The Indian Premier League, which was born in 2008, has turned 10! And look at how things have changed around us - when Lalit Modi declared the tournament open on 18th April 2008, Barrack Obama was yet to become the US President, Virat Kohli was still 4 months shy of debuting for India and the Chinese economy was still growing in double digits. Lalit Modi is no longer associated with the league but during this time, IPL has grown to become one of the world's most valuable sports leagues (worth $4.2 billion and with revenue of $378m in 2016). It'll be interesting, however, to see how various on-field aspects of the game have evolved in these exciting nine years.
I've attempted to evaluate the performances, season-by-season, using tournament stats. Using this approach, we try to address the following questions:
Surprise! Surprise! I was expecting the team totals to climb and show an upward trend. However, if one observes closely, it's clear that there's an increasing trend of more scores between 176-200 vis-a-vis 200+ scores (more orange than grey!), esp. when compared to IPL 2008 and 2010. For example, in 2008, amongst the top 30 team totals, there were 11 scores above 200 while in the last two editions, that figure has come down to 6 (in 2016) and 7 (in 2015), respectively. The median score has remained fairly stable (except for 2009) even though the highest median score was achieved in the first season itself, and has been matched twice since. IPL 2009 season is a bit of an aberration since that edition was held in South Africa (due to general elections in India) where batting conditions are relatively tougher.
In IPL 2008 and 2010 (2009 again an exception), there were 5 and 4 instances, respectively, of match aggregates exceeding 400. This too has come down to 2 in each of the last two editions. However, the 351-400 bucket shows higher frequency in the last few years. The median match aggregate has hovered around 350.
To compute IPL HHI, I've used the points garnered by a team during a season and divided by the total points available during that season (comparable to market shares). The IPL HHI is charted out below:
The first thing that jumps out - more the number of teams, more competitive the league. In 2011, when two new teams were introduced, the HHI fell to 0.106. Similarly, in 2012 and 2013 when there were 9 teams in the league, the HHI was 0.121 and 0.125, respectively. For an 8 team format, 0.131 is the lowest HHI, achieved in 2010 and then again in 2016. It's fair to conclude, looking at these figures, that the tournament has managed to stay highly competitive over the last nine years.
I've attempted to evaluate the performances, season-by-season, using tournament stats. Using this approach, we try to address the following questions:
- Are teams scoring faster now?
- What is a safe score to defend?
- Are batsmen more consistent now?
- Are bowlers bowling better now?
- Are teams evenly balanced?
1. Are teams scoring faster now?
Highest team totals
For this analysis, I've shortlisted the top 30 team scores for each season. Further, I've created separate 25-run buckets for team totals (151-175, 176-200, 201-225, 226-250, 251-275). Then, I've plotted, season-wise, the number of occurrences of team totals (or frequency) within different buckets. Further, on the right side Y axis, I've plotted the median of the top 30 scores in each season. The chart is produced below:
 |
| Data source: ESPNcricinfo.com |
Surprise! Surprise! I was expecting the team totals to climb and show an upward trend. However, if one observes closely, it's clear that there's an increasing trend of more scores between 176-200 vis-a-vis 200+ scores (more orange than grey!), esp. when compared to IPL 2008 and 2010. For example, in 2008, amongst the top 30 team totals, there were 11 scores above 200 while in the last two editions, that figure has come down to 6 (in 2016) and 7 (in 2015), respectively. The median score has remained fairly stable (except for 2009) even though the highest median score was achieved in the first season itself, and has been matched twice since. IPL 2009 season is a bit of an aberration since that edition was held in South Africa (due to general elections in India) where batting conditions are relatively tougher.
Highest match aggregates
Is the above outcome an anomaly? It'll be useful to look at a similar analysis based on match aggregates.
 |
| Data source: ESPNcricinfo.com |
In IPL 2008 and 2010 (2009 again an exception), there were 5 and 4 instances, respectively, of match aggregates exceeding 400. This too has come down to 2 in each of the last two editions. However, the 351-400 bucket shows higher frequency in the last few years. The median match aggregate has hovered around 350.
Conclusion
Based on the above evidence, it's safe to conclude the following:
Teams are scoring more consistently between 175-200 instead of 200+. This is further demonstrated by looking at a statistical analysis of the top 50 team totals in each season. While the average score (of top 50 scores), at 185, has remained the same in IPL 2008 as well as in 2015 and 2016, the variance and standard deviation has come down in the last two seasons. The steeper normal distribution curves in IPL 2015 and 2016 clearly bear this out (comprising top 50 team totals).
Teams are scoring more consistently between 175-200 instead of 200+. This is further demonstrated by looking at a statistical analysis of the top 50 team totals in each season. While the average score (of top 50 scores), at 185, has remained the same in IPL 2008 as well as in 2015 and 2016, the variance and standard deviation has come down in the last two seasons. The steeper normal distribution curves in IPL 2015 and 2016 clearly bear this out (comprising top 50 team totals).
 |
| Data source: ESPNcricinfo.com |
This could be pinned down to batting strategy where teams, at the start of the innings, are now deliberately targeting 175-200 instead of taking additional risks to score 200+, believing that this score is safe to defend. This can have an impact on the team composition as well. Of course, teams will have different strategies for different venues - at Bengaluru, teams typically target 200+ while at Eden Gardens, 160-170 is considered as a good score. But alternately, its also possible that the bowlers have become adept at keeping the runs down as well as picking up wickets at regular intervals.
