Numbers and US

Story that numbers tell us

Archive for the ‘Sports Analytics’ Category

Australia’s Bowling Worries

leave a comment »

Look at below chart, plotting three important bowling performance matrix for all major cricket playing nations from year 2002 to year 2008. It’s surprising to know that Australia has the worst economy rate , the worst bowling average, and the worst strike rate. Yes, I know , you must be thinking about the great Mcgrath and the great and entertaining Warne. While Mcgrath retired in year 2007 with career statistics: Average,22.02,Strike rate,34,Economy rate,3.88; Warne retired in year 2005 with career figure of Average,25.73,Strike rate,36.3,Economy rate,4.25.

Bowling Matrix

Please note that economy rate is plotted against secondry axis.

The possible explanation could be that, may be they score high,  and so does the opposition; or may be all teams raise their level high against Aussies, and lower down their standard against other teams.

To test this hypothesis, let’s see how do rest of the countries fare against Aussies in these three measures.

Bowling All against Australia

Let’s see how Australia’s bowler has done compared to their counterparts.

All against Australia

It seems, we have no choice but to reject the earlier hypothesis. Moreover, looking at the above bar chart, one is tempted to conclude that Aussies must have lost more mathes than they would have won in this time period.

Written by SK

August 16, 2009 at 5:21 am

The logic of Duckworth-Lewis

leave a comment »

No one has control over inclement weather. Neither are we good at weather forecasting!

A cricket one day match that usually takes eight hours of play, has to withstand the whims of rain and bad light. To counter it, to have a result at end while being fair to both teams has been challenge for cricket administrators. More so because, it is like forecasting the result of matches somewhere in the middle of action, and this goes against the grain of sports whose charm lies in it’s unpredictibility.

Cricket tried few methods before settling in with D/L method. Most notorious of methods is maximun-score over concept. This costed SA the berth in WC 1992 final.

  • Maximum-score over concept: This method asked the second team to chase down the best overs ( in terms of runs) by first team in stipulated over.  Consider the popular match, SA vs Eng, WC 1992. England scored 252/6 in 45 overs, in reply SA scored 231/6 in 42.5 overs. Rain interrupted the match, and 2 overs were reduced.  Now since, out of 45 overs that England played, in best 43 overs they scored 251 runs.

Then in 1997 came D/L method, most scientific method developed by two statisticians, Duckworth and Lewis.

The essence of the D/L method is resources. The two resources are ‘overs left’ and wickets in hand. Feeding these two values, P(O,W) is calculated. The function p(o,w) has been kept confidential for commercial reasons. They have reached at this function applying statistical techniques on historial data.  They keep updating it, Most recenly it was updated on 2004 to accomodate the fact the one day cricket had become high scoring compared to last decade. The p values signifies resources left.

               p(50,10) = 100%          p(0,10) = p(0,0) = p(48,0) = 0%

Below chart plots p values.

DuckworthLewisEng

 Here is it in tabular format. It’s in concise form; the complete table can be found hereDLtable

 

 

 How to use the table?

Broadly, there can be three kind of possibilities.

The most simple 3 in 1 scenario would be team batting first, bats for 20 overs for 110/3, then there is an interruption, and the game is now curtailed to 30 over game. Team 1 bats for 30 overs with final score of 210/6. On reply team 2 is on 90/4 in 10 overs when the again rain plays its part. For bad light game is again curtailed to 20 over game, now what would be target for team 2.

Solution:

From above table,  Team 1 on its 20th over would have left with p(30,3) resources.  It’s   61.6% .                                                                                                     Now  when play resumed,  team 1 had p(10,3) resources left. it ‘s  29.8% . So total resources utilized by team 1 is 100-60.6+29.8=69.2%.  At lunch time the game was still of 30 overs, so resources available to team 2 is 75.1% . 

So at lunch the target for team 2 would be 210+225(75.1-69.2) = 223 . The value 225 is G value. When the team batting first looses some overs, sometimes it’s favours  batting side or bowling side, depending on wickets lost by batting side. For target calculation for team 2 in these situation we use

T = S*R2/R1 or S or S+G(R2-R1) for R1>R2 or R1=R2 or R1<R2 respectively. The value of G is 225.

Now back to our imaginary problem, after ten overs resources left for team 2 is p(20,4) = 44.6%, but there is a rain and they loose 10 overs more so resources left after rain 28.8%. Because of rain they loose 44.6 – 28.8 = 15.8% resources.  Hence revised target would be (210+225(75.1-69.2))*(100-15.8) = 188.

 There is one more method put forward by an Indian Vasudevan to set target. ICL has adopted Vasudevan’s method, and it seems to give more practical target compared to D/L. The criticism of D/L has been that it gives too much weightage to wicket, and wicket of , say Tendulkar and Ishant is given equal importance.

 

 

Written by SK

August 5, 2009 at 7:56 pm

Humor, Sarcasm from & on Silicon Valley

Let's have a laugh together

Product Thinking

Peeling the layers of products that delight

WordPress.com

WordPress.com is the best place for your personal blog or business site.

%d bloggers like this: