A few years ago I used the Boston Consulting Group (BCG) charts to get insights into how ODI batsmen fared along two contrasting measures: the batting strike rate (SR) and Runs per Innings (RpI). In this article, I will refine that comparison to a more meaningful one.
It is also a good time* for this analysis since the ODI arena is stuffed with many all-time great batsmen. What was considered outstanding a few years back is now par.
Long-time readers will remember that I have done many presentations of contrasting measures using the very powerful BCG charts. For newer readers, let me describe the BCG charts briefly.
They are the invaluable contributions by Bruce Henderson, founder of BCG, to the complex area of marketing analysis. He created these charts in 1970 to present how businesses needed to classify their products between two contrasting marketing measures: market share and market growth. Over the years, the BCG charts have become valuable tools for any market-driven company.
In a nutshell, a BCG chart is drawn by dividing a space into four quadrants, placing entities into these four quadrants and naming the quadrants appropriately. Henderson named the first chart's quadrants as Stars, Cash Cows, Dogs and Question Marks. Each of us has their own naming conventions. When the chart is completed, an appropriate mix could be developed by a judicious mix of products from the key quadrants - in our case, the products are batsmen.
An ODI batsman has two distinct metrics applied to him across his career: strike rate and Runs per Innings. The strike rate is uncomplicated and stands tall as a clear and pure measure. However, the basic Runs per Innings metric is nearly as flawed as the batting average. It's as unfair to late-order batsmen as the average is to top-order batsmen.
So I developed a new measure, which will henceforth be called the Runs per Weighted Innings (earlier called Runs per Adjusted Innings). The idea is simple. The problem with the batting average is that for two contrasting scores of, say, 175 not out and 2 not out, both innings are treated as non-existent. The first is a three-hour masterclass and the latter is a three-minute effort. I differentiate between these two in the following manner: the runs scored (numerator) stays the same as when deriving the average, but the denominator, the innings count, is a weighted measure.
- For all dismissals, the innings count is increased by one.
- For all not-out innings where the score is above the overall career RpI (say, 41.3), the innings (175 not out, 85 not out, 41 not out et al) the innings count is increased by one, which is fair because the batsman has already crossed his own mean score and thus completed the innings, so to say.
- For all not-out innings below the overall career RpI, the innings count is increased proportionately. Continuing the example of RpI of 41.3, the few examples given will illustrate this. The value in brackets is the innings count for that particular innings. 40 not out (0.97), 27 not out (0.65), 5 not out (0.12) etc. The most positive aspect of this computation is that there is no "golden number" usage. The batsman's own RpI is used.
- The Runs per Weighted Innings (RpWI) is determined by dividing the total runs scored by the weighted innings value. The RpWI will be a figure between the batting average and RpI.
My first inclination was to do the BCG charts using strike rate and RpWI. However, I realised quickly that the RpWI is itself a component of two elements: the strike rate and Balls per Weighted Innings. As such, the two measures are not independent of each other. So for my chart to be effective, I had to draw it between strike rate and Balls per Weighted Innings (and not RpWI). As long as I maintain the same consistent denominator (Weighted Innings) while determining the BpWI, the overall integrity will be maintained.
In summary, the strike rate is a measure of how effectively the batsman strikes. The higher this value, the faster the batsman scores. Au contraire, the BpWI is a measure of how many balls he plays, a completely independent measure. A good team will have a mix of these two contrasting skills, and the use of the BCG charts will help achieve this objective.
It is time to peruse the tables listing the top batsmen by these two measures. The qualification criterion is 2000 ODI runs.
Let us take Glenn Maxwell. His strike rate is an astounding 123. But he plays, on average, only 23 balls. And so in general do the other batsmen who have strike rate values exceeding 100. AB de Villiers is an exception. He not only strikes at 101, he lasts 44 balls. That makes his RpWI an incredible 45.
David Warner, Quinton de Kock, Shikhar Dhawan and Viv Richards are quite close to de Villiers. However, Virat Kohli is an even bigger star. He strikes at 93 and stays for 53 balls. That makes his weighted run contribution a huge 49. I hate the word "super", but I have to say Kohli is a superstar. This table lists the batsmen who strike at 90 and more. Barring Viv Richards, many of the greats of the previous generation do not clock in. But that is today's game.
