..

What the stats say: is Steve Smith the second-best Australian cricket batsman ever?

• Written by Christopher Drovandi, Senior Lecturer in Statistics, Queensland University of Technology

The Ashes may be over once again, but one of the biggest talking points during the cricket test series between Australia and England was the continued meteoric rise of Australian captain Steve Smith as a test batsman.

There was much speculation as to whether Smith is the best Australian test batsman – bar Donald Bradman – ever to have played the game.

“The Don” Bradman is widely regarded as the greatest Australian cricketer – and was voted the greatest cricketer of the 20th century – with an unrivalled Test batting average of 99.94.

Don Bradman is almost run out at 101 in his innings of 169 in the Fifth Test at the Melbourne Cricket Ground, 1936/37. Flickr/National Library of Australia

So what is the ranking of Australian test batsmen since Australia’s first ever test match in 1877?

Read more: Video explainer: How cricket captains make good decisions

Let’s crunch the numbers

We need to use some statistical modelling to answer the question. The dataset includes only batsmen who played a substantial number of tests for Australia, defined here by batting at least 30 innings.

“The Don” is removed from the dataset to allow the other players to be ranked without being affected by his exceptional performance.

Steve Smith batting in the Third Test match between Australia and England, at the WACA in Perth. AAP/Dean Lewins

All players who were considered predominantly bowlers have been removed from the dataset. All-rounders and wicketkeepers were retained, as these types of players might be expected to make substantial contributions with the bat. That leaves 108 players in the dataset.

A statistic commonly used to compare batsmen is their average: the total number of runs scored in their career divided by the number of times they have been dismissed.

If a player is not dismissed in an innings – the player is “not out” for whatever reason – then their score would effectively carry over to the next innings, at least from the point of view of the way the average is calculated.

For example, if a player scores 5, 57, 34 (not out) and 3, then this is the same as the player scoring 5, 57 and 37, giving an average of 33.

Steve Smith has been compared to the great Don Bradman.

A changed game over the years

Just simply comparing averages may give misleading results, however. Cricket has evolved greatly over more than a century due to factors such as pitch conditions, professionalism, coaching, popularity and technology, to name a few.

Each circle represents the batting average of an Australian test batsmen. The solid line is the mean batting average estimated from the statistical model applied to the data. Christopher Drovandi, Author provided

The graph above shows that test batsmen found run-scoring relatively difficult in the late 1800s and early 1900s compared with more recent times. So we need the statistical modelling to account for the decade in which the players played most of their test cricket.

The solid line in the graph shows that the statistical model can detect the general upwards trend of test batting averages. We can then adjust the results to account for this trend.

Steve Smith batting in the Third Test match between Australia and England, at the WACA in Perth. AAP/Dean Lewins

This gives us the top 15 Australian test batsmen of all time (bar Bradman) as determined by the statistical model. Note that only players who have played at least 25 matches are considered in the final ranking.

In the graph above, the black circle on the plot represents the relative score for each player. This is the score the statistical model predicts for a player’s batting average relative to the model predicted average from all players (where the model adjusts for the decade the player played).

For example, a relative score of 10 suggests that the player scores on average 10 more runs than a typical batsman in the same era. The horizontal grey line for each player shows the level of uncertainty in that figure, but we’ll come to that later.

Read more: Video explainer: Batting expertise and decision-making in cricket

The second-best batsman is…?

We see from the circle plot that Steve Smith has indeed received the highest relative score. Greg Chappell stands out as the next best.

The statistical model naturally shrinks the scores related to players who have played relatively few matches and/or scored inconsistently.

This is actually an advantageous feature of the model. A player who has managed to maintain a high average and play consistently over a long period of time should be rewarded.

Steve Smith (left) raises his bat as he celebrates scoring a century at the Third Test match between Australia and England, at the WACA in Perth. AAP/Dean Lewins

And now the uncertainty

The statistical modelling approach not only provides an estimated relative score, it can also produce an interval that quantifies the uncertainty in the estimated score.

