Executive Summary
By all rights, Australia should be feeling on top of the world heading into the WT20 Cup in India next week. Recently reinstated as the top ranked test side, Australia are also ranked number one in the world in ODIs and are the reigning ODI world cup holders. However the reality is that Australia have struggled to transfer their dominance to the shortest form of the game. Currently ranked 8th in the world in T20 cricket, it would be no surprise if Australia once again fail to advance to the knock out stages of this year’s tournament.
Why is such a proud and successful cricketing nation languishing in a format that has been growing exponentially in popularity over the last few years through the success of the domestic Big Bash League? It would be hard to blame the players themselves for failing to adapt – Australian players have constantly been among the most sought after imports in the IPL.
In this article we suggest that team selection may be a major factor contributing to Australia’s sub-par performances.
We analyze the composition of the Australian squad for the 2016 WT20 using Contribution Score (CS) analysis. CS analysis allows us to assign a single number to measure the performance of a player that takes into account the match state at which the player makes his contribution, such as batting first or second, overs remaining, wickets lost and runs scored. CS for batsmen and bowlers are calculated on the same scale so that performances by batsmen and bowlers can be compared directly to see who has had the biggest impact on the result of a match, series etc. Refer to the Appendix for a more detailed explanation of the Contribution Score.
The CS analysis conducted on the Australian WT20 squad as well as on the squads of their known opponents in the group stages suggests that Australia will lose to India and possibly New Zealand in the group stages of the competition. Australia should overcome Pakistan and the qualifier to finish on two wins from four matches. Whether that will be enough to progress in the tournament remains to be seen.
By using CS as the basis of analyzing the strengths and weakness of individual batsmen against different styles of bowlers, and of individual bowlers against different types of batsmen, some clear gaps appear in the Australian WT20 squad. A heat map is a good visual tool to highlight the strengths and weakness of a country’s T20 cricket current squad.
How to Interpret the Heat Map?
The heat map shows how each player in a squad has performed over the last two years in T20 cricket, taking into account their performances (if any) in Asia. Firstly, the heat map summarizes each player’s bowling performances against:
- Left Handed Batsmen (v LHB)
- Right Handed Batsmen (v RHB)
- Left Handed Batsmen in the critical last five overs of the inning (L5 v LHB) and
- Right Handed Batsmen in the last five overs of the inning (L5 v RHB)
Secondly, the heat map shows each player’s batting performance against:
- Left arm fast bowling (v LAB F)
- Left arm fast medium or medium fast bowling (v LAB FM)
- Left arm medium bowling (v LAB M)
- Left arm finger spin, or orthodox, bowling (v LAB FS)
- Left arm wrist spin, or googly, bowling (v LAB WS)
- Right arm fast bowling (v RAB F)
- Right arm fast medium or medium fast bowling (v RAB FM)
- Right arm medium bowling (v RAB M)
- Right arm finger spin, or off spin, bowling (v RAB FS)
- Right arm wrist spin, or leg break, bowling (v RAB WS)
The shading in the heat map indicates how each player has performed, where:
A good heat map for a squad would have lots of green shading and few areas of red. Furthermore a column with lots of red and grey shading with little green shading would suggest the squad has an exploitable weakness against the type of bowler or batsman that column describes.
The heat map below highlights the strengths and weaknesses of Australia’s 2016 WT20 squad.
The heat map for the Australian squad highlights the following points:
Bowling
- Neither FS (Maxwell nor Agar) is a penetrative bowler. Furthermore they are both weaker against LHB and neutral against RHB. Given that Zampa is also weaker against LHB, there is no clear spinning option to counter LHB.
- Tye is the only FM bowler with positive CS against RHB
- No all-rounder has a positive bowling CS against either LHB or RHB
- All all-rounders except Faulkner have a negative CS against LHB. Faulkner’s CS against LHB is neutral
- There is no all-rounder or bowler in the squad who has a positive CS against RHB in the last 5 overs
Batting
- The top five batsmen have few weaknesses against the various bowling options. LAB M is the only style to which only Watson and Maxwell have negative CS. LAB M is not a common bowling style and this should pose little impact on the Australian line-up
- Australia will likely field two LHB and five RHB in the top seven. Three LHB and four RHB in the top seven would be a more balanced line up. Furthermore Steven Smith has negative CS against both LAB FS and RAB FS which is likely to be exploited in Asian conditions in the WT20. A LHB with no major weaknesses against any bowling type should be considered as a possible replacement for Smith
- In addition to having negative bowling CSs, Marsh and Faulkner have weaknesses with the bat that can be exploited
- Is Nevill’s batting CS strong enough for him to be the Australian keeper in the WT20?
The issues raised above were analyzed point by point, again using CS as the basis for the analysis. The analysis suggested that an alternative Australian squad could have been selected which would have closed many of the gaps in the original squad. The tables below show the squad changes that could have been made and the resulting revised heat map.
