usually accompanied by especially salty language). But, as the dust has settled from the disappointment that was the cup final — I decided to attempt a focused study on what happens in the final third when Rangers saw a great deal of success & then the other end of the spectrum — what happened against Hibs.
I decided that it was imperative to narrow the focus to when Rangers enter the final third with controlled entries (direct passes to a teammate or when a player dribbles the ball into the area). The numbers suggest that this style of play was emphasized by Mark Warburton & the results all season have supported that this was the best approach for a team like Rangers (who tend to have more talent then their opponents).
For example, in an examination of eleven matches from the tail end of the season – Rangers entered the final third 776 times. Of those entries, 520 were controlled (67%) & they led to a positive result 65% of the time. The other 256 entries were classified as non-controlled entries (long balls, passes into a crowd of defenders or thru-balls) & they led to a positive result 30% of the time (that’s a 70% failure rate). In those controlled entries, Rangers generated 83 shots & only 15 shots when they had a non-controlled entry.
Furthermore, when Rangers were playing within their system – building from the back & entering the final third with control – the results were consistently better. In the matches in which at least 70% of their entries were controlled – they averaged 11.8 shots from those entries. When their build-up play was being stymied & they didn’t reach that 70% rate – they only averaged six shots a game on controlled entries. One of those games was against Peterhead. If you remove those results (they are a League One team) – Rangers averaged 4.6 shots in matches with a less than 70% rate of controlled entries.
So, I decided to review a match in which Rangers had a great deal of success on their controlled entries & compare the results to the loss against Hibs. I decided to hone in on a 4-0 victory over Morton from late September.
In that game, Rangers spent 8 minutes & 48 seconds in the final third & in that time they scored three goals, generated 14 shots, earned six corners (one of which led to a penalty), & two free kicks. It should be noted that still had twelve turnovers, ten failed crosses, & had the ball dispossessed 18 times. It wasn’t all marvelous – but the ultimate results were quite impressive.
The objective was to identify trends of what was happening when Rangers entered the final third – both in their style of play & also seeing which players were having the biggest impact on the results. This why it was critical to only focus on controlled entries into the final third. The team had control of the ball – what happened when they successfully retain that possession & turn it into something positive? Who were the catalysts for that success? Also, who were the biggest culprits of losing control of possession in the final third?
In both matches, I tracked each controlled entry & tallied: how long Rangers kept the ball in the final third after the entry, successful passes, successful crosses, failed crosses, shots, free kicks earned, turnovers, & when players were dispossessed. I then combined that data to calculate total possessions by players in the final third & determined their success rate (pass, shot, successful cross), along with their rate of losing the ball on turnovers & being dispossessed. (Much of this work was motivated by the ‘microstats’ that Tempo Free Hockey have been applying to their coverage of the Stanley Cup Playoffs).
Even though the matches were played eight months apart, the lineups were pretty similar. James Tavernier, Danny Wilson, Rob Kiernan, Lee Wallace, Andy Halliday, Jason Holt, Gedion Zelalem, Martyn Waghorn, & Barrie McKay started both matches. Zelalem was subbed out at the same time in both games & was replaced by Dean Shiels both times. In the game against Morton, Nathan Oduwa started instead of Kenny Miller, even though Miller did enter the match for the last half hour.
So, the defense & midfield alignments of the 4-3-3 were identical, while Waghorn was the main striker in September & then replaced Oduwa out right for the Scottish Cup Final. Miller was the striker against Hibs.
Nicky Law played the last 15 minutes of the match in September, while Nicky Clark featured against Hibs for the final 15 minutes.
September 27, 2015
Rangers 4 Morton 0
Individual results: you may have to scroll to the right to see all of the data.
|Player||Min||Pass||Successful Cross||Failed Cross||Shot||Free Kick Earned||Turnover||Dispossessed||Total Possessions||Success Rate||Turnover/Dispossessed rate|
That’s a lot of info – but here are some observations that stick out:
- Rangers trio of midfielders (including the subs) had a combined 46 passes in the final third (41% from Zelalem), while the front three (again, including the sub) combined for 49 successful passes (45% from Oduwa). If you look at total possessions, the split is 66 to 71. That really seems to symbolize the six attacking players supporting each other, along with Lee Wallace & James Tavernier charging up the flanks.
