January 18, 2013

Analysing Test dismissals across the ages

A detailed look at various forms of dismissals in Tests, and their frequency over the years

The percentage of bowled dismissals has dropped steadily over the years in Tests © AFP

It has always been my desire to carry out a complete analysis of run-outs in Test matches covering every aspect of this fascinating form of dismissal. Since I have never looked at Test dismissals as a single topic, I decided to take an overall look at dismissals and subsequently go to the next level.

The initial analysis is an in-depth look at dismissals across various periods following the usual classification. The Pre-WW1, Between-wars, 1950s, 1960s and 1970s were automatic selections. On this occasion, I have tweaked the remaining years. Tendulkar, who frequently gives his wicket to debutants, was also the first victim of TV umpires who made their debut in 1992 at Durban. Hence I decided to have 1980-1992 as a period followed by three more: 1993-1999, 2000-2006 and 2007-2012. This ensures decent number of Tests in each period.

Let us begin with the base table of number of dismissals in each period. This analysis covers the period until 31 December 2012. Even though three Tests have since been played, all my table work was done before Test no 2069 was incorporated into the database.

Number of dismissals by period

Dismissal type187719201946196019701980199320002007All Tests
WK Caught3734328599201058184815851946132210343
Fldr Caught18091639233223252647396834924436319925847
Run Outs1791472412322183252923722262232

This table has been shown just for information. There is very little insight to be drawn from this since the number of Tests varies between periods. However the data is available if anyone wants to do some work.

Now the first of two special tables. The first shows the % of dismissals by dismissal type across periods.

Dismissals analysis: % of total dismissals

Dismissal type187719201946196019701980199320002007All Tests
WK Caught8.3010.3413.5015.8917.2218.5818.8618.2117.4516.27%
Fldr Caught40.2439.2236.6640.1643.0939.8941.5641.5042.2240.65%
Run Outs3.983.523.794.013.553.273.473.482.983.51%

Bowled: This type of dismissal has varied drastically across the ages. During the first period about 37% of batsmen were bowled. The figure dropped below 30% soon and remained around this figure until after the WW2. The next three periods exhibit further drops till it stabilized around 16% during the past three periods covering 20 years. The current figure is well below half the initial figure. How does one explain this? Can we infer that the batsmanÂ’'s defence was suspect? Or that the bowler was looking for the most direct form of the dismissals? Could it have been the effect of the uncovered pitches? Or the need to score quickly in 3-day matches? I will let the readers have a field day.

LBWs: This type of dismissal has moved in the other direction. Starting from a very low 6%, the LBW share doubled in the next period. Afterwards the figure remained reasonably steady but took a higher turn in 1980s. It was around that figure until a significant drop during the last period. What does it convey? Possibly that until the change of LBW laws, one could pad away with impunity. Did technique have an effect on these changes? During the past 5 years there is an increase of about 0.4%. This could be due to the partial implementation of DRS. Again let the readers come in with their comments.

WK Catches: Similar to LBWs, at 8% in the first period, followed by a slight increase between the wars and then steady increase until the peak was reached during the 1990s. Then, inexplicably, there has been a drop and now the figure is around 16%.

Fielder Catches: It is a great surprise that, irrespective of the period and whether the batsmen were out bowled or LBW, the fielder catches figure remains either side of 40%, the variation no more than 5% during most of these periods. Taking cognizance of the fact that 40% represents the highest share of all dismissal types, I am as surprised with this revelation as the drastic movement in Bowled and LBW dismissals.

Now we come to two very similar dismissals based on a batsman straying out and failing to reach home.

Stumpings: During the first three periods, the stumpings dismissal share remained at just over 3%. Then it dropped to nearly half and reached a low of 1.3% during 1980-92. It has picked up since and is around 2% these days. Can the current increase be attributed to the more attacking instincts of the batsmen? Third umpire came to the party during 1992. This meant a significant increase during the next three periods. Good spinners have always been around during all these periods.

Run Outs: Surprisingly the Run Outs % has remained steadily at around 4% right through the ages. The value dropped to 3% during 2000-2006, despite the introduction of third umpire. Perhaps the third umpire reversed more decisions in favour of the batsmen.

An analysis of dismissal percentages in Tests
© Anantha Narayanan

Dismissals analysis: Dismissals per match

Dismissal type187719201946196019701980199320002007All Tests
WK Caught2.
Fldr Caught13.511.711.212.513.411.712.712.913.212.5
Run Outs1.

Now for an alternate form of representation. This one relates to the specific numbers instead of percentage values. Though quite similar, they offer different insights.

The total number of dismissals was quite high to start with. 33 dismissals per Test. Then the value dropped and has stabilized at around 30. So these numbers are relevant indeed.

Notice how the average number of bowling dismissals has dropped over the periods and has stabilized now. On the other hand, LBW dismissals kept increasing and are slightly above the Bowled figure now. Wicket keeper catches increased steadily with a slight drop in recent years. Look at fielder catches which have remained either side of 12 catches per Test. Stumpings have dropped significantly and plateaued. The average number of run outs has more or less remained unchanged.

The combined values across all the periods are - 12.5 fielder catches, 6.6 bowled to 1.1 run outs and 0.6 stumpings per match totaling to about 30 dismissals per match including a few others but excluding those fancy dismissals like hitting the ball twice, obstructing the fielder, hit wicket, handling the ball etc.

An analysis of dismissals per match in Tests
© Anantha Narayanan

Here we are to where we started with. A complete look at run-outs. I wanted to look at how run outs have panned over the years: across periods, by countries, against countries et al. Nothing is gained by looking across 137 years and saying that Australia has lost 1.7% due to run-outs or has effected 1.9% of dismissals in run-outs. So the real need is to develop a four dimensional matrix: by dismissal type, by country/for country and by period. Then present the tables in an easy-to read and understandable form.

I plan to do that next. But why stop at run-outs? Every form of dismissal will throw some light on this fascinating aspect of the game. In fact I have a sneaking suspicion that this follow-up analysis could be very very interesting. We may be able to relate the numbers to famous names such as Harvey, Bland, Harper, Rhodes, Wasim, Waqar, McGrath, Alderman et al.

Anantha Narayanan has written for ESPNcricinfo and CastrolCricket and worked with a number of companies on their cricket performance ratings-related systems