This is what those horizontal grey lines show in the graph above. They show there is a 90% chance that the player’s relative score is in that interval.

Steve Smith loses his grip on the bat during play at the Fifth Ashes Test between Australia and England, at the Sydney Cricket Ground. AAP/Paul Miller

Although Smith has been allocated the highest relative score, there is substantial overlap in the score intervals of other players. This suggests that it is difficult to determine, from a statistical point of view, that Smith is a significantly better batter than the other players shown in the graph.

But the intervals do demonstrate that the players on this list are significantly better than a typical player from the corresponding era, as the intervals do not include 0.

The main reason for the large uncertainty is that a batter’s scores in individual innings fluctuate wildly. A batsman may get out for 0 runs in one innings and then score more than 100 runs in the next.

It is therefore difficult to home in on a batsman’s real average unless they were able to bat an extremely large number of times.

Read more: Video explainer: Bowling strategies and decision-making in cricket

The other reason for the wide intervals is that because a game of test cricket is so long (typically four or five days), each player is not able to bat many times, even if they have a long career.

A more rigorous statistical analysis would attempt to model the individual run scores of all the players directly rather modelling the average, which is simply a summary of a player’s batting career.

But the analysis has justified why many commentators, pundits and punters consider Smith to be the best Australian test batsman after Bradman – so far. Smith is still very much in the game so it will be interesting to see if he can maintain or improve on his recent performances.

Steve Smith raises his bat to applause as he leaves the field at the First Test match between Australia and England, at the Gabba in Brisbane. AAP/Dave Hunt

Authors: Christopher Drovandi, Senior Lecturer in Statistics, Queensland University of Technology

Politics

Central and North Queenslanders overwhelmingly back coal sector

More than 80 per cent of young people in Central and North Queensland believe the coal sector and the jobs it creates are critical to the state’s economy, particularly regional Queensland.   Mi...

Senator Canavan

AWU National Secretary Daniel Walton on steel industry

Ambitious expansion plans make mockery of those who wrote off Australian steel industry GFG Alliance's planned rejuvenation of the Whyalla steelworks, and the Australian steel industry more generall...

Media Release

George Neophytou supports roads funding and more

Independent candidate for Gippsland East George Neophytou has pledged to support completion of duplication of the Princes Highway between Traralgon and Sale and to upgrading of the Sale alternative ...

George Neophytou Media Release

10 Excellent Business Ideas for Young Entrepreneurs in 2019-2020

Top 10 Low-Budget Business Ideas for Students in 2019 Whoever said that you should wait for 10 or even more years after your graduation to launch your own business? 69% of US citizens start their ...

Elsa Avendano

The 1-Page Marketing Plan

SKIP BUSINESS SCHOOL. READ THIS NEW BOOK INSTEAD. Named as one of the “10 best marketing books for small business owners” by The Huffington Post.   MEDIA RELEASE: MELBOURNE, DECEMBER 13, 2018 —...

Claire Marshall

How to Empower Your Employees in 2019

Employee empowerment is the focus of many employers in this present day. The reality is that although you may hire competent and qualified professionals, you still need to invest in building them up...

News Company

Travel

Road Trip On Your Mind? Keep These Aspects In Mind As Well

Take a Road Trip! The great road trip. Movies are made of this type of freedom. The roads are calling and you are feeling the need for a road trip. Join millions of other people who enjoy the great o...

News Company

Best Short Term Accommodation in Brisbane

Brisbane, the capital city of Queensland in Australia is known for its museum and Science center. With a lot of creative spaces and gardens on the riverside, it is a magnificent and popular destinat...

News Company

Best Short-term Accommodations in Sydney

If you are looking to spend a night or two in the sprawling city of Sydney, you might be in need of a lodging place or hotel to stay. We would like to give you here a list of the best short term acc...

News Company