In the inaugural edition, the average runs per innings (RPI) by all the top batsmen who aggregated 300+ during the season, was 31.3 and the average strike rate was 137.3 per 100 balls. The RPI has gone up from around 29 between 2012-14, to 33.4 in IPL 2016. This is the highest across all editions. Further, the strike rate has dipped only marginally, from 137.3 in 2008 to 136.2 in 2016. Hence, it would be safe to conclude that the batsmen are managing more consistency now without compromising too much on the strike rate. Also, increasingly there are more top order batsmen in the Top 10 batsmen list than in the past.
Let's look at how bowlers' performances have evolved since IPL 2008. For this, I've taken two important parameters - economy rate and bowling strike rate (i.e. no. of balls / wicket). Bowling average has deliberately not been considered since in a T20 match, balls remaining in the innings are a more important resource vis-a-vis wickets in hand. So, I've first taken the bowlers who've taken at least 10 wickets in a season and then tweaked the economy rate computation - instead of runs conceded per over, I've considered runs conceded per 100 balls (let's call it RP100B). This change in computation thus makes it directly comparable to batsmen's strike rate. Here is what we get:
While the RP100B had dipped to 120-125 between 2010 and 2014 editions, the figure has again risen up and is touching 131 in the last two editions. The average strike rate has, more or less, remained in the 18-20 balls / wicket range.
Further, I slice the above by bowling type i.e. pace or spin. This is how it looks like:
Over the years, except for IPL 2015, spinners have done better than pacers in keeping the runs down. On the other hand, as is the norm across formats, pacers have managed slightly better strike rates than spinners. During 2009-2014, the spinners were significantly better than the pacers in keeping the runs in check and they were close to the pacers in terms of strike rates too. This isn't surprising given that this period saw some world-class spinners operate - Warne, Kumble, Muralitharan, Harbhajan, and also saw some excellent talent emerge - Ashwin, Jadeja, Sunil Narine (before his modified action), Amit Mishra, Chahal, Axar Patel etc. Over the last couple of seasons, apart from ever-green performers like Malinga and Bravo, there's a group of seamers who've impressed - Bhubhaneshwar Kumar, Mustafizur, McClenaghan, Bumrah, Russell, Chris Morris etc.
2. What is a safe score?
Based on empirical evidence over the last 9 IPLs, let's try to deduce a safe target when batting first. For this, we've adopted two approaches. First, we've looked at all matches where the team batting first has won with a margin of 15+ runs i.e. comfortable victories. Here's what we get:
 |
| Data source: ESPNcricinfo.com |
In IPL 2008, the average target set when the team batting first won by a margin of 15 runs or more, was 193 and 12 such wins were recorded during that season. The corresponding figures for IPL 2016 were 193 runs and 11 wins. The average target across all editions is 180.
Next, we've taken those matches where the chasing team has won a close match with only 3 balls (or less) remaining. Here's how it looks:
 |
| Data source: ESPNcricinfo.com |
In IPL 2008, the average target which was successfully chased down with 3 balls or less remaining, was 174 and 9 such victories were recorded in that season. The corresponding figures for IPL 2016 were 173 runs and 8 wins. The average target across seasons is 163.
Comparing this with the above figure of 180 (winning comfortably when batting first) and taking into account the 3 balls in hand, it's reasonable to assume (based on empirical evidence) that 180+ is generally a safe target when batting first. For the above analysis, abandoned matches, ties and matches by Duckworth-Lewis method have not been considered.
Comparing this with the above figure of 180 (winning comfortably when batting first) and taking into account the 3 balls in hand, it's reasonable to assume (based on empirical evidence) that 180+ is generally a safe target when batting first. For the above analysis, abandoned matches, ties and matches by Duckworth-Lewis method have not been considered.
3. Are batsmen more consistent now?
How about batsmen? The record for most runs in a season has been improved from 616 runs in IPL 2008 (Shaun Marsh) to 973 runs last year (Virat Kohli). While Marsh scored at a very impressive strike rate of 139.68, both Kohli and Warner (who got 848 runs in 2016) got their runs at an incredible strike rate of nearly 152 per 100 balls. While only three centuries were struck in 2008, Virat Kohli himself scored four centuries in 2016 (overall six centuries were scored). What do the numbers tell us? On a season-by-season analysis of the runs per innings (RPI is slightly different from batting average) and strike-rate (runs per 100 balls) of all batsmen who've scored 300+ runs, we get the following:
 |
| Data source: ESPNcricinfo.com |
In the inaugural edition, the average runs per innings (RPI) by all the top batsmen who aggregated 300+ during the season, was 31.3 and the average strike rate was 137.3 per 100 balls. The RPI has gone up from around 29 between 2012-14, to 33.4 in IPL 2016. This is the highest across all editions. Further, the strike rate has dipped only marginally, from 137.3 in 2008 to 136.2 in 2016. Hence, it would be safe to conclude that the batsmen are managing more consistency now without compromising too much on the strike rate. Also, increasingly there are more top order batsmen in the Top 10 batsmen list than in the past.
