HomeTeams - PLChelseaSteven Gerrard and Frank Lampard in 2013/14 | A Statistical Comparison

Steven Gerrard and Frank Lampard in 2013/14 | A Statistical Comparison

There has been considerable talk about the 35-year old Frank Lampard leaving Chelsea at the end of the current season as part of the Jose Mourinho house-cleaning that might also include John Terry and Ashley Cole. In fact, rumours of Lampard joining the Los Angeles Galaxy in the MLS have been spread over the past two seasons.

Furthermore, the recent run of injuries Steven Gerrard has been accumulating over three of the past four seasons serves as a nagging reminder that at 33 years old, the next injury could also be his last.

It is not inconceivable that this season might be the final chance to examine how these two celebrated British footballers are currently performing on a head-to-head basis. Are they still producing numbers worthy of their prominent history and does another EPL season seem realistic for the future of either or both of these two players?

Editors Note: This article was written before the Chelsea-Tottenham game so the statistics are up to 7th March 2014

Gerrard Vs Lampard 13-14

Gerrard Vs Lampard: Comparing the Numbers

It is important to recognise from the outset of this analysis that there is a very large discrepancy in the game minutes that each of these players has accumulated during the first 28 league matches this season.

Steven Gerrard has played 80.3 percent of the minutes available while Frank Lampard has been able to accumulate only 55.4 percent over the same period. The reason for this incongruity is found in the very different roles they are playing for their respective clubs.

Steven Gerrard is clearly the linchpin for Brendan Rogers in connecting the wingers and forwards in the aggressive Liverpool attack. Gerrard is selected for the Reds starting eleven whenever he is healthy.

Frank Lampard has not been so fortunate. Jose Mourinho’s voluminous midfield selection has often excluded Lampard as a starter for league play in favor of a sporadic substitute role or in a more prominent role only during cup play. As a result Lampard’s opportunity for “minutes” has been comparatively limited this season for EPL games.


In order to accommodate the large advantage of Gerrard’s game minutes over that of Lampard’s, the quality of metric performance will be added to the cumulative metric numbers in the performance comparisons. So, as an example, when I examine passing performance, the number of successful passes as well as the efficiency of passes completed will be included.

Gerrard Vs Lampard: Establishing Statistical Difference

Additionally, in comparing performance metrics, we are going to supplement some of the more amenable metrics with statistical analysis. And by amenable, I mean if the sample size of measures for the specific metric is sufficiently large. If so, the statistical difference reflected by the numbers will be presented in terms of the odds or chances of these differences normally occurring. This will elevate the findings beyond an opinion of whether one player outshines the other to a clinical analysis looking for statistical superiority. Here is what I mean:

In sports analytics, as in most statistical analyses, the most common accepted value for establishing a statistical significance is when the difference or relationship you are examining could have normally occurred by chance only 5% of the time or less. Or, to say it another way, it would have to be so rare that the odds of it occurring are 20 to one. Here are a few examples of common chance events:

The ability to accurately call the flip of a fair coin would be 50% (1 in 2) or to predict the correct suit of a card would be 25% (1 in 4), or to predict the exact card to be drawn out of a full deck would be less than 2% (1 in 52).

So, the standard of statistical significance is quite conservative and is not a normally encountered event. It suggests special circumstances. That is, very different from normal circumstances. It is more natural to expect no statistically significant difference.

Here are findings across the game metrics examined for Steven Gerrard and Frank Lampard that include passing, creativity and attacking.

Gerrard Vs Lampard: Passing

Gerrard completes 1253 of 1465 open play passes compared to Lampard’s 725 of his 880 attempts. Their open play passing accuracy seems very similar: 85.5% to 82.4%, respectively. However, this 3.1% difference is based upon a very large sample of data—enough to establish that Gerrard performs at a significantly superior level than Lampard. In fact, the chance that this 3.1% difference would normally occur is only one time in 23 or approximately 4%.

Additionally, for midfielders who typically play in a central role or just behind the front line like Gerrard and Lampard, the most critical passing occurs in the final third of the pitch. In this situation Gerrard is better than Lampard by an 11.3% difference — 73.3 to 62.0%, respectively. The chance of this large difference occurring normally is only one time in 1639! Again, Gerrard’s performance is significantly superior to Lampard’s in passing accuracy in the final third of the pitch.

Gerrard Vs Lampard Creativity

Although Gerrard is achieves 12 out of his 17 successful dribbles compared to Lampard’s 6 out of 10 attempts, this 70.6% to 60% advantage is not statistically significant because of the small number of measures. The odds of this outcome are less than 2 to 1 and are quite common. Remember the importance of sample size? There is simply not enough information to be conclusive.

A similar outcome occurs when contrasting crossing skills. Although Lampard’s crossing accuracy of 41.0% (26 completions out of 64 attempts) is superior to Gerrard’s 28.0% (37 completions out of 131 attempts), it also fails to achieve the standards of statistical significance. It can be viewed as a statistically interesting difference although not clinically significant (close but no prize). The chance of this difference occurring is about 8% or one in 12. It’s different, but not different enough to conclude that Lampard’s crossing is significantly better than Gerrard.

