Batting July 7, 2009

ODI Strike Rates - a fresh look (and a preview of Test Bowler Analysis)

A comparison of batsmen career strike rates with the strike rates of the rest of the team in the matches played by the batsman indicates how quickly he scored compared to his team-mates

Since I need some time to complete the Test Bowler Analysis, I have come out with an article on ODI Strike Rates. What started as an interim article has turned out to be a very interesting one.

Whenever we compare measures across years we always have problems since the relevant period strategies, pitch/ground conditions, quality of bowling (or batting), prevailing laws etc vary significantly. Shahid Afridi's 100+% strike rate cannot be blindly compared to Viv Richards' sub-90 strike rate since everything has changed over the years.

I have created a new factor comparing the Batsman career strike rate with the rest of the team's strike rate, in the matches played by the batsman. The great thing with this measure is that this stands firm across decades, across different types of pitches/conditions and across different types of bowling skills and strategies.

If the average scoring rate of the period was way below currently acceptable values, no problem, this condition applies to all the players in that match. Was the pitch unplayable, no problem, this condition applies to all the players in that match. Was the pitch a belter, no problem. Were the grounds small or huge, no problems. Was there a devastating bowling attack, no problem. Was it the East African or Canada bowling attack, no problem, all should have helped themselves to the buffet lunch. And so on. Our comparison applies only to matches played by the batsman so these are completely valid.

The analysis has also evolved. My first idea was to compare the batsman's career strike rate to the team's overall strike rate. Then I changed to the concerned match strike rate of the team but this had an element of overlap since the player's own performance is embedded in the team's performance. Finally I came out with the idea of taking into account the other players' strike rates. This has worked out very well.

Now let us look at the tables. The criteria is that the concerned batsman should have scored a minmum of 2000 ODI runs. Even this means that there is a sample size of 146 batsmen. This table is current upto match no. 2855, the fourth ODI between West Indies and India.

Table of Career strike rates to Concerned match team strike rates

SNo Batsman           Cty Mat  Runs Balls  S/R OBRuns OBBalls   S/R  BSRF
1.Shahid Afridi Pak 276 5642 5083 1.110 52937 65461 0.809 137.3% 2.Kapil Dev N Ind 225 3783 3979 0.951 35676 49298 0.724 131.3% 3.Powell R.L Win 108 2085 2157 0.967 18941 24678 0.768 125.9% 4.Richards I.V.A Win 187 6721 7451 0.902 28195 38757 0.727 124.1% 5.Sehwag V Ind 205 6592 6472 1.019 40230 46569 0.864 117.9% 6.Wasim Akram Pak 356 3717 4224 0.880 55541 73789 0.753 116.9% 7.Jayasuriya S.T Slk 431 13151 14443 0.911 77876 97706 0.797 114.2% 8.Klusener L Saf 171 3576 3978 0.899 27976 35034 0.799 112.6% 9.Gilchrist A.C Aus 287 9619 9923 0.969 56114 64341 0.872 111.1% 10.Flintoff A Eng 141 3393 3819 0.888 22790 28419 0.802 110.8% 11.Chappell G.S Aus 74 2331 3088 0.755 11416 16449 0.694 108.8% 12.Pollock S.M Saf 303 3519 4059 0.867 43168 54126 0.798 108.7% 13.Cairns C.L Nzl 215 4950 5879 0.842 36554 47167 0.775 108.6% 14.Zaheer Abbas Pak 62 2572 3216 0.800 9520 12863 0.740 108.1% 15.Tikolo S.O Ken 117 3213 4214 0.762 18721 26291 0.712 107.1% 16.Gower D.I Eng 114 3170 4222 0.751 19486 27765 0.702 107.0% 17.McCullum B.B Nzl 153 2984 3353 0.890 24937 29918 0.834 106.8% 18.Pietersen K.P Eng 92 3127 3576 0.874 15244 18585 0.820 106.6% 19.Botham I.T Eng 116 2113 2816 0.750 19731 27866 0.708 106.0% Silva P.A Slk 308 9284 11497 0.808 51495 67537 0.762 105.9% 21.Rhodes J.N Saf 245 5935 7310 0.812 42228 54993 0.768 105.7% 22.Trescothick M.E Eng 123 4335 5086 0.852 21661 26647 0.813 104.9% 23.Symonds A Aus 198 5088 5504 0.924 34568 39054 0.885 104.4% 24.Tendulkar S.R Ind 425 16684 19481 0.856 76047 92266 0.824 103.9% 25.Moin Khan Pak 219 3266 4011 0.814 37111 47228 0.786 103.6% ... 40.Gibbs H.H Saf 244 8038 9647 0.833 45073 54128 0.833 100.0% ... 142.Yasir Hameed Pak 56 2028 3029 0.670 11363 12777 0.889 75.3% 143.Wessels K.C Saf 109 3367 6057 0.556 16626 22456 0.740 75.1% 144.Tillakaratne H.P Slk 200 3789 6544 0.579 31601 39951 0.791 73.2% 145.Mudassar Nazar Pak 122 2653 5067 0.524 19282 25900 0.744 70.3% 146.Marsh G.R Aus 117 4357 7721 0.564 20183 24649 0.819 68.9%
Note: The OB figures reflect the aggregate of the runs/balls of the other batsmen who batted in all the innings in which the concerned batsman has batted. If the concerned batsman did not bat at all, the figures for that innings are not included in the aggregate.