4. Are bowlers' bowling better now?
Let's look at how bowlers' performances have evolved since IPL 2008. For this, I've taken two important parameters - economy rate and bowling strike rate (i.e. no. of balls / wicket). Bowling average has deliberately not been considered since in a T20 match, balls remaining in the innings are a more important resource vis-a-vis wickets in hand. So, I've first taken the bowlers who've taken at least 10 wickets in a season and then tweaked the economy rate computation - instead of runs conceded per over, I've considered runs conceded per 100 balls (let's call it RP100B). This change in computation thus makes it directly comparable to batsmen's strike rate. Here is what we get:
 |
| Data source: ESPNcricinfo.com |
While the RP100B had dipped to 120-125 between 2010 and 2014 editions, the figure has again risen up and is touching 131 in the last two editions. The average strike rate has, more or less, remained in the 18-20 balls / wicket range.
Further, I slice the above by bowling type i.e. pace or spin. This is how it looks like:
 |
| Data source: ESPNcricinfo.com |
Over the years, except for IPL 2015, spinners have done better than pacers in keeping the runs down. On the other hand, as is the norm across formats, pacers have managed slightly better strike rates than spinners. During 2009-2014, the spinners were significantly better than the pacers in keeping the runs in check and they were close to the pacers in terms of strike rates too. This isn't surprising given that this period saw some world-class spinners operate - Warne, Kumble, Muralitharan, Harbhajan, and also saw some excellent talent emerge - Ashwin, Jadeja, Sunil Narine (before his modified action), Amit Mishra, Chahal, Axar Patel etc. Over the last couple of seasons, apart from ever-green performers like Malinga and Bravo, there's a group of seamers who've impressed - Bhubhaneshwar Kumar, Mustafizur, McClenaghan, Bumrah, Russell, Chris Morris etc.
5. Are teams evenly balanced?
One of the more remarkable and less-talked-about things with respect to IPL, is the concept of "equal purses" for all franchises. It started out as "equal auction purse" in 2008 ($5m for each team) i.e. all teams were allowed to spend upto $5m on buying players in the auction. This figure was increased to $9m in 2011. However, there was no limit on what teams could spend outside the auctions (i.e. trading players between each other). This was then further regulated in 2013 when IPL governing body stipulated an upper limit of $12.5m for the entire squad ("equal purse"). This equal purse concept has ensured a level playing field.
I've tried to analyze the competitiveness of the league for each season, using a measure called Herfindahl Index (Economics students will be aware of this concept). It's a measure of the size of firms in relation to the industry and an indicator of the amount of competition among them. HHI is an economic concept widely applied in competition law, antitrust and also technology management. It is defined as the sum of the squares of the market shares of the firms within the industry (sometimes limited to the 50 largest firms), where the market shares are expressed as fractions. The result is proportional to the average market share, weighted by market share. The results are to be interpreted as follows:
An HHI < 0.01 indicates a highly competitive industry.
An HHI < 0.15 indicates an unconcentrated industry.
An HHI 0.15 - 0.25 indicates moderate concentration.
An HHI > 0.25 indicates high concentration.
I've tried to analyze the competitiveness of the league for each season, using a measure called Herfindahl Index (Economics students will be aware of this concept). It's a measure of the size of firms in relation to the industry and an indicator of the amount of competition among them. HHI is an economic concept widely applied in competition law, antitrust and also technology management. It is defined as the sum of the squares of the market shares of the firms within the industry (sometimes limited to the 50 largest firms), where the market shares are expressed as fractions. The result is proportional to the average market share, weighted by market share. The results are to be interpreted as follows:
An HHI < 0.01 indicates a highly competitive industry.
An HHI < 0.15 indicates an unconcentrated industry.
An HHI 0.15 - 0.25 indicates moderate concentration.
An HHI > 0.25 indicates high concentration.
To compute IPL HHI, I've used the points garnered by a team during a season and divided by the total points available during that season (comparable to market shares). The IPL HHI is charted out below:
 |
| Data source: ESPNcricinfo.com |
The first thing that jumps out - more the number of teams, more competitive the league. In 2011, when two new teams were introduced, the HHI fell to 0.106. Similarly, in 2012 and 2013 when there were 9 teams in the league, the HHI was 0.121 and 0.125, respectively. For an 8 team format, 0.131 is the lowest HHI, achieved in 2010 and then again in 2016. It's fair to conclude, looking at these figures, that the tournament has managed to stay highly competitive over the last nine years.
Conclusion
The above is simply an ex-post analysis of IPL data over the last nine years. There are multiple ways to analyze the same data and I've tried to take a different approach as compared to the traditional methods. Some of the conclusions may appear to be intuitively obvious. Yet, the humble attempt is to see whether empirical data leads to the same conclusions.
No comments:
Post a Comment