One of cricket fans’ favourite pastimes is comparing Test batters within and across the generations.
These debates over who is better are never-ending and inconclusive at best. The common refrain used is that it is impossible to compare players from across generations as the circumstances were different. It is imprudent even to compare performances of the same batsman across two innings of the same Test match. As we saw in the Oval Test between England and India, the first innings’ conditions were drastically different from the second innings’.
The imprudence of these comparisons is not because it is impossible to compare these performances but because we don’t capture the necessary data and apply the relevant analysis to deduce it. With the availability of technology and data, cricket must analyse and start showcasing a player’s skill in a new light.
A batsman’s career is measured based on the cumulative runs he scored in his career in today’s world. By that measure, we regard Sachin Tendulkar as the best Test batsman in the world. The moment I say that, the typical response is that Don Bradman scored close to 7000 runs on uncovered pitches. Or someone will point out that Sunil Gavaskar scored most of his runs against the fearsome West Indies fast bowlers. The list of arguments is never-ending. To stop these arguments and make cricket an objective sport, I propose adopting a par score method like what they use in golf.
In golf, every hole in a course has a par score which is arrived at based on different factors such as the playing length, altitude, terrain, obstacles and so on. A player’s performance is measured against this par score as the number of strokes above or below par. In an 18-hole game, the scores are totalled and the net score above or below par for the course determines the winner of the game.
The concept of par score is not new to cricket. We have seen umpteen times when ex-players assess the playing surface and declare a range or a specific number as a par score for that pitch. The ex-player makes this inference based on his experience and past performances on the ground. It is a crude number arrived at based on a simplistic assessment of data. Now, there is enough data and technology to make this crudely-derived number into a scientific number.
How can we measure par score?
We know at a high level that the number of runs scored by a team is impacted predominantly by the skill levels of its batsmen, the quality of the opposition’s bowlers, the nature of the pitch and overhead weather conditions. All these factors are usually represented crudely by batting averages, bowling averages, amount of grass on the pitch, cloudy or clear weather conditions, amount moisture on the pitch, the type of ball used and so on.
Apart from the batting and bowling averages, the impact of the other factors has not been formally studied and quantified in the game. I know that Cricviz collects and presents data related to seam deviation, swing, spin deviation and others. This data can be used to arrive at a formula to calculate a par score for a pitch. As we know, the pitch and the overhead conditions change during a Test match. So, this par score should be adjusted at multiple intervals during the game, and a final number be formalised at the end of the game. Once this number is formalised, the team’s performance can be measured as some runs above or below par.
(Photo by Adam Davy/PA Images via Getty Images)
The beauty of this measure is that the par score will not be the same for the two teams playing in that match. As the teams have differently skilled players and may have played under different conditions during the game, the end par score for each team will be different. Even though the Test result will be decided based on the usual parameters of runs and wickets, the par score will help coaches understand how well their teams performed vis-à-vis the expected level.
The other advantage of measuring vis-à-vis the par score is that we can maintain a cumulative above or below par number to track a team’s performance over the years and until the end of time. A coach and a captain can be objectively rewarded based not only on win-loss ratio but also on the trajectory of their team’s cumulative par scores during their tenure.
The concept of the par score can be applied at an individual batter level as well, but with some adjustments. In addition to the earlier discussed factors, the number of runs scored by a batter is influenced by his position in the batting order, the number of runs the opposition has scored, the number of balls he faced during the innings and the bowlers faced during his innings. All these additional factors are also easily quantifiable. As we did for the team par score, the batter’s performance can be quantified as the number of runs above or below the calculated par score. Just like the team par score, a batter’s par score can be cumulated across his career.
(Photo by Hagen Hopkins/Getty Images)
The advantage of the par score system will be evident to the reader now. The cumulative par score represents the reality of the matches the batsman played under and illustrates how far ahead or below he ended up from the expected performance. Now, this cumulative par score can be compared across generations to say who performed better. There will be no more argument that Vivian Richards did not get to face his bowlers or his par score would have been higher than that of his opposing batsmen. That Tendulkar played during an era of flat pitches would lead to him being measured against a higher expected par score than today’s batsmen who play on more ‘demanding’ pitches.
What I have presented is an idea, and I am sure there are holes in it. I also know that past data for many of the factors will not be available. So, the par score for past matches and past players cannot be computed today. However, for the current and future players, I hope that cricket moves in such a direction of objectivity that fans stop making impossible to prove statements.
It is our divine right as cricket fanatics to know objectively who the best batsman in the game is then. We have the technology and data to make this happen. I hope such a rigorous measuring system becomes a reality soon.
Pope Paul VII
Roar Rookie
Thanks RC. Hill had a diabolical debut series 1896 – 30@6. 1897/98 included 188 out of 302 while he was at the wicket, 1/1 to 9/303 I think they played around 41 matches together, in which Trumper got 8 x 100s , 11 x 50s to Hilly 6 x 100s and 16 x 50s. Including the trio of 90s. Maybe they missed each other in 1896 and 1909? I’m not challenging Trumper’s brilliance, just that Hill was a fine batsman and just as valuable. A back foot lefty, hooking, pulling and cutting. That’s poetry to me. And gutsy. Getting 160 with the flu in stifling heat. Though fielding is often described as pedestrian in this period, Hilly ran a mile and to take an outstanding outfield catch in Trumper’s famous 100 before lunch match. Interesting Hilly was a very gregarious and popular character on and off the field. Loved a laugh with his friends and team mates. Except for Pete McAllister, whom he tried to throw out the window. Trumper, irrepressible with the bat, was quiet and sober off it.
Renato CARINI
Roar Rookie
Pope, FYI: Hill, Much higher average in draws Much higher average when leading Also, Hill played in a very high scoring series, 1897-98, which Trumper missed And then didn't play in the low scoring series of 1909. Check the numbers for the games where they both played.
