THE CORDON HOME

BLOGS ARCHIVES
SELECT BLOG
April 26, 2010

Batting

Test batting position averages: a follow-up

Anantha Narayanan

In the article on Test batting positions, I looked at the highest averages in each batting position from Opening to No.7. There were a number of useful comments and some of the readers wanted me to create additional tables to throw more light and create a better insight into the fascinating topic. Hence this follow-up analysis.

1. The first table is a very important one asked for by Abhi. This is a matrix of Decades and Batting Position Averages.

Decade Tests  <---------------Batting average---------------->
Opening  BP 3   BP 4   BP 5   BP 6   BP 7   Op-7

1930s: 99 38.10 51.93 40.93 35.84 31.09 26.71 37.85 1940s: 44 44.13 42.62 52.71 40.71 33.34 25.68 41.02 1950s: 165 33.42 37.10 40.64 33.02 25.75 22.45 32.51 1960s: 186 36.38 41.55 41.87 38.42 33.20 24.89 36.36 1970s: 197 38.29 40.16 40.23 38.19 31.56 28.90 36.76 1980s: 267 34.79 38.10 41.64 36.43 35.14 29.24 35.85 1990s: 347 35.51 36.00 40.88 38.13 33.37 26.77 35.35 2000s: 477 37.34 43.51 44.11 41.11 34.37 30.32 38.47

Total: 1782 36.46 40.51 42.17 38.18 32.83 27.69 36.54

My gut feel is that this is going to be a very important table which will be used by many of us quite regularly. Let us see the salient numbers. First a brief explanation. For reasons which are obvious the first Test I considered was Test # 176, which began on 30 Nov 1928 (no prizes for guessing why). Hence the 10 Tests during these 13 months are clubbed with the 1930s. Similarly the 13 Tests which were played during the current year are clubbed with the 2000s decade.

Let me first explain the two 50+ averages. The very high average at BP3 during the 1930s is solely because of Bradman's 98 average until end of 1939. The 50+ average during 1940s at BP4 is mainly because of the 50+ averages in this position of Hassett, Compton and Hammond. Morever only 44 Tests were played during this decade.

My thanks to Abhi for an excellent suggestion. A few comments, not necessarily a complete list. Readers can add their own observations.

- Barring the Bradman-centric 1930s, the 2000s have had the best averages in the positions, BP3, BP4 (again ignoring the 1940s with only 44 Tests), BP5 and BP7. Truly a batsmen-dominated decade.
- The best Opening figures have been during 1970s (Gavaskar, Boycott, Lawry, Glenn Turner et al).
- The best BP6 figures have been during the 1980s (led by Border).
- The change from 1990s to 2000s is truly amazing. A 10% increase in overall average value.
- Note also the very high BP7 average of the 2000s.
- It can also be seen that BP4 has a higher overall average than BP3. This is a slight deviation from the earlier discussions.
- Note also the discernible correlation between the Opening average and the overall average.

2. Now for a table which I thought of to provide additional insight to the way an individual batsman has batted. I have identified the top 3 favourite batting positions of batsmen based on runs scored and created a table of runs scored, % of total runs, batting average in this position and a comparison to the overall batting average.

SN0 Batsman           Top Bat position   Next Batpos     Third Batpos
Pos Runs   Avge  Pos Runs   Avge  Pos Runs  Avge
%Car ToBtAvg    %Car ToBtAvg     %Car ToBtAvg