The rationale for the changes has already introduced and is described in detail in the full article. It is difficult to omit the captain, Steve Smith, from the squad however the benefit of having an additional LHB in the top six together with Smith’s historical weakness in this form of the game against FS makes Sean Marsh a better candidate. The recommendation to remove Hazelwood is mainly due to the fact that McKay’s form in the BBL 2015/6 was too irresistible to ignore. Add to that Hazelwood’s limited participation in T20 recently and his heavy workload during the Australian summer makes McKay’s inclusion all the more reasonable.
Alternative Squad Heat Map
Is the Alternative Squad better than the Original Squad?
Visually the heat map of the alternative squad has an overall greater concentration of green shading. Furthermore, each column has a dominance of green shading which indicates no exploitable weaknesses against any one particular batting or bowling type.
CS analysis can be used to quantify how many runs better the alternative squad is compared to the original squad. The analysis also shows that with a different squad and with a flexible approach to match day team selection, Australia’s performance could improve by 20-30 runs per match (combined batting and bowling impact) depending upon the make-up of the opposition team. The analysis suggests this would be sufficient to defeat New Zealand and bring Australia close to beating India and therefore guarantee them a place in the knock out stages.
This article is not an attempt to suggest that the winners of cricket matches can be predicted with a computer. However the science of team selection and game day strategy is common practice in many sports today. Baseball managers are acutely aware of which pitchers to use and what fields to set against different batters. Again these strategies do not guarantee success, as part of the beauty of sport is that anything is possible. However over the long run, trends emerge that if followed can increase the likelihood of success.
Analytics is already widely used in T20 cricket in India to drive team selection and game day strategy. India is currently ranked number 1 in the world in T20 cricket. What lessons can Australia learn from India?
APPENDIX
An Introduction to Contribution Score
There are a few different analytical approaches to assessing player performance in cricket. The difficulty in assessing and comparing player performance is that the match state is a major component of what drives a good or a bad performance. Suppose a batsman scores 50 runs off 50 balls. This is a reasonable achievement however given that the average strike rate in T20 cricket is 7.5 runs per over, then at face value, 50 runs off 50 balls is a below average performance. Now take two different game situations in which this score might be made:
- An opening batsman in a team batting first scores 50 off 50 balls and is the first wicket to fall, in the 18th over with the score at 115
- An opening batsman finds his team at 6-20 chasing 80 for victory. He scores 50 off 50 balls and although he is dismissed in the last over, his team hangs on for a 2 wicket win
Which inning is more important for the team? Clearly the second innings was more important and should be rated higher than the first. Most available metrics in cricket will not recognize the difference in value of these two innings. The Contribution Score (CS) however identifies every match state in every game, and assesses the players’ performance accordingly. As a result, over the course of an inning, a match, a series, a year or a career, we can assess how a well a player has performed given the situations he or she has been in.
CSs are measured in terms of runs. A CS of 0 for a ball means the player performed in line with average. A CS of 1 for a ball means they contributed 1 more run than expected while a CS of -1 means they contributed one less run than average. These scores are calculated and measured over any period you require and gives you the ability to rank and compare batsmen across different positions etc.
The same approach is applied for bowling. A score of +1 for a batsman automatically means a score of -1 is given to the bowler. As CS are measured on the same scale, a batting CS of +10 means exactly the same as a bowling CS of +10 in terms of the players contribution to the team’s performance.
Because Contribution Scores are measured in terms of runs they are easy to understand. For example a batsman who averages a CS of +10 effectively contributes +10 runs above average to the team performance. If the average score for a team batting first in T20 cricket is 151, then this players contribution is the increase the team’s average total by +10 runs to 161. Similarly, a bowler who averages -5 runs a game effectively gives away 5 runs on average to the opposition per game. If the average score for a team batting first in T20 cricket is 151, then this players contribution is the increase the opposition team’s average total to 156.
Any contribution score greater than zero should be interpreted as above average performance. The higher the CS, the more above average.
Examples Using Contribution Score from the 2015/6 Big Bash League
BBL 2015 Best and Worst Five Batsman (minimum 50 balls faced)
The players are ranked in terms of the Average Contribution Score per innings. Khawaja’s average of a CS of +34.9 means that he contributed 34.9 runs above the average players’ performance, given the situations in which he batted, per inning.
Lynn, the 2015/6 BBL Player of the Series, is ranked third in this list with an average CS of +13.6 runs per inning. Does that mean Khawaja should have been awarded player of the series instead of Lynn? Lynn played eight inns while Khawaja only played four. Given that CS can be negative or positive, had Khawaja also played eight innings and had he failed in some of these, his average CS would have dropped. The player of the series award, quite rightly, should be given to players who have proved themselves across the whole series, playing the full – or close to the full – quota of games.
Two of the three Australians in the top five list are in the WT20 squad.
