THE CORDON HOME

BLOGS ARCHIVES
SELECT BLOG
June 25, 2010

Batting

Significant Test innings, and their architects: a follow-up

Anantha Narayanan

Brian Lara has a high Significant Innings percentage of 45.69 © AFP

A few days back I had come out with an article on the significant innings in Test cricket. It received, arguably, the best responses I have received for any of my articles in this web log. The readers appreciated that there is a completely new measure to evaluate Test innings. The fact that it was off the beaten track of averages, centuries, strike rates et al was a very important factor. The comments and suggestions were some of the best I have received and I was determined to come out with the follow-up article sooner than later.

I will summarise the changes below.

1. As many readers have suggested, I have used the innings as the basis for determining the significant innings rather than the two team innings together. This takes care of the many Tests where the two innings by a team are as different as chalk and cheese. If we take the famous Calcutta Test of 2001, the two Indian innings were 171 and 657. The 59 in the first innings was an outstanding innings considering the 171 for 10 as the basis, probably not if we take 728 for 17 as the basis.

2. This is one lapse which was missed by all the readers. And for that matter myself. In the base analysis, I had taken the wickets as the base for determining the runs and balls cut-off. This is quite wrong. I should have taken the number of batsman who batted as the basis. Take the West Indian innings of 790 for 3. The base should be 5 (which includes Sobers) and not 3 (the number of wickets). If a team is all out, the base will be 11. Of course batsmen who did not bat will be excluded, but batsman who retired hurt will be included. This is absolutely the correct method.

3. Raise the multiplier values for two reasons. One is the consideration of innings as the base and the other is the taking of batsmen as base rather than wickets. I have also introduced a graded multiplier. The multiplier is highest at 2.00 for low rpb/bpb (runs per batsman and balls per batsman) values for 1-7 batsmen and stays at 1.00 for high rpb/bpb values ford 8-11 batsmen. The capping of run-cutoff at 100 and balls-cutoff at 200 is also retained.

4. I will ignore all not out innings below 10, if they have already not become SI, from the total innings. This is a very relevant suggestion. This is necessary since quite a few batsmen, especially in the late order and in later innings remain unbeaten on low scores. Since they have not been given the opportunities to further their innings, these innings are excluded from the total.

5. Now that we have the single innings as the base and have raised the cut-off values, there is no need to have the one-third criteria. Even in the 26 by New Zealand, the 11 by Sutcliffe does not really warrant being considered as a SI. On the other hand, Hutton's 30 out of 52, Tancred's 26 out of 47 and Flintoff's 24 out of 51 must be included. This is done by keeping the lower limit for runs cut-off as 20.

6. Finally one very important addition. I have done a weighting of the innings by determining a Situation innings index value. A 100 out of 200 and a 100 out of 500 are both significant innings. However the first innings is far more significant than the later. This measure indicates the extent of significance. It is possible that this factor can very well be used to determine the influence of batsmen. So there is an additional table based on the average SI Index value. The SI Index value is a simple calculation. The innings measure, runs or balls, is divided by the runs cut-off or balls cut-off, as required. Thus the minimum value for this is 1. Where a player has crossed both cut-offs, the higher index value is taken.

Let me conclude this section by saying that the user responses have been outstanding revealing a very incisive way of thinking. Let us now look at the tables now.

First the table of players, ordered by the % of SIs played. This is a reflection of the consistency of the players. Players such as Dravid, Border et al are likely to be at the top. They are likely to score two 75s in successive innings.

List of players, ordered by the % of SIs achieved

SNo Batsman              Cty Mats  Runs Inns  SIs % of SI

1.Bradman D.G Aus 52 6996 80 40 50.00 2.EdeC Weekes Win 48 4455 81 38 46.91 3.Hobbs J.B Eng 61 5410 101 47 46.53 4.Barrington K.F Eng 82 6806 129 59 45.74 5.Lara B.C Win 131 11953 232 106 45.69 6.Dravid R Ind 139 11395 236 106 44.92 7.May P.B.H Eng 66 4537 105 47 44.76 8.Sutcliffe H Eng 54 4555 83 37 44.58 9.Hutton L Eng 79 6971 137 61 44.53 10.Chanderpaul S Win 124 8710 210 93 44.29 11.Hammond W.R Eng 85 7249 137 60 43.80 12.Younis Khan Pak 63 5260 111 48 43.24 13.Gavaskar S.M Ind 125 10122 211 91 43.13 14.Umrigar P.R Ind 59 3631 93 39 41.94 15.Flower A Zim 63 4794 110 46 41.82 16.Compton D.C.S Eng 78 5807 129 53 41.09 17.Kallis J.H Saf 138 10911 232 95 40.95 18.Javed Miandad Pak 124 8832 187 76 40.64 19.Richards I.V.A Win 121 8540 180 73 40.56 20.Tendulkar S.R Ind 166 13447 269 107 39.78

The top three remain the same. A few minor changes down the table. Chanderpaul moves down a few places. Sutcliffe also moves down. Lara, Dravid and May move up. Andy Flower moves down a few places.