As expected Shahid Afridi is at the top. He has out-scored his team-mates by an amazing margin of 37.3% although his team-mates themselves score at a fair clip, 80.9. This underscores his value to the team. He outperforms his team-mates by such a wide margin, I fail to understand how the selectors could ever drop him, I am not even referring to his bowling.

Look at the second entry, also a proof that this measure cuts across years with ease. Kapil Dev has outperformed his team-mates by over 26%. His team-mates have been sluggish. However this understandable since those were the times. It was outstanding performance by Kapil Dev to score at a great strike rate of over 90% during those days when 70 was the norm.

Third player in the table is Ricardo Powell, who has out-scored his team-mates by over 25%. Whatever happened to Powell.

Now comes two interesting entries. Viv Richards' value to his team cannot be exemplified more than by this measure. He has outscored his team-mates by over 21%, day in and day out. This, coupled by the achievements of those mean and fiery fast men, was primarily responsible for the West Indian successes of the 1970s/80s.

Then comes the modern great, Sehwag. His team, India itself, has scored at a pretty good rate, 86.4. Sehwag has still managed to outscore his team-mates by 18%. This single factor has been one of the main reasons for the Indian team's recent successes.

In the next 5 places we have Wasim Akram, Jayasuriya, Kluesener, Gilchrist and Flintoff who have all outscored their team-mates by over 10%. All are great strikers.

Tendulkar has managed to outscore his team-mates by around 4%, mainly because the rest of the team, with a number of attacking batsmen, including Sehwag, Yuvraj et al, have scored at a good rate of 82.4. But his contributions, in the opening position, have been outstanding. Note the relatively lower placement of Symonds, just over 4%, indicating, a la Tendulkar, the higher scoring rate of his team-mates, in this case a very high 88.5.

Gibbs is the only batsman who has almost exactly mirrored his team-mates' achievements.

At the other hand we have mostly defensive batsmen of olden years, led by Geoff Marsh whose team-mates have outscored him by over 30%. The only modern batsman is Yasser Hameed who has scored at an amazing 25% below his team-mates, accepting that this group includes Afridi.

To view the complete list, please click here.

The above table includes the team extras in the runs scored. Thus the rest-of-the-team strike rates is slightly higher. I have given below the same table, this time excluding the team extras. No major changes.

