Tuesday, September 27, 2011

On The Importance Of Bowling In A Strong Attack

The table below shows a summary of Tests played in the last since January 1, 1960. It shows the distribution of bowling averages per Test Match for each side, taken cumulatively. The summary is broken into 3 groups - Matches where a team has conceded at least 35 runs per wicket; Matches where a team has conceded between 25 and 35 runs per wicket; and matches where a team has conceded less than 25 runs per wicket. Teams conceding at least 35 runs/wicket in a Test have had a 2% chance of winning, and a 55% chance of saving a Test, teams conceded between 25-35 runs/wicket have had a 36% chance of winning and a 69% chance of saving a Test, while Teams conceding less than 25 runs per wicket have a 79% chance of winning and 90% chance of saving a Test. In this post, I compare the records of the 7 most prolific wicket takers amongst Test spin bowlers by dividing each player's record in these three groups.

Tuesday, September 20, 2011

On Tendulkar's Proposed Revisions to the ODI Format

It is reported that Sachin Tendulkar has proposed revisions for the ODI format. As people have been at pains to point out, he has discussed some of this stuff before, but now it has found its way into a letter from him to the ICC Chief Executive Haroon Lorgat. It is alarming to see that Tendulkar seems to have bought into corporate gibberish like "value creation", but his idea, such as it is, makes ODI Cricket significantly more batsman friendly.

T20 Cricket, My Way

I don't like T20 Cricket. I find it to be quite boring and with very very rare exceptions, of poor quality. But I can see why BCCI has wanted to make money by producing cricket for an evening's entertainment. So here is my attempt to design an evening's cricketing entertainment. My interest is in producing a contest between bat and ball that is not held hostage by the clock. It should last for about three to four hours, and it should allow teams to win and lose and have stars.

Here is how it works.

Monday, September 19, 2011

Inside Harbhajan Singh's Recent Performances

Over the past few months, I have been trying to argue that Harbhajan Singh a really good bowler who is down on form, and that this may have something to do with the way in which the fast bowling has tended to collapse around him in recent Test Matches. It has been an uphill argument, given that the vast majority of the blogosphere seems to think he's "crap". This is a peculiar species of blogger and cricket fan - one that cannot be bothered to look up a player's record on Cricinfo's statsguru, but nevertheless are happy to play their hunches. A player's record is a good place for the examination of perceptions about a player.

Ponting compared to his Australian contemporaries

Please click on the image to see it at full size. This image should explain how the "Value" of a player changes over time. Michael Hussey had a phenomenal start to his career and averaged 84 over his first 20 Tests.

By mistake, I have not added Damien Martyn to this graph.

Sunday, September 18, 2011

Comparing Dravid, Kallis and Ponting

These three batsmen made their Test debuts with 6 months of each other in the mid-nineties, and have been amongst the champion batsmen of their era. Here is a graph showing how their "Value" to their respective teams compares over the years. Please click on the graph to see it in full size.

In the graph, Dravid is red, Ponting is yellow and Kallis is blue.

A New Batting Statistic

I have recently been working on a way to determine the value of a player to his team in Test Cricket. In this post, I will propose a new batting statistic based on this work. I call this statistic "Value". The value of a player an indication of the importance of a particular player to his team.

For a batsman in Test Cricket
Value = (Batting Average) x (Performance) x (Share of Team Runs)

Batting Average is already well understood. It is the number of runs made per dismissal.
Share of Team Runs is the total runs made by a batsman divided by the total number of runs made by his team in Test Matches that he played in.
Performance attempts to measure when a batsman makes runs. It is given for each player for each Test Match by -
Performance = (1st innings Runs + 2nd innings Runs) - (Number of Dismissals)*(Team Batting Average)

The number of dismissals can therefore be 0, 1 or 2.

Here is an example. In the recent India v England series, Ian Bell made 504 runs at an average of 84 in 4 Tests, while Rahul Dravid made 461 at an average of 76.83 in 4 Tests. Here is how their respective "Value" can be computed:

Bell had a very fine series. He made 18% of England's runs, including 2 big centuries. He had more support than Dravid did. Both Pietersen and Cook made big scores as well. Prior, Broad and Bresnan made a significant runs lower down the order. Dravid made nearly a quarter of India's runs and had very little support. His Value score for the series is (1356) nearly 3 times that of Bell's (458). In the table above "VALUE" is a cumulative average score, as is "PER." (Performance). "Bat P" is the Performance for each match.