1.Tendulkar S.R | 4:11239- 57.34 | 5:1331- 55.46 | 6: 745- 43.82 | 84% 1.03 | 10% 1.00 | 6% 0.79 2.Lara B.C | 4: 7535- 51.26 | 3:3749- 60.47 | 5: 536- 41.23 | 63% 0.97 | 31% 1.14 | 4% 0.78 3.Ponting R.T | 3: 9417- 59.60 | 6:1989- 49.72 | 7: 208- 26.00 | 79% 1.08 | 17% 0.90 | 2% 0.47 4.Dravid R | 3: 8970- 55.71 | 4: 957- 53.17 | 1: 489- 32.60 | 79% 1.04 | 8% 0.99 | 4% 0.61 5.Border A.R | 4: 3792- 49.89 | 5:3062- 52.79 | 6:2556- 52.16 | 34% 0.99 | 27% 1.04 | 23% 1.03 6.Waugh S.R | 5: 6754- 56.28 | 6:3165- 51.05 | 7: 543- 33.94 | 62% 1.10 | 29% 1.00 | 5% 0.66 7.Kallis J.H | 4: 6943- 61.99 | 3:3335- 49.78 | 5: 409- 37.18 | 64% 1.13 | 31% 0.91 | 4% 0.68 8.Gavaskar S.M | 1: 9607- 50.30 | 4: 236-236.00 | 5: 144- 36.00 | 95% 0.98 | 2% 4.62 | 1% 0.70 9.Jayawardene M | 4: 7290- 59.75 | 5: 897- 33.22 | 3: 798- 49.88 | 80% 1.11 | 10% 0.62 | 9% 0.92 10.Gooch G.A | 1: 7811- 43.88 | 5: 419- 32.23 | 3: 347- 43.38 | 88% 1.03 | 5% 0.76 | 4% 1.02 11.Javed Miandad | 4: 6925- 54.10 | 5:1468- 54.37 | 6: 221- 24.56 | 78% 1.03 | 17% 1.03 | 3% 0.47 12.Inzamam-ul-Haq | 4: 4867- 52.90 | 5:2144- 51.05 | 6: 887- 36.96 | 55% 1.07 | 24% 1.03 | 10% 0.75 13.Chanderpaul S | 5: 4409- 52.49 | 6:2235- 65.74 | 3: 925- 34.26 | 51% 1.08 | 26% 1.35 | 11% 0.70 14.Hayden M.L | 1: 8626- 50.74 | | | 100% 1.00 | | 15.Richards I.V.A | 3: 3508- 61.54 | 5:2720- 47.72 | 4:1566- 41.21 | 41% 1.23 | 32% 0.95 | 18% 0.82 16.Stewart A.J | 1: 3348- 44.64 | 6:1421- 34.66 | 3:1307- 43.57 | 40% 1.13 | 17% 0.88 | 15% 1.10 17.Gower D.I | 4: 3223- 38.37 | 3:2619- 49.42 | 5:2131- 49.56 | 39% 0.87 | 32% 1.12 | 26% 1.12 18.Boycott G | 1: 8091- 48.16 | 4: 23- 11.50 | | 100% 1.01 | 0% 0.24 | 19.Sobers G.St.A | 6: 2614- 53.35 | 5:1895- 59.22 | 4:1530- 63.75 | 33% 0.92 | 24% 1.02 | 19% 1.10 20.Waugh M.E | 4: 6662- 42.43 | 5: 700- 35.00 | 6: 589- 53.55 | 83% 1.01 | 9% 0.84 | 7% 1.28 ... ... 37.Bradman D.G | 3: 5078-103.63 | 6: 681- 97.29 | 4: 485- 53.89 | 73% 1.04 | 10% 0.97 | 7% 0.54

I am not going to make many statements. Let the readers do the talking.

Just to explain something. First ignore the opening batsmen like Hayden who has a perfect 100% in his opening position. Take Tendulkar and Lara. Tendulkar has scored 84% of his runs in a single batting position indicating a reasonably settled career. In fact he has the highest top position % amongst all top batsmen. On the other hand Lara has moved between BP3 and BP4 quite a lot. Note also how much of a movement Border, Chanderpaul, Inzamam and Sobers have had.

Note also the very high batting averages of batsmen like Steve Waugh, Kallis, Richards and Jayawardene have had in their favourite batting positions. Contrast this with Border, Sobers and Lara who have paid for their movements with below-par average in their batting position.

Bradman's distribution does not show any surprise other than the very low (you are kidding !!!) average of 53.89 at BP4, possibly during the body-line series, if I am not mistaken.

To view/down-load the complete tables, please click/right-click here.

3. This is a table asked for by Marees who wanted a summarized analysis of the 9-10-11 positions. Certain criteria. First the batsman must have scored a minimum of 500 runs in these three positions. The other added criteria is that the batsman should have scored over 50% of his career runs in these 9-10-11 positions. This is to prevent players with higher level batting qualifications, such as Shaun Pollock (534 runs at 41.08), Oldfield (658 runs at 34.63) and Vaas (804 runs at 26.80) et al.