SNo Batsman           Cty Mat  Runs Balls  S/R OBRuns OBBalls   S/R  BSRF

1.Shahid Afridi Pak 276 5642 5083 1.110 49132 65461 0.751 147.9% 2.Kapil Dev N Ind 225 3783 3979 0.951 32898 49298 0.667 142.5% 3.Powell R.L Win 108 2085 2157 0.967 17332 24678 0.702 137.6% 4.Richards I.V.A Win 187 6721 7451 0.902 25859 38757 0.667 135.2% 5.Sehwag V Ind 205 6592 6472 1.019 37006 46569 0.795 128.2% 6.Wasim Akram Pak 356 3717 4224 0.880 51127 73789 0.693 127.0% 7.Jayasuriya S.T Slk 431 13151 14443 0.911 70806 97706 0.725 125.6% 8.Klusener L Saf 171 3576 3978 0.899 26076 35034 0.744 120.8% 9.Flintoff A Eng 141 3393 3819 0.888 20940 28419 0.737 120.6% 10.Gilchrist A.C Aus 287 9619 9923 0.969 52125 64341 0.810 119.7% 11.Tikolo S.O Ken 117 3213 4214 0.762 16758 26291 0.637 119.6% 12.Cairns C.L Nzl 215 4950 5879 0.842 33299 47167 0.706 119.3% 13.Zaheer Abbas Pak 62 2572 3216 0.800 8669 12863 0.674 118.7% 14.Chappell G.S Aus 74 2331 3088 0.755 10480 16449 0.637 118.5% Silva P.A Slk 308 9284 11497 0.808 46393 67537 0.687 117.6% 16.Gower D.I Eng 114 3170 4222 0.751 17751 27765 0.639 117.4% 17.McCullum B.B Nzl 153 2984 3353 0.890 22785 29918 0.762 116.9% 18.Botham I.T Eng 116 2113 2816 0.750 17981 27866 0.645 116.3% 19.Pollock S.M Saf 303 3519 4059 0.867 40335 54126 0.745 116.3% 20.Pietersen K.P Eng 92 3127 3576 0.874 14069 18585 0.757 115.5% 21.Trescothick M.E Eng 123 4335 5086 0.852 19830 26647 0.744 114.5% 22.Lamb A.J Eng 122 4010 5290 0.758 19026 28691 0.663 114.3% 23.Rhodes J.N Saf 245 5935 7310 0.812 39173 54993 0.712 114.0% 24.Tendulkar S.R Ind 425 16684 19481 0.856 69447 92266 0.753 113.8% 25.Crowe M.D Nzl 143 4704 6464 0.728 20206 31581 0.640 113.7% ... 77.Inzamam-ul-Haq Pak 378 11739 15827 0.742 60323 81270 0.742 100.0% ... 142.Taylor M.A Aus 113 3514 5867 0.599 18912 25762 0.734 81.6% 143.Yasir Hameed Pak 56 2028 3029 0.670 10522 12777 0.824 81.3% 144.Tillakaratne H.P Slk 200 3789 6544 0.579 28664 39951 0.717 80.7% 145.Mudassar Nazar Pak 122 2653 5067 0.524 17685 25900 0.683 76.7% 146.Marsh G.R Aus 117 4357 7721 0.564 18347 24649 0.744 75.8%

Test Bowler Analysis

I have given below a brief write-up on the Test Bowler Analysis. If you want to send in your comments on this, please do so, as a separate comment, titling the same, "Test Bowler Analysis".

1. Period Separation: These periods have been identified with lot of thought and deliberation with inputs from a few interested readers. Many related factors have gone into this process. Separate tables will be prepared for different periods. I have considered, and rejected, a separation of Pace and Spin since there will be too many tables and we will not have the charm of a Murali/Warne vs Hadlee/Lillee comparison.

- The bowling era: 1877-1914 (134 Tests and 370 players)
- The batting era: 1920-1969 (535 Tests and 980 players)
- The balanced era: 1970-2009 (1251 Tests and 1220 players).

2. Match Performance: This is a very important aspect of any such analysis. Many readers have completely forgotten that this is not career-based and takes into account every ball bowled and wicket captured weighted by the match conditions and the opposition. Those who are shouting at the rooftops that the career-end figures are not favourable to one player over the other should take the trouble of understanding this aspect of analysis carefully. This will incorporate the following factors.

- Wickets captured (Base)
- Balls bowled (Base) - to recognize long spells
- Batsmen dismissed - based on his score at time of dismissal (Base)
- Overall quality of batting team - primarily top-7 batsmen
- Bowling accuracy - relative to the innings scoring rate
- Own team's bowling quality (to take care of very strong bowling sides)
- Match-related pitch characteristics
- Match situation (incl first day spinners' performances, defending low/high totals in innings 2, innings 3 situation, levels of fourth innings totals defended, win margins et al.
- Home/Away - incorporating relative team strengths
- Result - incorporating relative team strengths.

3. Career Achievements: This is an equally important aspect of any such analysis. It also encompasses aspects of bowling which do not require consideration of the match conditions or situation. The only longevity measure is the "Career wickets captured" measure with no more than a 10% weight. This will incorporate the following factors.

- Career Wickets captured
- Bowling Strike rate (BpW)
- Bowling accuracy (RpO)
- Average Quality of batsmen dismissed - based on CtD batting average
- Type of wickets captured - Top order / Middle order / Late order
- % of wickets with own efforts - Bowled/Lbw/C&b (Still undecided on this).

Once again reminding the readers to send separate comments on this topic.

