Note about Batting Averages
Ever since averages were introduced as a method for comparing the relative performance
of players, statisticians have had the problem of the never out batsman.
The Batting average is calculated by dividing the number of runs scored by the amount of times
the batsman is out. This works well if the batsman is out at least once. If he isn't, then
you face the prospect of dividing the runs by zero. As any good mathematician will tell you, the average
is then infinity, irrespective of the number of runs scored. This of course is unsatisfactory as
a method of comparing performance. For example a player who gets a streaky 1 not out will have
a better average than a player who gets out after scoring a century.
General accepted practice amongst cricket statisticians to circumvent this problem is to
say that the batsman has 'no average' if he is never out. This partially solves the problem as the
player in the above example would be placed at the bottom of the list of players rather than the top.
I however, prefer to go one step further. If a player scores a not out century he would be placed
below the player who is out after scoring a single. To my mind this is a greater inaccuracy than the
original example. The 'accepted' system penalises a batsman who is good enough to protect his
wicket and favours those that cannot hold out against the opposition bowlers. In my view that
has got to be wrong!
The way I overcome this is to divide the runs scored by the number of times the batsman is out
or one, if the batsman is never out.
This then means that every batsman will have an equitable average with which he can
compare (and be compared) with other batsmen.
Simon Smith
Statistician
Bar Hill Cricket Club