A possible transparent quality measurement solution for BCB’s selection challenges
Our current team selection process is subjective at best. There is no transparency as to what are the criteria for a team selection. The selection process of the BCB may not be the worst nor may it be very different from the other boards as well since it gets advice from the CA, but any wrong selection by the BCB hurts the most since we are unquestionably with very limited number of experienced talents at the moment. However, I must say that making the selection process efficient alone will not improve the quality of our national team and its performance to the fans’ desired level. A quality team needs quality performers. Quality performers can be produced on a consistent basis only with appropriate planning and executing the right plan. However, even after producing quality players, if the selection process fails to pick the correct team, any quality improvement of our players will be wasted. How to improve our cricket industry as a whole, including players, coaches, umpires, infrastructure, etc., is another discussion. But assuming we always want to select the best team based on the available resources, we must have the best possible selection process. Hopefully, I am not “reinventing the wheel” here, but I am unaware if any board is currently using a quantitative quality measurement system for the team selection purposes although it’s not a rocket science either and the method has been successfully used in many industries for evaluation, selection and other purposes.
Some of the Team Selection Challenges
There may be many different challenges while selecting a team. Without analyzing or listing them all, my focus is to provide just a few examples here as I believe that by analyzing with appropriate subject matter experts, the proposed solution can identify and address each of the challenges while implementing the proposed solution.
To be an efficient team, it needs not only just the best individual performers, but the best suitable performers who can perform their best together as a team. This means, the players of a team has to be good team players where each has his role. Once the role has been changed, the individual may or may not perform to his capabilities. For example, let’s assume that we have 5 excellent opening batsmen, whose averages are higher than all our available batsmen of any positions. Can we select them all and let them play in different positions? The answer can be yes or no based on many other facts that can easily be overlooked without a system that can provide the needed facts. What about chemistry between individuals? For example, people who follow the NBA will agree that Kobe Bryant of LA Lakers is one of the best technically skilled guard in the NBA even though many of us may not like him for his selfishness. But putting him in the same team with many very good players may or may not win the team games because of his style of play. Yet, he may still be very effective with appropriate team members who are suitable to his style. What about that a bowler or batsman who has the highest average, but also has the lowest or not so good average for a particular pitch type, in which the team will be playing? Can we still pick the same bowler or batsman even though we may have an alternative player who is suitable just for that type of pitch? To be able to capture these types of data, we must have a system that can provide the necessary data at the most appropriate granular level. Otherwise, the selection may always be based on personal preferences and understandings of the selectors.
For a team like ours, whose selectors aren’t as experts as many of the top sides because of their lack of experiences, the selection challenges are even greater. Any partiality or even impartiality can add one more dimension to the dilemma. For example, no doubt that Tamim Iqbal and Nafees Iqbal are two great talented players. Now that their uncle, Akram Khan, is a selector, can questions be raised if both are selected for the team? I am not implying that Akram Khan is biased. But it’s not just about him since the same type of scenario can happen again, and what if one in another such scenario isn’t as honest as Akram Khan. As talented as Tamim and Nafees, there may always be a case that someone or two are better than them. But what if they are the best and we still question their inclusions because of their uncle’s status? Only transparency can help in such cases. How can we make it transparent without appropriate supporting data? This is where quantitative quality measurement can help. However, we also should understand that there will be new challenges if the new systems and processes are put in place. The biggest one is that our selectors will have to be trained so that they understand each and every quality element and their values as well as how those quality factors can be used while creating appropriate selection criteria.
What’s Quantitative Quality Measurement?
Quantitative quality measurement is quantifying a quality element. To simply put, as the name suggests, quantitative quality measurement is expressing a quality in terms of number, which is sometimes more useful than the qualitative expressions. For example, from a scale of 1 to 10, if 9 and10 are equal to “excellent” then instead of saying that A is an excellent batsman, who might have gotten the score of 9 or 10, it may be more meaningful just to use the actual number of 9 or 10 while expressing A’s quality as a batsman. However, from this example, one shouldn’t conclude that what’s been proposed here is same as placing a player’s overall quality by a single ranking indicator. This numbering should also not be used for performance evaluation purposes to award the best performers such as the best batman or best bowler so that the system ranks the players to 1, 2, 3, etc. Instead, the quantitative quality measurement should be used in measuring many different quality elements of each player.
Therefore to use the quantitative quality measurement system appropriately, BCB will need to create a system that can collect all the qualitative elements of BD players and score those attributes so they can be expressed in numbers. And these numbers can be used later to identify the best matche(s) for the selected criteria. If we would have known that the selection criteria would remain constant, it would have been much easier to use a top-down approach to identify all the elements related to the selection criteria. However, selection criteria should never be constant since each tournament has different needs. So a bottom-up approach is necessary in this case. Below are examples of the processes that can be used:
1. Quality Measurement Process
A. Identify and define the quality elements
A detailed analysis needs to be performed to identify all necessary quality elements that may be necessary to evaluate a player’s quality. Quality elements for a bowler can be factors such as best speed, lowest speed, average speed, average against right handed batsmen, left handed batsmen, against #1 batsmen, against #2 batsmen, average while bowling with new balls, with old balls, average in bouncy wickets, in slow wicket, etc. Appropriate quality elements need to be identified for batsmen, bowlers, and keepers. The more quality elements are captured, the more expensive it will be to maintain, but the more granular it is, the more accurate the measurement will be as well. While identifying the quality factors, the formulas and procedures need to be defined so that each quality element can be expressed quantitatively. Everyone knows of some of the standard quantitative quality data definitions such as strike rate, average runs per inning, bowling economy rate, etc., and those have been used forever. However, they aren’t granular enough. They also can’t be directly added or subtracted since some are positive data and some are negative data. For example, a higher number of economy rates of a bowler can give a negative image of a bowler whereas a higher strike rate can give a positive image of a batsman. It will also cause difficulty to scaling. However, similar (to those mentioned above that we know of) data of many individual quality elements need to be stored. Each then needs to be converted to a score in a scale of 0 – 10 or 0 – 100 or of another range that can easily be manipulated arithmetically. As a simple example, using a 0 -10 scale as the quality indicator, we identify the following data elements that are required for a batsman selection:
Average Test Career Runs, Last 5 Test Matches, Test Run Career against OB, Test Run against OB Last 5 Matches, Test Career Runs against Fast Bowler, Test Run against Fast Bowler Last 5 Matches, Test Runs in Bouncy Pitch, Test Runs in Bouncy Pitch Last 5 Matches.