Anantha Narayanan has written for ESPNcricinfo and CastrolCricket and worked with a number of companies on their cricket performance ratings-related systems

Comments have now been closed for this article

  • testli5504537 on July 15, 2009, 16:24 GMT

    TEST BOWLER ANALYSIS Ananth, looking forward to it. Two suggestions : (a) it will be great if you could analyze the best/most effective bowling "combinations" - either pairs, or three-somes, and (b) instead of ranking them serially, and creating a "how could x be better/worse than y?" arguements, create "bands". Thanks! [[ A few months back I did an article on the exact bowling-pair combinations. If you to the "It Figures" archives you will see that. Bands is a good idea even though it is only a matter of presentation. Let me see. Ananth: ]]

  • testli5504537 on July 15, 2009, 10:08 GMT

    In response to Karthik's "India has never gone 12 months without a Test"- i would have thought they did not play a test between the one-off Chandigarh test (where they won by an innings after scoring 280-odd !)vs sri lanka in aug / sep 1990 till their next test series in Australia which started late 1991.Apologies for this but just trying to put record straight- being anally retentive, you see:). In between, i think India managed a couple of visits to Sharjah (The IPL equivalent of the previous decade !!).Maybe Ananth one day will do an analysis of the impact of such cricketing legends like Packer, Bukhatir, Stanford (sadly only a brief foray into cricket until other pressing matters took over), Subhash Chandra & of course, Modi.Maybe Bacher could come into this list as well. [[ I used to think that the Sharjah was an "un"necessary evil. However that is nothing compared to the IPL which, because of the combination of money and power, is going to do unlimited harm. Mr.Modi will say one thing, even that with a forked tongue, but quietly do another thing. See the way the IPL3, FOR THIS YEAR, was thrown up with no warning. Incidentally would it be right to say that Flintoff's retirement from Tests has been prompted by the "1.5 millions for 6 weeks" opportunity Ananth: ]]

  • testli5504537 on July 14, 2009, 11:12 GMT

    hiiii ananth, u made a very good analysis. I am surprised at, how sachin tendulkar made that much of runs with excellent average and such strike rate... even sanath jayasuriya also lack of average..

  • testli5504537 on July 14, 2009, 4:29 GMT

    Shahid Afridi and Jaysuriya are the best. India in general and BCCI in particular are very greedy.For there lust of money they are destroying the game.

  • testli5504537 on July 13, 2009, 17:08 GMT

    why is it that of the 146 batsmen evaluated only 40 have their revised strike rates above 100% and 106 have below 100%? since all the players in the list have been teammates to each other shouldn't the stike rates average each other out to 100%? or has the inclusion of extras in the team strike rate played spoilsport (esp. since the rates excluding extras shows inzy having 100% at 77 on the list i.e. roughly halfway down)? also wouldn't playing for a stronger team like say australia reflect badly on a player like symmo or gilly's strike rate since the they bring EACH OTHER'S strike rates down inspite of playing brilliantly individually. could including the opposing team's strike rate along with rest of the same team help? [[ Anurag I would appreciate your comments after perusing the follow-up article. Also there is a table in the main article which excludes the Extras. Ananth: ]]

  • testli5504537 on July 11, 2009, 6:49 GMT

    OK, as long as there ARE actual Tests that India will be playing, I am placated.

    I was quite worried that it appeared that there would be no Tests from 3/09 to 3/10. India has never gone 12 months without a Test in the 23 years of my cricket watching.

    So our fear that Dravid and Tendulkar will have little to add to their Test stats may be premature.

    I sincerely hope there is no second IPL this year. I have totally lost interest in the IPL. It was fun the first time, but now it is just tedious.

    Also, if the BCCI is clever, they will realize that a Test Series and the IPL can happen concurrently. Only 11 players are utilized in the Tests, which will hardly dent the IPL pool. The opponents for India can be a team that does not send many players to the IPL, like England or Bangladesh. Tests are during the day and the IPL is in the evening, so the TV airtime need not be cannibalized.

    It is entirely reasonable to treat Tests and the IPL as different audiences. [[ Kartik That is not a bad idea. The players could also mix the two. Dhoni could play an IPL match, come to play in a test (missing 2 IPL matches) and go back to IPL. Anyway there does not seem to be complaint about the fatigue factor. Also players such as Tendulkar, Dravid, Laxman, Ponting, Zaheer, Pietersen et al could select Tests over IPL although I very much doubt whether anyone will forego the easy money offered by IPL. I have not published your other comment since that related to Test Batsmen and does not belong to this article. Ananth: ]]

  • testli5504537 on July 10, 2009, 12:20 GMT

    Ananth - to build on Arjun's idea, you could add a "Containment Measure": [%balls bowled]/[%runs given]. That will help bring out bowlers who bowled long spells of tidy bowling (mostly spinners and slow-medium pacers).