- The data also supports Mark Warburton’s decision to remove Zelalem & Jason Holt at the 60 minute mark. Jason Holt was barely involved at all in the play in the final third & when he was – he was dispossessed 22% of the time (the third highest rate). While, Zelalem was a real focal point in the attacking third (only Oduwa had more possessions – despite the American playing 30 less minutes), the teenager had three turnovers & was dispossessed four times. His success rate was third worst on the team.
- Barrie McKay’s success rate was quite low & this was largely due to three failed crosses & four turnovers. The crosses would not be a point of concern given that Warburton’s system does not stress getting crosses into the box. Only 7% of the 184 possessions in the final third involved crosses.
- I’m not sure what to make of Martyn Waghorn’s statistics. He is hard-charging & assertive when on the ball – which in this match led to him losing the ball 41% of the time he had it. The trade-off was that he scored two goals in the run of play (he added a penalty as well).
Based on this:
- Lee Wallace & Nathan Oduwa were the most important players in the final third & a large part of Rangers attack was going through Gedion Zelalem – but his results were mixed. That seems to represent a real sense of balance to Rangers attack given that Wallace attacks on the left, Zelalem centrally, & Oduwa was predominately out on the right in this match.
Yes, this is an isolated game from what now feels like a long, long time ago – but in some ways it sets a standard. Rangers were effective in the final third & while it should be noted that Morton’s defence was quite porous this season when it comes to allowing shots – we all know that Rangers are at their best when they are dictating the terms of play.
Before we look at the results from the Hibs match — let’s identify a few other trends from the data.
Number of passes
- Rangers had 6+ successful passes on a controlled entry four times, this led to two shots & a positive result 100% of the time
- Average time spent in final third: 17.85 seconds
- Average time player spends with ball (time per possession): 2.30 seconds
- They had 3-5 passes on a controlled entry 15 times, leading to five shots & a positive result 80% of the time.
- Avg. time spent in final third: 12.19 seconds
- Avg. time per possession: 2.67 seconds
- They had 1-2 passes on a controlled entry 28 times, leading to five shots & a positive result 54% of the time.
- Avg. time spent in final third: 7.19 seconds
- Avg. time per possession: 3.08 seconds
- Rangers had zero successful passes on a controlled entry 14 times. Basically, this is when a player enters the zone on his own & either turns the ball over, is dispossessed, or truly goes solo on the attack. This led to two shots, & a positive result 57% of the time.
- Avg. time in final third: 4.77 seconds
- Avg. time per possession: 4.53 seconds
Against Morton, a team that struggled to prevent shots this season – the best tactic clearly was to break down the defence by retaining possession in the final third, through a quick series of passes. Considering that Nathan Oduwa was the most effective player in the final third in this match, you wonder what impact his pace had on these results?
When Rangers passed the ball more in the final third they were more likely to generate a shot & the more direct they tried to be – the less likely they were to get a shot on goal.
Just cross the ball into the bloody box? C’mon!!
This simply isn’t Rangers approach to their play in the final third. That’s not to say they won’t cross the ball – they just won’t rely on it to generate scoring chances. Of Rangers 17 crosses in this match, only 17% were successfully completed to a teammate. Compare that to the 138 attempted passes along the ground in the final third – the completion rate was 91%. I don’t think we need a graph to show the difference here!
But, again, crosses can have their impact – especially if the opposition isn’t necessarily expecting you to play them. Of Rangers shots against Morton, 14% were generated because of crosses — the only two successful crosses Rangers had in the match after a controlled entry.
Playing with a higher tempo
I think any observer this season would agree that Rangers were better when they played with a higher tempo. When play slowed down, so did the results. As the season went on, teams deliberately tried to stifle Rangers tempo with the layers of defence they would put in between the goal & the ball.
With the data tracked, I have attempted to measure tempo & apply it to a variety of results. For example, for all of Rangers shots against Morton – they averaged 3.07 successful passes in the build-up to the shot in the final third & players averaged 2.32 seconds per possession. Conversely, when Rangers turned the ball over in the final third (after a controlled entry, of course) – they averaged 1.32 successful passes in the build-up to the eventual turnover & 3.46 seconds per possession. A second barely registers in regular life (you’ve lost several just reading up to this point) but a second in football can make all of the difference in the world – think of how many plays have broken down because a player hesitated just for that one second & defenders were able to close down on the ball.