Now we come to the other end of the spectrum: the BpWI measure. This table lists all the batsmen who have weighted balls per innings exceeding 50. Geoff Marsh lasted more than 67 balls per weighted innings, but he scored at the rather pedestrian rate of 56. The net result was still an impressive 38 runs per innings. Gordon Greenidge lasted a few balls fewer but scored a little bit faster and contributed over 40 runs per innings. Desmond Haynes was not too far off his batting partner.
Two facts to be noted: Kohli is the only batsman on both tables. That is the sign of his true greatness. He is the only player to have lasted over 50 balls per weighted innings at a strike rate exceeding 90. Have a look also at Hashim Amla. He strikes at 87 and lasts as many balls as Kohli, making his contribution an astounding 48. A most underrated ODI player indeed.
It can be seen that for Marsh, Yasir Hameed and Mark Taylor, their Innings and Weighted Innings values are the same. This indicates that all their unbeaten innings are above their RpI values. This also substantiates the fairness of the Weighted Innings concept. Otherwise these batsmen, especially Marsh, would have got undue advantage.
The BpWI value is placed on the X axis. The range is 10 to 70 and the Y axis is drawn exactly at the midpoint, i.e. at a BpWI value of 40. The placement of the X axis is a little tricky. The exact midpoint is a SR of 90. But that would set the bar too high and the charts would be bottom-heavy. Hence, I have located the X-axis line at a SR value of 85.0 to get a decent collection of batsmen on either side. In any case, a strike rate of 85 is above average and only 49 batsmen (out of the 211 who have scored 2000 runs or more) have values exceeding 85. It also represents a reasonable par score of 255.
A typical BCG chart can accommodate around 25 entries. I can stretch this to about 35 if I don't display the descriptions (batsman names) on the main graph and display the number-name combination as a separate graph below. The players are represented by a number and the numbers are displayed centred on the exact player coordinates. These coordinates determine the quadrant in which the player falls.
There are two sets of BCG charts. The first shows the top 35 run scorers and the second the next 25 run scorers and ten other selected batsmen, to bring out the features of the charts.
The top-right quadrant consists of batsmen who have played over 40 balls per innings at strike rates exceeding 85.0. They are rightfully called High-Fliers. Six batsmen qualify. The leaders of this group of masters are Kohli (strike rate 92.1 and 53 balls per innings), de Villiers (101.1 and 44) and Amla (88.6 and 54). They are on an outside arc.
Is there a batsman here who does not deserve this lofty status? Chris Gayle certainly deserves his place, with numbers like 86.1 and 41. What about MS Dhoni? Well, his numbers, 88.4 and 44, answer this query. The feature of this group is that barring one batsman, the others are from the modern era. Tendulkar (86.2 and 48) represents the middle 15 years. I could have had floating, era-specific cut-off values, but that would have complicated matters. In any case, the greats of the previous eras are all represented here.
Moving clockwise, the bottom-right quadrant consists of batsmen who have played over 40 balls per innings at strike rates below 85. They are rightfully called Bankers: a dependable lot of players indeed. Nineteen batsmen qualify. Some of them are just below the qualifying SR mark of 85, but they are all well ahead of the 40-balls-per-innings mark.
This quadrant sees a mixture of modern greats, like Kumar Sangakkara (78.9 and 49), Brian Lara (79.5 and 46), Rahul Dravid (71.2 and 50), Ricky Ponting (80.4 and 48) and Sourav Ganguly (73.7 and 52), and quality players from previous generations, like Mark Waugh (76.9 and 48), Haynes (63.1 and 58), Javed Miandad (67.0 and 52) and Zaheer Abbas (84.8 and 51). Any great team has to have at least two of these bankers.
The bottom-left quadrant consists of batsmen who have lasted fewer than 40 balls per innings at a SR below 85. They are appropriately called Also Bats. Only four batsmen qualify. However, the presence of Mahela Jayawardene (79.0 and 39), Aravinda De Silva (81.1 and 39) and Steve Waugh (75.9 and 37) here is a bit of a surprise. But let us accept it. These stalwarts should have crossed the 40-balls-per-innings mark.