The "Value" score helps differentiate batsmen with similar averages. It puts their averages into perspective. Here are two tables showing the career Value of Test batsmen. It is also interesting to trace the development of a player's Value over the course of his career, but I will leave that for another post.

See Brian Lara in comparison to Ponting, Jayawardene and Miandad in this list to get a sense of how "Value" distinguishes players where batting average doesn't. Other comparisons could be between Greg Chappell and Ricky Ponting and George Headley and Graeme Pollock. Note Mohammad Yousuf's high Value for Pakistan. This is mainly due to his blistering form in the post-Inzamam years. See Graham Gooch and David Gower in comparison to Steve Waugh. Also find England's current World Number One batting line up of Alistair Cook, Kevin Pietersen and Ian Bell all with career "Value" in the low 100s.

Test Batsmen with a Career Value of at least 150 (Min: 2000 Career Runs)

Test Batsmen with a Career Value of from 100 to 150 (Min: 2000 Career Runs)

Saturday, September 17, 2011

Cricket Is In Trouble In India

The England tour was a total disaster. India won nothing, lost everything that could be completed, and lost nearly their entire first team to injuries. A great generation is coming to the end of their careers, and this series may just have hastened that end. It is far from clear how much longer Virender Sehwag intends to play serious cricket. It is even less clear how much longer he will be good enough to play serious cricket (serious cricket = unlimited overs games, or at least games where the bowlers have a chance to bowl a serious spell). For a player of Sehwag's method, losing that small fraction of a second in reaction time due to age could be fatal.

Tuesday, September 13, 2011

Walking is Dead

On the final day of the second Test between Australia and Sri Lanka at Pallekele, this happened. The Australian commentator couldn't see why Jayawardene didn't walk. According to this commentator, there could not have been any doubt about the legal completion of the catch by Clarke. It was a wonderful take - the ball was dying on Clarke, and was to his wrong side (Or was it? Clarke bats right handed but bowls left arm spin).

Two Classic Dismissals

Caught Sobers Bowled Gibbs. Notice how close Sobers stands Gibbs' leg trap. Sobers explained that he was able to stand so close at leg slip because he worked out a method of safely stepping behind the wicketkeeper if the batsman shaped to sweep. This had the added benefit of not requiring Sobers to turn his back on the action, so that in case of a miscued sweepshot, he would still have a chance to catch it.

The second catch is a classic.


Sunday, September 11, 2011

Inside India's Rise to World Number 1

2006 was an important year in India's Test history. The year began badly, on some very very flat wickets in Pakistan. Virender Sehwag made 254 at better than a run a ball at Lahore's Gaddafi Stadium in a quadruple century stand with Rahul Dravid. This innings marked the beginning of his decline as a Test player. In 2006, his batting seemed to slip from the carefully orchestrated mayhem of 2003, 2004 and 2005, into a deeper madness. He made 180 in St. Lucia, but little else, and was eventually dropped for the tour of England in 2007 after a terrible run in South Africa where he had to be removed from the opening slot. Zaheer Khan too appeared to have ended his Test career in 2006. After 5 indifferent years in Test Cricket, Zaheer appeared to have run out of ideas. He was dropped for the home series against England in 2006, as well as the subsequent tour of West Indies. In 2006, two promising fast bowlers made their Test debuts for India - Munaf Patel and S Sreesanth. Neither appears to have survived India's rise to the top.

The Curious Case of Shane Warne

A couple of weeks a go, I posted a method to determine the value of a bowler to his team in Test Cricket. The idea was simple:

So the most valuable bowler is the one who produces the best bowling performance by taking the highest share of the 20 available wickets, in the smallest share of overs at the lowest cost in term of runs.

I introduced a measure called 'performance', which produced, for each match, a figure comparing one bowlers performance to that of his colleagues.

A player's "Value" takes into account the relative strengths of the attacks he bowled in.

The adjoining table shows the world's top 24 wicket takers. The most curious result here belongs to the second name in that list.

One way to read this would be to say that Warne was less valuable to Australia than Harbhajan Singh has been to India, or Lance Gibbs was to West Indies. Yet, Warne's record is decisively superior to these bowlers'. Compare 708 wickets at 25.4 with 406 wickets at 32.2.

For that matter, compare McGrath to Marshall. McGrath's 563 wickets at 21.6 bring him a "Value" of 0.73, while Marshall's 376 wickets at 20.9 bring him a value of 1.08.

Tuesday, September 06, 2011

What does BCCI have to hide?