BBL 2015 Best and Worst Five Bowlers (minimum 50 balls bowled)
The CSs in the chart above are shown from the bowler’s perspective. A positive CS indicates that the bowler is better than average, while a negative CS indicates that the bowler is worse than average. Nathan Lyon’s CS of +11.0 runs per 24 balls bowled (4 overs) was the best in the 2015/6 BBL. Effectively, Lyon’s bowling contributions resulted in the opposition scoring 11 less runs per 4 over spell bowled than they would have done had an average bowler been used.
Of the three Australians among the top 5 bowlers, none are in the WT20 squad.
BBL 2015 Best and Worst Five Players (minimum 50 balls batted and/or bowled)
CSs can also be measured at a player level. Since CS for batsmen and bowlers are measured on the same scale, a batsman’s CS can be directly compared against a bowler’s CS and an individual player’s batting and bowling CS can be added together to arrive at his player level contribution to the team. As a player, Khawaja had by far the highest impact on his team’s performance across the four games he played in the BBL. Among players who played at least 8 matches, Lynn had the highest CS per match followed by Clint McKay. James Faulkner (who is in Australia’s WT20 squad) had the fourth lowest CS per match in this year’s BBL.
Best Bowler in Overs 17-20 (Minimum 20 balls)
The CS in the chart above is expressed as the CS per 12 balls, or 2 overs. Rashid was by far the most effective closing bowler in this year’s BBL, saving his team 9.5 runs against the average number of runs conceded across overs 17-20.
Most Successful Bowling Type* (minimum 100 balls)
* RAB = Right Arm Bowler
LAB = Left Arm Bowler
WS = Wrist Spinner
FS = Finger Spinner
F = Fast
M = Medium Pace
FM = Fast medium or medium fast
RAB WS was the most effective form of bowling in this year’s BBL, with a CS of +2.9 runs saved per 24 balls (4 overs) bowled. In fact the top four most successful bowling types in this year’s BBL were all spin, with wrist spinners taking the top two spots. Spin accounted for 33% of deliveries bowled in the 2015/6 BBL, while RAB FM, which accounted for 41% of deliveries bowled, had a contribution score of -1.3 runs per 24 balls (4 overs).
Best Player of RAB WS Bowling (minimum 20 balls)
Khawaja was Australia’s best player of RAB WS bowling in the 2015 BBL closely followed by George Bailey. Lynn, who was player of the series, played RAB WS bowling the worst of all players who faced at least 20 balls from this style of bowler.
Ranking by Contribution Score in the 2015/6 BBL Final
The BBL 2015/6 final was a match highlighted by two outstanding batting performances. Pietersen’s 74 off 39 balls earned him a batting CS of +37.3 runs while Khawaja’s 70 off 40 balls earned a batting CS of +28.0 runs. Despite almost identical runs/balls counts, Pietersen’s CS was higher due the fact he almost single handedly batted the Stars to a highly competitive total after they had lost two wickets within the first seven overs with just 50 runs on the board.
Watson and Stoinis were the picks of the bowlers on the night. Watson was responsible for the two early breakthroughs that initially put the Stars on the back foot, and his overall economy rate of 5.7 rpo was the best of the match. Watson’s performance earned him a bowling CS of +14.9 runs.
Worrall and Ahmed had disappointing nights for their respective teams, both going wicketless and conceding runs at the rate of 12 rpo. Worrall’s bowling CS of -15.7 was the worst of the match, given that his 2nd and 3rd overs went for 16 and 15 runs respectively. By the time he had finished his initial 3 over spell the Thunder were 54 for 0 off just 5 overs and all the momentum was with the Thunder.
Pietersen’s player level CS of +37.3 was the greatest single contribution of the match, however Khawaja’s CS of +28.0 was not far behind and, being on the winning team, Khawaja was rightly awarded player of the match honours. DJ Hussey for the Stars was the only player to register a positive CS with both bat and ball, but it was to no avail as the Thunder won a close final by three wickets with three balls in hand.
As can be seen, CS is a useful way of comparing and examining player performances in greater detail and from different views. Contribution Scores can be calculated real time and are a useful way of identifying the key contributor to a match or series.
CS also allows us to introduce the concept of all-rounders applied to batsmen and bowlers individually. Normally we think of an all-rounder as a person who can bat and bowl. The advantage of an all-rounder in the traditional sense is that two roles are filled by the one person, allowing the team to be strengthened by selecting an additional specialist bowler or batsman. However in the science of sports analytics, where batsmen are measured by their ability to play against a variety of bowling types and bowlers by their ability to perform equally well against both LHB and RHB, the term “all-rounder” can also be applied to a specialist player who has no weakness against any type of bowling or batting. Non-all-round bowlers, for example, that are strong against RHB but not against LHB are a potential liability to the team in that the team requires another bowler to be selected to counter the weakness against LHB. An all-round bowler on the other hand removes this requirement and frees up the team to select another player with a skillset to further boost the team’s chances of success.
Examining the strengths and weaknesses of players against different bowling and batting types is a critical component of team selection.