Renato CARINI
Roar Rookie
Micko You nearly gave me a heart attack!! If you interviewed ANY Test cricketer from that time You wouldn't find a single one who said Hill and Trumper were close. Not one.
Renato CARINI
Roar Rookie
That's why you're a guru, Rosie :happy:
matth
Roar Guru
I agree that his assumptions may not have been perfect. I also struggled with players moving up the rankings based solely on catches per match. But at least he set his assumptions and then reviewed the results, instead of setting his result and then searched for confirmation.
DaveJ
Roar Rookie
Very true Matth, but when I looked at it again, I didn't find some of the assumptions clear or convincing. I have no problem with the Imran Khan rating- averaged 19 with the ball and 44 (52 as captain) with the bat over his last ten years. But Frank Worrell is ahead of Malcolm Marshall on the greatest cricketers list with a higher batting rating than Marshall's bowling, which is also behind Greg Chappell's and Graham Pollock's batting ratings. Something smells funny. And I think 20 Tests is too small a sample to be bringing people into all time rankings, no matter how great people think Pollock looked (against some not so brilliant bowling attacks) - ditto for Headley, rated number 3 all time after DG and GP. The real problem is that the actual calculations used are not spelled out, either for space reasons or because he wanted to keep it to himself. E.g. at one points he says his standardised averages are based on relative strength of opposition as well as the "overall standard of the time", with batting averages adjusted downwards by 5% in the 1898-1915 era and 8% in the 1920-49 era. But later he says the standardised averages reflect pitch conditions and differences in scoring between countries. Trumper's average gets boosted from 39 to 45.5, Hobbs goes down from 57.6 to 55.5, but its hard to see how it all works. (I'd have thought Trumper might go higher.)
matth
Roar Guru
That sounds about right.
matth
Roar Guru
That's partly why I liked the book. He developed a methodology and let the chips fall where they may. too many others start with a proposition they want to 'prove' and then use their assumptions and selective samples to achieve the result they were looking for. Confirmation bias.
matth
Roar Guru
Great info Lurch, thanks.
sheek
Roar Guru
That, Or we could do what we’ve mostly done, which is accept the averages as they are, & use our gut feeling, perception (right or wrong), experience & knowledge to determine who are the best. It’s an imprecise science, but so is life with many things. Here’s my latest Baggy Greens all-time XI. I say ‘latest’, because I do change it regularly (especially the openers). I have no idea if it’s 100% accurate, or 90%, or only 60%. I try to be definitive, but comprehensive might be elusive. Only God knows the real answers. Victor Trumper, Matt Hayden, Don Bradman (c), Steve Smith, Greg Chappell, Keith Miller (vc), Adam Gilchrist (wk), Shane Warne, Dennis Lillee, Bill O’Reilly, Glenn McGarth. New ideas are always welcome, especially when they add value to what we already know. Alternately, we don’t want to fall into the trap of continually re-inventing the wheel.
Micko
Roar Rookie
So many things to consider isn't there? I think you're spot on though. Warnie's best ever series was the 2005 Ashes, where he had to shoulder the burden from the second test onwards and lift even more from his usually high standards.
Jeff
Roar Rookie
What if the bowler is having to carry the attack on his own, without a bowling partner at the other end softening up the batsmen or keeping things tight and building pressure for example? Should that bowler's performance rating be increased in that case?
Jeff
Roar Rookie
"Andy’s team mates don’t count here.." But aren't partnerships an important element that add to a batsman's individual performance? Without good partners at the other end, both putting pressure on the fielding time to adjust their game plan as well as giving Andy more freedom to play the game he wants, wouldn't Flower suffer? In which case shouldn't Flower's performances rightly be enhanced?
Micko
Roar Rookie
I’m going to start Renato’s book in the next couple of days, but his articles here showed already a unique personality displayed by Trumper, where personal stats didn’t bother him. He saved his best for the toughest situations, but wasn’t ruthless about run-scoring. Kind of how Bernie showed we might need to re-evaluate Mark Waugh’s impact, as he had a similar tendency.
Pope Paul VII
Roar Rookie
Maybe Vic flatterred to deceive. A prettier bat for sure but was he anymore valuable? Playing an identical number of innings Trumper got accolades. Meanwhile Hill hit `7 x 100s, 6 x 90s (take that Slats and S R Waugh) and 5 x 80s. 18 scores over 80 compared to Trumper's 10 - 8 x 100s and 2 x 80s.
All day Roseville all day
Roar Guru
I like to compare batsmen (and partnerships) using median or "middle" scores. They are what a batsman or pairing has achieved in at least half of their innings. Medians reduce the skewing caused by not outs, and by occasional huge innings surrounded by many small ones. They have something in common with Renato C's views on the value of runs scored after reaching 100. For example- aggregate 200, average 50.00, and median 50- from the five individual scores 20, 30, 50, 50, 50no. same aggregate, same average, but median only 20- from five scores 15, 20, 20, 30, 115no. I'd prefer the team-mate who reaches 50 more than half the time, to the one who gets past 20 less than half the time.
All day Roseville all day
Roar Guru
Duckworth-Lewis-Stern still isn't quite right for limited-overs cricket, after almost 20 years of modelling, testing and refining. So it might take a while to get a par score rating system for the far more complex four-innings, five-day cricket right...
Micko
Roar Rookie
I meant Hill was NEVER considered Trumper's equal. In hindsight people now claim Hill was as good as Trumper.
Micko
Roar Rookie
2002 or 2003 was it? (Warnie's drug ban?)
Micko
Roar Rookie
Apparently the comparisons only come from people long after the fact. When both were around Hill was considered Trumper's equal.