    A combination of this measure and Arjun's measure is: [%balls bowled * %wickets taken]/[%runs given]. This is a more effective measure, I think. After all, spinners are useful in helping a team bowl 90 overs/day (and avoid a fine) as they take fewer minutes per over. In early 80's, I think there were instances when WI bowled less than 68 overs/day ... that situation improved for them only after Larry Gomes became a regular. [[ Alex/Arjun In fact I was thinking about the same. I have all three pieces of data, viz., Team Balls, Team Runs and Team Wkts. It is only a matter of looking at the ratios and interpreting these correctly. The other beneficial nature of this is that it helps the spinners who could do with a few such helping ideas. Thanks. Ananth: ]]

  • testli5504537 on July 10, 2009, 9:16 GMT

    Building on Avi Singh's comment and Ananth's response, you need to take into account the limited number of balls available in an inns.

    Take Dravid, if he averages 40 with a strike rate of 71, he's using up 19% of his teams resources (balls) but only scoring 16% of their runs (I estimated this from the figures in Ananth's table) - therefore he's not very efficient.

    Afridi, on the other hand, if he averages 23, uses up 7% of his teams resources and scores 9% of their runs - meaning he's pretty efficient.

    Assuming that the average ODI innings end with a team 7 or 8 down (I don't know the exact number but feel it will be close to this)then a player like Dravid potentially costs his team runs. It might be better to pick a player who averages less but scores quicker.

    Eg. On average, Dravid scores 40 runs in 56 balls (SR 71)

    If Player X scores 35 runs in 44 balls (SR 80) this leaves 12 balls for the tailenders. If they manage 6 runs between them, the team is better off by 1 run

  • testli5504537 on July 10, 2009, 7:16 GMT

    India hosts SL for 3 tests in Nov '09, plays 2 tests in BD Dec '09, hosts SA for 3 tests in Feb '10, and plays 2 tests in Zim May '09. Schedule:

    Ananath - I think you might be wrong about where SRT and Ponting may finish at the end of their careeres. Australia plays a lot of test crickt. Ponting is doing OK and will likely play many tests in a weakened Aussie side. Looks like, he will become No. 1 in terms of #centuries, #runs, and as per your method as well.

    Someone in media and BCCI should take a note of your statistics and realize how well Indian batsmen have fared in tests, and do something to make sure India plays more test cricket. [[ Alex Many thanks. Unfortunately these did not find their place in another Future fixtures page of Cricinfo possibly because the exact dates and venues have not yet been finalized. Australia, all said and done, makes sure that they play enough tests. The other thing is that the Zimbabwe Tests are the perfect candidates for cancellation for any one of 5 reasons. The second IPL of the year is on the anvil. Ananth: ]]

  • testli5504537 on July 9, 2009, 21:23 GMT

    On an unrelated but important note :

    When are India next scheduled to play a Test? There appears to be no more Tests for India in the next 6-9 months, as per Cricinfo's 'Future Series' section.

    This could mean that we have already seen Rahul Dravid's final Test appearance.

    Does anyone have more info on when India will next play a Test? [[ Kartik The only news we have got recently is the 7-ODI (I may be wrong) series between India and Australia. No reference to any Test tours. It is obvious what is the reason for this. Mr.Modi, in his infinite wisdom (or lack of) and omnipotence would have made sure that the Indian players would be available for the Champions' League, IPL-3 et al. And the BCCI would have happily concurred. Notwithstanding all the pious statements made, the senior Indian players would not mind this. Who would scoff at 6 weeks of effort and a million plus dollars in the bank. Unfortunately the Australian/SAF boards are also party to this. Why should India and Australia play an ODI tournament so soon. And is it not a shame that the only country in which Bangladesh has not played a Test series is India. Such arrogance. Money talks, unfortunately it shouts. In the bargain many a Test career is going to be cut short, be it, as you say, Dravid's illustrious one or Ishant Sharma's chequered one. And tell me, will Ishant Sharma reach even 100 Test wickets at the current rate. I am sorry for the outburst. But your mail touched a raw nerve within me. Despite the wonderful Test matches being played, Test cricket is being smothered slowly and surely. Unfortunately not a whimper from the players who carry clout, other than the wonderful Mahila Jayawardene. Ananth: ]]

  • No featured comments at the moment.