The above quality data elements will be used while exampling the selection a team member in the selection process section below. Again this is purely for demonstration purposes while actual quality elements can be many more and the scaling and score can also be customized.
B. Identify the players’ eligibility for selection
Identify which players are eligible to be selected in the team. For example, are the players playing club crickets eligible to be selected? What about FC players? What about players of different age group matches? Once a player becomes eligible, he needs to be registered in the system so that all his performance data can be collected. For continuing the example through selection process, we determine that the only players are eligible for a Test team selection are the ones who have Test playing experience (although in real life this will be so impractical that we will never have any new Test team member).
C. Collect data
Identified quality element data of all eligible players from all of their games need to be collected and stored appropriately. This data will be used accordingly while selecting a team based on the team selection process and criteria. For example, based on the eligible criteria above, we collect the following data that will be used later for the selection purposes:
X: Average Test Career Runs = 7, Last 5 Test Matches = 6, Test Run Career against OB = 9, Test Run against OB Last 5 Matches = 8, Test Career Runs against Fast Bowler = 4, Test Run against Fast Bowler Last 5 Matches = 5, Test Runs in Bouncy Pitch = 5, Test Runs in Bouncy Pitch Last 5 Matches = 5.5.
Y: Average Test Career Runs = 7, Last 5 Test Matches = 6, Test Run Career against OB = 9, Test Run against OB Last 5 Matches = 8, Test Career Runs against Fast Bowler = 4, Test Run against Fast Bowler Last 5 Matches = 5 Test Runs in Bouncy Pitch = 3.5, Test Runs in Bouncy Pitch Last 5 Matches = 3.
Z: Average Test Career Runs = 6, Last 5 Test Matches = 7, Test Run Career against OB = 6, Test Run against OB Last 5 Matches = 8, Test Career Runs against Fast Bowler = 4, Test Run against Fast Bowler Last 5 Matches = 5 Test Runs in Bouncy Pitch = 0, Test Runs in Bouncy Pitch Last 5 Matches = 0 (since in Test, he never played in bouncy pitches).
2. Selection Process
A. Identify team selection criteria
Team selection criteria can be varied depending on various factors such as objectives or goals of a tournament, type of pitches that will be played on, opposition teams, etc. However, whatever those factors are, selectors need to identify all the quality factors that are necessary for the team prior to each team selection. This should be done independent of who is available, which can be addressed later at the time of actual selection so that if the best choice isn’t available, the immediate best available player can be selected based on the same criteria. Selection criteria can be and should be identified separately for each different position (e.g. batsman #1, #2, or pace bowler #1, #2, etc.). This way, we can have one obvious choice for each spot, and the backup ones as well.
In consistent with the examples of above sections, consider this that a # 5 batsman needs to be selected for an upcoming Test series that will be played on a bouncy pitch and the opposition team also has a killer OB bowler. So the selectors identify the following quality elements as criteria:
Last 5 Test Matches, Test Run against OB Last 5 Matches, Test Run against Fast Bowler Last 5 Matches, Test Runs in Bouncy Pitches Last 5 Matches.
Selectors could have decided to use a weighted formula to give different weights to different quality elements, but they thought equal weight will do the work.
B. Select the team
Once the quality measurement process can provide all the up-to-date quantitative data for each quality elements, and the selection criteria have been identified, selection of the team will be as simple as to finding the best numbers for each position and selecting a player for it. Using the selected criteria above, the selectors select X as the best fit for the # 5 spot by calculating the followings:
X: 6 + 8 + 5 + 5.5 = 24.5
Y: 6 + 8 + 5 + 3 = 22
Z: 7 + 8 + 5 + 0 = 20
A new team can be selected or the old one can be kept as desired by the selectors. The selector should also have the flexibility to alter any selection based on the special needs with appropriate justification and assuming all risks. Although this sounds like a hybrid solution initially, it may be handy until the system becomes mature. However, the criteria and all results while selecting the team need to be transparent to public. This way, selectors can be held responsible of any wrong doing.
3. Re-evaluate Quality Measurement and Selection Processes
For continual quality improvement, quality measurement and selection processes need to be re-evaluated periodically and be updated as necessary.
Finally, we must understand that with the quantitative quality measurement system and this advanced selection process in place, the human selectors won’t and shouldn’t go away. However, their job responsibilities will change from selecting the names to managing and enhancing the selection criteria and managing the selection system and process. Therefore, subjectivity may remain to some extent. And we may argue with the selected criteria instead of the actual selected names, but the transparency factor will always make the selectors think twice even when they won’t be perfect.
Last edited by sharifk; July 2, 2007 at 05:12 PM.