Here is a graphical representation of a variety of results in the final third, along with the number of successful passes leading to that result & the average time players averaged per possession.
The magic number here seems to be that three second mark per possession. When players moved the ball quickly (under three seconds) the results were either a Very High Danger shot or a High Danger Shot. These are shots coming from the heart of the penalty area on down to the goalie’s six yard box. This is where most goals are scored. Also, you an see the shots they either beat the keeper for a goal or hit the post or forced the goalie into a save also came when the average time per possession was under three. Missed shots & blocked shots also fell into this category.
The times in which Rangers averaged more than three seconds per possession resulted in shots from Low & Medium Danger areas, turnovers, or the ball being dispossessed.
First the shots, shots are good – but not all shots are created equal. Low Danger shots come from outside the penalty box, while Medium Danger shots come from the flanks within the box. These shots have low (really low) success rates. The visual below represents Rangers shots from this past season (they do not include headed shots).
It is interesting to see when Rangers slowed down the play, they had to rely on lower percentage shots against Morton. Also, notice that the number of passes in the build-up is also much less than the more dangerous shots.
The fact that a slower tempo led to a higher rate of turnovers & being dispossessed should not be a surprise. The slower the ball moves, the more time a defender has to read the play.
Still with me????
We isolated one match in which Rangers were very successful in the final third & set that has a baseline to measure the calamitous performance against Hibs.
In their limited time in the final third on controlled entries (remember it was just 3 minutes & 24 seconds), Rangers only generated two shots (both goals) & 24% of their possessions (touches of the ball) ended with a turnover or being dispossessed (compared to 16% against Morton).
Their pass completion rate in the final third was only slightly less at 89% & (if you’re curious) the cross conversion rate was 17% (identical to the Morton match & again the only successful cross led to a goal). So, there’s no real variance here.
Before we continue the search for the magic bullet – let’s see how each player did against Hibs:
|Player||Min||Pass||Successful Cross||Failed Cross||Shot||Free Kick Earned||Turnover||Dispossessed||Total Possessions||Success Rate||Turnover/Dispossessed rate|
- The only player with a high number of possessions in the final third after controlled entries who also had a high success rate was Barrie McKay. The others with a lot of possessions: James Tavernier, Martyn Waghorn & Kenny Miller had the three highest turnover/dispossessed rates on the team.
- Gedion Zelalem, who was singled out for his solid play in midfield by most observers, did not have nearly the same impact in the final third as he did against Morton.
- Remember against Morton, there was a real balance between the front three & the three midfielders when it pertained to successfully passing the ball in the final third. In the cup final, the front three (including the sub) had 25 successful passes, while the midfield (including the sub) had 14. When it comes to total possessions in the final third the split was 38 from the front three & 19 from the midfielders.
As you can see the balance of the middle three players working with the front three disappeared against Hibs. Additionally, Rangers relied on Lee Wallace & James Tavernier more than they did against Morton. This brings Waghorn & Miller’s rate of losing the ball into an even brighter spotlight. Waghorn had the same issue against Morton but Kenny Miller certainly did not.
Next I applied the same average rates of total possessions & success ratios from the Morton game to the Hibs match to see what players were close to the producing at the same rate. You may want to look away….
Using the Morton match as the baseline, you can see the only player approaching the Critical to Success area of the graph would be Barrie McKay. You can see that Lee Wallace was effective in the final third, but not nearly as involved as he was against Morton. James Tavernier, in particular, had a poor match. Of his 12 total possessions he only had four successful passes & one successful cross (which did lead to Kenny Miller’s goal…so it wasn’t all bad). The rest of the time he was giving the ball back to Hibs, either with via a turnover, a failed cross, or being dispossessed. His turnover/dispossessed rate was the highest on the team.
Also, when you compare the graphs from the two matches – it is interesting to see that the only player who saw little, to no variance in his performance was Jason Holt. The most drastic dip in relevance in the final third was certainly Gedion Zelalem.
Of course, the biggest difference between these two matches comes down to sheer volume. Rangers spent 8+ minutes in the final third because of controlled entries against Morton & only 3+ against Hibs. They had 184 total possessions/touches in the final third in the Morton game & only 75 in the Cup Final.
Rangers had a possession rate of 67% against Morton in that September match & 60% against Hibs. That’s not a huge difference, but was it enough or was it a case of Rangers build-up play failing to get the ball into the final third on controlled entries?