Batsman           Team   BPA  Runs  Inns NO   Avge % of total

Swann G.P Eng 9.00 507 18 4 36.21 90.1% More K.S Ind 8.33 693 34 9 27.72 53.9% Allen D.A Eng 8.63 511 29 10 26.89 55.7% Boje N Saf 8.10 804 37 7 26.80 61.3% Pollock P.M Saf 9.29 509 33 11 23.14 83.9% Lee B Aus 8.72 1122 69 16 21.17 77.3% Tayfield H.J Saf 8.58 577 41 8 17.48 66.9% Cairns B.L Nzl 9.02 737 51 8 17.14 79.4% Gillespie J.N Aus 8.86 867 76 25 17.00 71.0% Edmonds P.H Eng 8.80 514 43 12 16.58 58.7% Sarfraz Nawaz Pak 8.97 824 57 7 16.48 78.9% Abdul Qadir Pak 8.66 603 45 8 16.30 58.6% Hall W.W Win 9.65 784 64 14 15.68 95.8% Doull S.B Nzl 9.76 562 47 11 15.61 98.6% Harbhajan Singh Ind 9.11 997 81 17 15.58 62.9% Srinath J Ind 9.55 932 79 19 15.53 92.4% Mohammad Rafique Bng 8.75 574 41 4 15.51 54.2% Trueman F.S Eng 9.26 868 72 14 14.97 88.5% Collinge R.O Nzl 10.10 533 50 13 14.41 100.0%

Swann's position as an outstanding low-order batsman is blostered by this table. Brett Lee's 1000+ runs at an average of 21+ is also quite commendable.

4. The final table is one asked by Unni. He wanted a table on batting position value weighted by the runs scored in that particular position. I will present this table without comments and let Unni have his say.

To view/down-load the complete tables, please click/right-click here.

SNo Batsman         Team  Inns   BPA    Runs WtBPA(R)

1.Tendulkar S.R Ind 271 4.28 13447 4.23 2.Lara B.C Win 232 3.78 11953 3.71 3.Ponting R.T Aus 243 3.82 11924 3.61 4.Dravid R Ind 240 3.19 11395 3.26 5.Border A.R Aus 265 4.70 11174 4.65 6.Waugh S.R Aus 260 5.42 10927 5.33 7.Kallis J.H Saf 231 3.81 10843 3.76 8.Gavaskar S.M Ind 214 1.26 10122 1.20 9.Jayawardene M Slk 182 4.10 9123 4.04 10.Gooch G.A Eng 215 1.46 8900 1.38 11.Javed Miandad Pak 189 4.22 8832 4.21 12.Inzamam-ul-Haq Pak 200 4.65 8830 4.62 13.Chanderpaul S Win 210 4.80 8669 4.97 14.Hayden M.L Aus 184 1.00 8626 1.00 15.Richards I.V.A Win 182 4.16 8540 3.93 16.Stewart A.J Eng 235 3.58 8465 3.30 17.Gower D.I Eng 204 3.98 8231 3.97 18.Boycott G Eng 193 1.03 8114 1.01 19.Sobers G.St.A Win 160 5.04 8032 4.92 20.Waugh M.E Aus 209 4.24 8029 4.22

To view/down-load the complete tables, please click/right-click here.

The curtain rings on a fascinating subject in which the reader comments have been very illuminating. My thanks to all of them.

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

RSS Feeds: Anantha Narayanan

© ESPN Sports Media Ltd.

Posted by Ramaswami Subramanyan on (May 29, 2010, 3:31 GMT)

Dear Mr. Ananthanarayanan,

First, thank you - truly wonderful and enjoyable columns.

I have been trying to some number crunching myself - but I am not sure how to download information from statsguru to study the results. Please do write a column on the techniques to so so. If you prefer, an email to me would be fine too...

Thanks, RV

P.S. Please do respond either way. :-)

Posted by Request on (May 11, 2010, 4:46 GMT)

Can we have an analysis done quality of runs and wickets earned by batsmen and bowlers respectively. The quality being based on "where" they were earned. I am not a fan of accumulation, if you play long enough you will get x amount of runs and wickets I prefer to add some weight based on where they were earned. It's one thing to get loads of runs at Adelaide, Sydney, Capetown etc and another to get runs at Brisbane, Melbourne, Trinidad etc.

Posted by Mub on (May 3, 2010, 21:17 GMT)

Hi Anantha,

Regarding the batting position functions, what I thought was, if we could get the average of the partnership of each wicket and then take an average of the instances of those wicket partnerships (meaning we won't have that many instances of a 9th wkt partnership as the 3rd wkt partnership), we might arrive at the typical Test Match. I guess that would give some indications of at least what each position is doing as opposed to what we think it should be doing, wouldn't it?

