A Mean Average

Someone wise once said, “There are three kinds of lies: lies, damned lies and statistics.” Cricket has bucketfuls of statistics but how many of them are misleading? Actuaries Sanchit Maini and Sumit Narayanan recently addressed (here) what they considered as a flaw in calculating the kingpin of cricketing stats – the batting average.

Apparently the fault resides with the not-out innings. Their runs are counted into the total but the innings themselves are not added into the divisor. This seems fair enough since we cannot know how many runs the batter would have scored if the innings had continued. But this method can lead to some freakish results. For example, perennial rabbit Bill Johnston topped Australia’s tour averages in 1952 with 102 because he was dismissed just once in 17 visits to the crease. Another, more controversial, example could be Michael Bevan’s outstanding one-day average of 53.6.

Maini and Narayanan remove this bias by using a formula that sometimes counts a not-out innings into the divisor, depending on how many balls were faced in the innings. The results are that everyone’s average decreases but lower order batsmen are much more greatly affected than their higher order teammates because they are more likely to be non-out at the end of the innings. So, Virender Sehwag’s one day average drops only from 32.4 to 31.2 because he is unlikely to last a full 50 overs but Bevan’s revised average is a whopping 14.9 runs less since he (as every Australian opponent experienced to their cost) was so often still there at the end of the innings.

Who tops the list Highest Revised ODI Batting Averages? None other than Sir Viv Richards, followed by Sachin Tendulkar, then Ricky Ponting and then Michael Bevan. My only query is that the formula seems to punish players for scoring high not out scores, but then I’m a natural skeptic; I heard that 88.2% of all statistics are made up on the spot.