This is the remarkable title of an article by Kunal Pradhan in the Mumbai Mirror on September 6, 2011. Pradhan is the reporter who "broke" the story about Sunil Gavaskar and Ravi Shastri being paid by BCCI for doing commentary in matches played in India.

At the outset, let me say that bringing National Sports Federations under the purview of the Right To Information Act may well be an excellent idea. Where public money is at stake, scrutiny is welcome. But it is ironic that a reporter like Pradhan who's reportage of the Gavaskar story can only be described as being either incompetent or wilfully malicious, is asking us what the BCCI has to hide?

A Problem with Pitch Mats

See the two screenshots below. They are both from the same delivery - the first ball that Kumar Sangakkara faced in the 2nd innings of the Galle Test. Cricinfo's commentator recorded it thus: 

Monday, September 05, 2011

On Cricket's Technology Fetish

Cricket has acquired a technology fetish. Fans, administrators and players alike demand "foolproof" technology. Some of the same people (and a few others) also want technology to be "100% accurate". Many of these people think that these two demands are interchangeable. They are wrong about this. To be foolproof is to merely be immune to human incompetence or misuse, where as being "100% accurate" is precisely that - to be perfectly accurate. That is, to be 100% accurate is not only to be foolproof, but to also be absolutely right. This technology fetish is to the detriment of good decisions. I am increasingly convinced that it is not the treatment of the howler, but the treatment of marginality that is the central issue. The thing is this - if all you want to do is sort out howlers, then you don't need DRS. By definition, a howler should be palpably obvious, or at least, the error should be very clear. The elaborate charade of players asking for reviews, and 4 different technologies giving their reading should be unnecessary. What DRS does is blow-up marginality.

Saturday, September 03, 2011

On the Philip Hughes LBW - II

In my previous post on the Hughes review I suggested that one of the possibilities could be the the distance from pitching point to point of impact could be less than 2 m. I sent my post to Hawkeye Innovations, mainly to see if my hunch about why the prediction went wrong was reasonable. They responded by saying that since they are currently talking to the ICC about Umpire Taufel's comments, they will not comment further on the issue. They also suggested that I should have looked at the ICC's revised DRS regulations before writing my first post. The Managing Director of Hawkeye Innovations, Stephen Carter wrote to me that "If you want to understand the reason why the 40cm graphics weren't displayed I suggest you consult the revised icc playing conditions, which I would have thought would have been a minimum level of research you would have taken before writing your blog."

Friday, September 02, 2011

On the Phillip Hughes LBW

The dismissal in question is at 2:30 in the video.

The incident has gotten plenty of attention. The ICC's senior Umpire, Simon Taufel is on tour in Sri Lanka currently, and has referred the Hughes decision to the ICC as a serious question mark against the accuracy of Hawkeye, the ball-tracking technology being used in the series. For a general sense of how ball tracking technology works, see this.

Umpire Taufel has brought up the question of independent testing, something which I cover in that long survey of the technology. He also brings up the question of whether or not it is fair for Umpires to be asked to make decisions first, which are then subject to review. I have brought this up before. In general, I think it is good that an Umpire has gotten publicly involved in the technology.

Cricinfo's report quotes Umpire Taufel as saying the following:

Thursday, September 01, 2011

Most Valuable Players - Batsmen

A few days a go I posted a method to determine the value of a player to his team in Tests. I posted a single measure for bowlers. In this post I focus on batsmen. The method of calculating "Performance" is as described in my first post on this subject. Performance is a measure of when a player scored his runs as well as how prolific he was. In addition, I use every batsman's share of his team's runs. A players "Value" to his team, is calculated as follows:

Value = (Share of Career Runs)*(Performance)

So the most valuable player is one who is furthest ahead of his compatriot contemporaries. The fact that Lara's "Value" is 7.07, while Ponting's is 2.00 does not say that Lara was a better batsman than Ponting, but it does say that Lara was about three and a half times more value to West Indies than Ponting was to Australia.

The highest rated players in this list, involve some players who played in weak teams and were world class players in modest batting line ups (like Andy Flower, Brian Lara and Vijay Hazare). Dudley Nourse, who is 4th on the list makes Cricinfo's All time South African XI. He once made 208 in 9 hours at Trent Bridge with a broken finger. Other high rated players like Bradman, Pollock, Barringhton and Hammond appear at the top through sheer weight of runs. Jacques Kallis and Sachin Tendulkar have nearly identical worth.

Have a look at this list, and do post questions, comments, criticisms or suggestions in the comments section.