Posted by Ananth on (May 2, 2010, 3:04 GMT)

Dear readers I started writing for "It Figures" during late 2007 and since then have done a total of nearly 100 posts without a break. This is the one project from which I derive a lot of pleasure and satisfaction. The exchanges from readers all over the globe makes the hours I spend well worth. Of late the blog takes a lot of time possibly because I am the type who reads each and every comment and responds to at least 50% of these. There is no doubt that the blog is made because of this. In order to avoid getting stale and being repetitive I need a short break. It is not that I am short of ideas or reader suggestions. I have at least a dozen such analysis possibillities on my plate. Hence I would like to inform the readers that I will take this break during the month of May. My next article will be at the beginnin g of June when I return hopefully refreshed. Of course, I will continue to answer the reader comments. Ananth

Posted by Ron on (May 1, 2010, 7:20 GMT)

These tables partially explain Tendulkar's sub 60 average. The batsmen with a greater or equal part of their careers in the '90s as opposed to the 2000s would all have seen their averages impacted. [[ Ron Not just Tendulkar's but Lara's who spent more than his career during the 1990s. Ananth: ]]

Posted by Boll on (April 30, 2010, 17:06 GMT)

...but Bradman made these sorts of comparisons irrelevant, by scoring so many more runs per innings than anyone else. In the words of Bill Woodfull, ` for 20 years he was worth three men to the Australian side`. Or the great Walter Hammond `None of us had the measure of him, and that`s the plain fact.`For mine, SRT has been the best amongst equals, DGB has made any argument about the greatest of all time superfluous.

Posted by Mub on (April 29, 2010, 23:02 GMT)

Hi Anantha, This might be a bit off-thread, but regarding the batting positions, I suppose each position has its attendant function/responsiblity. Openers are expected to negotiate the tricky conditions so that the better batsmen to follow at 3/4 make the runs, etc. What I was wondering is, is it possible to find a Test(s) that played out true to these roles? And then maybe find out if the batting positions are really designed truly or there are huge deviations. I don't know if this is even possible, but just a thought... [[ Mub Unfortunately these parameters are not clearly definable. It is not easy to set the parameters for analysis. Ananth: ]]

Posted by Steve on (April 29, 2010, 2:27 GMT)

We talk about how individual batsmen have impacted the stats. One wonders how much effect great bowlers have had on these tables. The West Indies & Lillee of the 80s, McGrath, Donald and Warne of the 90s, Akram & Yunis etc I'm sure McGrath and Akram had their impact on the opening and higher order slots and Warne certainly caused headaches for the middle order. The great West Indies of the 80s stitched up just about every position LOL

Posted by Sandeep on (April 28, 2010, 15:51 GMT)

How many innings did Bradman actually play at no 4? Perhaps that was a mistake on Bradman's part- he should have batted at no 3, and would have ended up with a career average of 100+! [[ Sandeep You can determine that yourself. He played 9 dismissed innings. Anyhow one four in his last innings would have given him an average of 100.00. For that matter if he played in a later test and retired at 4, he would have reached 100. I don't think he bothered about this at all. Ananth: ]]

Posted by Murali on (April 28, 2010, 14:06 GMT)

Just want to know. If Table 1 is split into two columns - average of batsmen at each position (winning team) vs. average of batsmen at each position (losing team). probably it would indicate how much some teams (wi or aus) dominated during certain decades and scored mountains of runs. [[ Murali This table is at the highest macro level. As such it is certain that if the columns are split into "winning" and "losing", nothing much will be revealed. The "winning" average will necessarily be higher than the "losing". We will only get an insight into the difference in % at each position level. On the other hand a broader analysis, bringing in a third dimension of Team would reveal more since Bangladesh will get into a separte plane thus not diluting the Australian figures. What you have suggested is certainly food for thought, an euphemism for adding on more to the plate. Ananth: ]]

Comments have now been closed for this article

ABOUT THE AUTHOR

Anantha Narayanan
Anantha spent the first half of his four-decade working career with corporates like IBM, Shaw Wallace, NCR, Sime Darby and the Spinneys group in IT-related positions. In the second half, he has worked on cricket simulation, ratings, data mining, analysis and writing, amongst other things. He was the creator of the Wisden 100 lists, released in 2001. He has written for ESPNcricinfo and CastrolCricket, and worked extensively with Maruti Motors, Idea Cellular and Castrol on their performance ratings-related systems. He is an armchair connoisseur of most sports. His other passion is tennis, and he thinks Roger Federer is the greatest sportsman to have walked on earth.

All articles by this writer