Remember that stat about when Rangers had 70% of their total entries be controlled that they averaged 11.8 shots on those entries & when the didn’t their average shots on their controlled entries went way down? Yeah, from like an hour ago when you first started reading (skimming) this post. Well, against Hibs 53% of their entries were controlled. That means 47% of the time they got into the final third it was via a long ball, a pass into a crowd of defenders, or a thru-ball.
My first instinct was to blame the midfielders & they are partly to blame. Jason Holt only had one entry into the final third (it was controlled), while Andy Halliday had two (both non-controlled). Gedion Zelalem had six total entries & five were controlled. Dean Shiels (channeling his inner 2013-self) had zero entries into the final third. The fact that the midfield only combined to generate eight entries into the final third (12% of total entries) highlights the fact that they were outplayed by Hibernian’s midfield & I doubt any subjective (eyeball test) opinion would differ there. The stats support what we all saw.
But a good deal of the criticism should also come down to James Tavernier & even, Barrie McKay as well. They were responsible for a combined 38 entries into the final third (56% of the team’s total). However, both players were forced into (or chose) non-controlled entries the majority of the time. Eleven of Tavernier’s 20 entries were non-controlled, while ten of McKay’s 18 entries were. They combined for a positive result 29% of the time on those non-controlled entries (including one shot) which is actually not that bad. Of course, their controlled entries resulted in a positive result 53% of the time (not that great but it includes both goals & the only two shots generated on controlled entries).
Credit should also go to Hibs defence. They clearly were forcing Tavernier & McKay into getting the ball into the final third before they may have wanted to. But Hibernian were also able to limit Rangers ability to breakdown the defence once they were in the final third. You can see this when you compare the number of passes & time spent in the zone per entry.
Here are the stats against Hibs (with Morton in parentheses to save you from scrolling back up)
- Rangers had 6+ passes on a controlled entry zero times (4 against Morton)
- They had 3-5 passes on a controlled entry eight times (15), leading to zero shots & a positve result 63% of the time (5 & 80%)
- Average time spent in final third: 10.17 seconds
- Average time player spends with ball (time per possession): 2.57 seconds
- They had 1-2 passes on a controlled entry 14 times (28 against Morton), leading to one shot & a positive result 64% of the time (5 & 54%)
- Average time spent in final third: 5.50 seconds
- Average time per possession: 2.41 seconds
- Rangers had zero successful passes on a controlled entry ten times (14), leading to one shot (off of a cross) & a positive result 20% of the time (2 & 57%).
- Average time spent in final third: 3.60 seconds
- Average time per possession: 3.42 seconds
Of Rangers controlled entries into the final third, Rangers were limited to two or less successful passes 75% of the time & their time in the final third never lasted more than 15 second (even against Morton only six of their entries lasted longer than 15 seconds). You don’t get much time on the ball in the final third per entry – so any build up play needs to be done quickly & with a high tempo.
Speaking of tempo – Rangers averaged 1.78 seconds per possession on their two goals. This shouldn’t be a surprise since they were both ‘bang-bang’ plays that saw no real build up in the final third. Miller’s goal came from a cross as Tavernier entered the zone & Halliday’s was a first time shot from a Barrie McKay pass into the attacking third (which came after a throw-in). So, tempo really can’t be measured on these two plays.
But if you apply our same measure of tempo to all of the entries that ended up in a positive result (shot, corner, cross into box, or retaining possession while leaving the final third) you’ll see that once again the quicker Rangers move the ball – the better the results. When you compare the tempo on the positive results to when Rangers had a turnover or were dispossessed – you’ll see again that the magic number is three seconds.
This may feel like the longest article ever to you (I sure know it does for me) & ironically, we can’t put too much weight on the results because only two games are examined. One really good match & one really bad one. They are, however, isolated snapshots that really do highlight how you can apply statistics to not only support what your eyes told you but also to hone in on some objective data that can be used to find entry points to improvement on the pitch.
We may have also learned why Mark Warburton has put such an emphasis on improving his pool of midfielders for next season & also why he may want to look into a speedy, attacking player for the front three given the impact Nathan Oduwa had on the results from the Morton match.
All right, I’m done. Thanks to all of you who actually read all of this…
You can follow Rangers Report on Twitter @TheGersReport