Sunday, March 3, 2013

Progression of Test Batting Averages of Various Positions

Are you curious to take a look at the batting averages by position (for Test Matches), over the entire history of Test Cricket starting from 1877, to the present? The batting average is calculated the usual way (runs scored minus not outs, divided by innings batted). Below is the graph depicting the results.

No real surprises here, with the average reaching a peak at around No. 4, and taking a dip from there. Now, why would the openers have a lower average than No. 3 and No. 4, or even No. 5, for that matter? It's probably because the openers have the responsibility of seeing off the new ball, and have to face the brunt of the hot pace thrown by the fresh fast bowlers in the beginning of an innings. By the time, No. 4 comes in to bat, the bowlers are no longer fresh, the ball is no longer new, and the conditions are slightly better to bat. This is not to take anything away from the middle order batsman. But let's appreciate what the openers have to go through at the start of an innings.

From these figures it looks like it's most advantageous for a team to have its best batsman play at No. 3 or 4, provided of course that the openers are good enough to safely see off the new ball.

Let's now see if the pattern holds true for all the different countries, over the years starting from 1877. With minor exceptions, the pattern remains very similar. The chart is shown below.

From the above, it looks like England and Australia have had some of the best openers over the years. Also, England and Australia seem to have very good No. 11 batsmen, compared to the other countries!

We will now do the same exercise for different eras.The results are show below in the chart.
Once again, the pattern of higher middle-order batsman having a higher average is established, especially over the later years. What is also very interesting and can be observed above is that the difference between the best averages and the lowest averages seems to be increasing as the years progress. While the averages for top-order batsman have increased, the averages for lower-order batsmen have remained the same over the years. What may be the reason for this?

Maybe the players have become more specialized in their roles these days, with the batsmen becoming really good, and bowlers preferring to specialize in their own craft?

Another interesting thing to note is that the batting averages got really high in the 1925-1949 era, and even surpassed the averages over the next 50 years. It may be remembered that Don Bradman played in this era. But could one batsman make so much difference? Further research needs to be done on this point.

Stay tuned for a Part 2, which will dig deeper into each country and how they fared in various eras.

Note : The above statistics cover all Test Matches until end of February 2013. 

Copyright 2013, Deepak Vishwanathan


  1. Hi Deepak,

    Where did you get this data from? I'd like to look at it myself.

    Also, why do you think positions 1 & 2 are different? That doesn't make much sense to me.


    1. Hi Mark,

      Thanks for stopping by. Just saw your comment today.

      I got this stuff from ESPN Cricinfo statsguru, a gold mine for cricket statistics. It is still not super-comprehensive but if you dig deep enough, you can usually find what you want.

      For example : for batting position 3, click this -;batting_positionmin2=3;batting_positionval2=batting_position;class=1;filter=advanced;groupby=overall;orderby=runs;template=results;type=batting

      .. and so on.

      Good point about difference between pos. 1 and 2. The only thing I can think of is - maybe the better of the two openers usually likes to take the first go ?

      - Deepak