So, what different performance metrics might Lampard have delivered a statistically significant difference in crossing success percentage? Here are two hypothetically possible ways:

  1. If Lampard had completed 28 instead of 26 of his 64 cross attempts, he would have improved his completion percentage to 43.8%. This higher 15.5% advantage would have established a statistically significant difference—one that would only occur one time in 32 occasions or about a 3% chance event.
  2. A second way would be for Lampard to have achieved exactly the same percentage of successful crosses (41%) but to have made more attempts. The exact breakdown would be 48 successful crosses in120 total attempts. This would yield a statistically significant outcome of occurrence of exactly a one chance in 20.

The chances created in open play (including assists) illustrates that Gerrard has created more opportunities than Lampard for his team-mates in terms of both quantity—53 to 26 total chances —and frequency of supply—38 to 54 minutes between chances.

The clear-cut chances supplied by the two are very sparse because these events, like goal assists, are relatively rare events. Clear-cut chances supplied represents goal scoring opportunities that the player provides for his teammates that is viewed as a one-on-one situation. Gerrard outshone Lampard 7 chances to 1 in this category. No statistical comparison is offered given the sparse number of events, however, the raw numerical difference is difficult to ignore.

For central midfielders, goal assists are one of the most important performance measures. Unfortunately, the small sample size allows very little deconstruction of this metric. Contrasting Gerrard’s 9 assists with Lampard’s 3 is, at first glance, overwhelming. However, the fact that 7 of the Gerrard’s assists were from set plays needs a bit of moderation in assessing. Gerrard’s role with Liverpool is to take almost every set piece with the exception of those he shares with Luis Suarez when the location is close to the area. Conversely, Lampard’s current role has been changed from the primary choice of taking almost all set pieces to being, at best, second choice to Eden Hazard. The opportunity differential is, again, an important point to note.

Gerrard Vs Lampard: Attacking

Shooting accuracy—the ratio of shots on target/total shots—gives Gerrard a 13.1% advantage over Lampard (65.6% to 52.5%). Even though Lampard played less than 70% of Gerrard’s game minutes, he shoota much more frequently. Lampard ventured a shot on goal roughly every 35 minutes while Gerrard’s frequency reflected a more economical rate of once every 63 minutes.

Regardless, Gerrard’s edge in accuracy (21 shots on target out of 32 shots taken) was not sufficient to establish a statistically significant margin over Lampard’s 21 on target shots out of a total of 40 shots taken. The chance that this 13.1% margin could have occurred for the sample of shots taken was about 26% or one chance in 4. There is no statistically significant difference between Gerrard and Lampard’s shooting accuracy. Repeat after me: “Damn those small samples.”

Shooting effectiveness or chance conversion—the ratio of the percent of goals scored/shots on target—yield very similar results to the shooting accuracy. Again, Gerrard converted 8 of his 32 shots (25.0%) into goals compared to Lampard’s 5 out of 40 (12.5%). The statistical chance of this difference occurring is less than one in six or about 17%. The conclusion, once again, is that Gerrard and Lampard do not display statistically different shooting effectiveness. Keep in mind the number of events in the sample. It is always an important and mitigating factor in the determination of statistical significance.

The final metric is clear-cut chances or shots that have been described as “can’t miss” (but sometimes are) or sitters. Steven Gerrard encountered and converted 6 out of 7 of these opportunities (85.7%) compared with Frank Lampard’s success rate of 2 out of 4 (50.0%). Again, we suffer the fate of small sample sizes when dealing with attempting to establish statistically significant differences and, alas, we cannot. The chance of this difference occurring given our miniscule sample sizes are about one chance in 5 or 20%.

Gerrard Vs Lampard: Conclusion

Even with the disparity of playing opportunity, with the exception of Gerrard’s superior overall passing skills, including those especially critical in the attacking third, it was not possible to establish a convincing statistical superiority in performance between the two warhorse midfielders. However, the consistent numerical advantage that Gerrard displays over Lampard across most measures begins to tell a story about whether it is statistically supported or not. It is important to not ignore the steady advantage of performance that the numbers illustrate even though they are only descriptively compelling. The story told does not seem to waver.

Nevertheless, the conditions of opportunity have played a major role in this season’s impact on the two players. While Steven Gerrard is currently enjoying one of his best seasons in years as Liverpool look to him as the centerpiece of its attack in each and every game, the conditional playing opportunities offered Frank Lampard paints a different and not so promising future at Chelsea.

It is a fair bet that at Anfield on April 27th, given the current title race, the seasons last game between Liverpool and Chelsea might be one of the most memorable in years between these two great players—and very likely the last with their respective teams for one of them.

Joel Oberstone
Joel Oberstonehttp://www.twitter.com/JoelOberstone
Joel Is an avid football and modern jazz fanatic. He sees the connection between the improvisational elements of each ... the connection between Andres Iniesta and Lionel Messi as well as Miles Davis and Bill Evans. He wrote a weekly column for the Wall Street Journal Sports Europe between 2010 and 2011 using a demystified style of sports analytics to explain the details of football performance. Joel is a professor of Business Analytics at the University of San Francisco, School of Management. He is also an ardent fan of writers Mick Dennis, Barney Ronay, and Jonathan Wilson and the never-ending word wizardry of former Newcastle United midfielder Ray Hudson in his La Liga match calls and commentary.
More News


Please enter your comment!
Please enter your name here