Tuesday, December 30, 2014

Cricket - Batting averages can lie

Continuing on Rohit Sharma…


His test record looks good with batting average of 40.71. Virat Kohli crossed 40 average mark only after 30th test match. 


Rohit Sharma has two not outs in his career 111 n.o and 31 n.o. If those two innings are excluded then his average falls down to 30.57. Now by considering those two innings and ignoring the not outs his average for 16 innings is 35.63. Even that seems to be quiet good for a batsman who has played 9 test matches only.

Further, analyzing his 16 innings statistically...


Mean - 36 is average number of runs he has scored in each innings. That means to say we can expect him to score 36 runs every innings.
Median - He has scores less than 22 in 50% (8 innings) of the innings he has played.
Mode - He has got out on 6 runs 3 times.
Standard deviation - He has scored runs anywhere between 0 to 82 runs in most of the innings.


This graph clearly suggest his two innings are the outliers and incidentally these two innings he has batted in India.

Interestingly these two innings account to 49.5% of the runs he has scored.

If we ignore these two innings that will constitute his overseas record.


His graph here is more consistent.




His average (mean) value dips to 20 runs per innings.

While his median value doesn't have much of an impact it is 20, i.e., he has 50% (7 innings) scores of below 20 runs per innings he has played. That is why median gives true indication than average which can go up and down based on one big innings or even more fluctuate when it is a big innings not out.

As the outliers are removed, his standard deviation also improves 20 runs. That means most of his scores are between 0 to 40 runs with mean .

The above analysis clearly shows that only average doesn't qualify to be a right measure of a batsman.

No comments:

Post a Comment