The top-left quadrant consists of batsmen who have lasted below 40 balls per innings at a SR exceeding 85. They are the Meteors, the crowd-pleasers, normally batting in the middle order. Because these are the top run scorers, only six batsmen qualify. They would certainly win quite a few matches but will also have early exits in many others. The major batsmen in this quadrant are Sanath Jayasuriya (91.2 and 34), Virender Sehwag (104.3 and 33), Adam Gilchrist (96.9 and 36), and Shahid Afridi (117.0 and 19). Readers should note that the high SR values are what keep them in this intriguing quadrant. The closest to miss this quadrant was Tillakaratne Dilshan (86.2 and 39.9). And the chart clearly depicts this.
Because these batsmen have played in fewer matches, it is easier for them to be High-Fliers. This is proved by the fact that no fewer than 12 batsmen qualify to the elite High-Fliers quadrant. Prominent here are Richards (90.2 and 46), David Warner (96.6 and 44), de Kock (94.0 and 46), Dhawan (93.4 and 46), Kevin Pietersen (86.6 and 41), Mike Hussey (87.2 and 42) and Joe Root (86.9 and 52).
The Bankers quadrant has only 11 entries. Geoff Marsh (55.9 and 68) is the most interesting of these. Dean Jones (72.6 and 53), Sunil Gavaskar (62.3 and 50), Michael Bevan (74.2 and 51) and Abbas (84.8 and 51.3) form an interesting quartet. Abbas just misses out on the top quadrant.
The most prominent of the six batsmen who fit the Also Bat category are: Younis Khan (75.3 and 39) and Allan Border (71.4 and 37).
Shane Watson (90.4 and 39 - just misses out on the High-Fliers' quadrant) leads the attacking batsmen group. Maxwell (123.1 and 24) and Jos Buttler (118.0 and 26) are two special selections, just to bring out the BCG concept. There are only five batsmen in this quadrant.
Very few teams have the luxury of having a number of High-Fliers in the side at the same time. India and South Africa currently have four such batsmen (Kohli, Rohit Sharma, Dhawan and Dhoni for India, and Amla, de Villiers, Faf du Plessis and de Kock for South Africa). Whether teams are blessed with such high-quality batsmen or not, they can fill the other positions with a judicious mix of Bankers and High-Fliers with shorter life spans.
The BCG chart for batsmen offers very good insights into the careers of the concerned batsmen. It is not just a matter of moving into the important groups (quadrants). For active players, it is more a question of moving right and up on a consistent basis.
However, it is clear that we also need a single measure with which to evaluate batsmen. That is already with us in the form of Runs per Weighted Innings (RpWI). Because we used the same dividing factor, weighted innings, it is clear that the RpWI is determined by multiplying the SR by BpWI. It is the most effective and useful batting measure as it encompasses both the key measures. Let me plot these on a single graph.
This is a self-explanatory chart. Kohli stands supreme with an RpWI value of nearly 50. It is slightly surprising that Amla beats de Villiers to second spot, with 47.6, clearing de Villiers' 45 by a substantial margin. Note the wide separation between Kohli and Amla, and Amla and de Villiers. Then the bunching of players starts. Dhawan and Kane Williamson are tied at the same value.
Root is in the top ten along with Dhawan, Williamson, Warner and de Kock. Abbas is the best among the older players. Three of the legends of the game - Tendulkar, Richards and Greenidge - are just outside the top ten. These 13 batsmen have RpWI values exceeding 40. Then come a slew of top ODI batsmen across the ages. The last two positions of the top 30 are taken by two of my favourite players: Mike Hussey and Lara.
For the bowlers, the contrasting measures are the Bowling Strike Rate and Runs per Over. These are two straightforward measures, not requiring any tweaking, and they lend themselves to BCG charting comfortably. I will tackle this sometime in the coming months, maybe as an article sandwiched between the Bowling and Batting Performance Analysis mega posts.
*I intended to have my two-part update (after 15 years) of the Wisden 100 Bowling Performance tables run in May and June. But I feel they won't get the attention they need during the IPL, so I'll do them in June and July, when the Test season will be in full swing.