The most significant change is that of Tendulkar who moves up quite a few places into the top-20 table.

Now the table of players, ordered by the % of SIs played. This is a reflection of the extent of significance once the cut-off is reached. This is likely to have players like Sehwag, Lara et al at the top. They are likely to score a 150 and 0 in two successive innings.

List of players, ordered by the average SI index value

SNo Batsman              Cty Mats  Runs Inns  SIs SII   Avge
Pts   SII

1.Bradman D.G Aus 52 6996 80 40 92 2.320 2.Sehwag V Ind 76 6691 129 40 80 2.009 3.Hanif Mohammad Pak 55 3915 95 33 65 1.995 4.Sangakkara K.C Slk 88 7549 146 52 102 1.965 5.Lara B.C Win 131 11953 232 106 207 1.959 6.Pietersen K.P Eng 62 5166 111 35 66 1.914 7.Amiss D.L Eng 50 3612 86 27 51 1.897 8.Flower A Zim 63 4794 110 46 87 1.892 9.Cullinan D.J Saf 70 4554 111 33 61 1.877 10.Crowe M.D Nzl 77 5444 129 42 78 1.865 11.Walcott C.L Win 44 3798 73 29 54 1.863 12.Atapattu M.S Slk 90 5502 149 40 74 1.853 13.Mitchell B Saf 42 3471 79 30 55 1.844 14.Ijaz Ahmed Pak 60 3315 92 25 45 1.836 15.Hill C Aus 49 3412 89 30 54 1.833 16.Saeed Anwar Pak 55 4052 91 35 64 1.832 17.Asif Iqbal Pak 58 3575 97 28 51 1.832 18.Gomes H.A Win 60 3171 87 21 38 1.830 19.Harvey R.N Aus 79 6149 134 48 87 1.821 20.Gooch G.A Eng 118 8900 215 75 136 1.820

The batsman non pareil, Bradman has an average SI Index value of 2.27. Then comes Sehwag, as expected. His string of high scores have propelled him to this second position. Now there is a surprise. Hanif Mohammad, the chalk to cheese (or vice versa) of Sehwag, closely follows Sehwag. His third position indicates how under-rated the great little master was. What he did for Pakistan cricket is unbelievable. That too on difficult pitches and often away. Now come two modern greats, Lara and Sangakkara. This confirms their penchant for out-performing often.

I have given below the best three innings as far as the SI Index is concerned. The first one is the Asif Iqbal classic. During 1967, Pakistan scored 216 in the first innings. England replied with 440. Then Pakistan slumped to 65 for 8. Asif Iqbal then played the greatest of all late order innings and one of the best ever. He added 190 with Intikhab Alam and took the total to 255. England won comfortably. Asif Iqbal’s innings has the highest SI index value ever of 5.41, based on a runs-cutoff value of 27.7 (255/11, multiplied by a factor of 1.333 (no 8-11) and adjusted downwards by 10% for being the second innings).

Pakistan 2nd innings
+Wasim Bari                              b Titmus              12
Mohammad Ilyas        c Cowdrey          b Higgs                1
Saeed Ahmed           c Knott            b Higgs                0
Majid Khan                               b Higgs                0
*Hanif Mohammad       c Knott            b Higgs               18
Ghulam Abbas          c Knott            b Higgs                0
Mushtaq Mohammad      c D'Oliveira       b Underwood           17
Javed Burki                              b Underwood            7
Asif Iqbal            st Knott           b Close              146
Intikhab Alam                            b Titmus              51
Saleem Altaf          not out                                   0
Extras                (b 1, lb 1, nb 1)                         3
Total                 (all out, 101.1 overs)                  255
FoW: 1-1, 2-5, 3-5, 4-26, 5-26, 6-41, 7-53, 8-65, 9-255, 10-255.
The next is one is another all-time great innings by Dennis Amiss. During 1974, in Kingston, England started their second innings, 230 in arrears. Amiss opened the innings, remained unbeaten on 262 and guided England to safe total of 432 for 9. This innings is reminiscent of the Laxman classic. Amiss' innings has the second highest SI index value ever of 4.52, based on a runs-cutoff value of 39.3 (432/11, multiplied by a factor of 1.667 (no 1-7) and adjusted downwards by 10% for being the second innings).
England 2nd innings
G Boycott             c Murray           b Boyce                5
DL Amiss              not out                                 262
JA Jameson            c Rowe             b Barrett             38
FC Hayes              run out                                   0
*MH Denness           c Rowe             b Barrett             28
AW Greig                                 b Gibbs               14
DL Underwood          c Murray           b Sobers              12
+APE Knott            run out                                   6
CM Old                                   b Barrett             19
PI Pocock             c sub              b Boyce                4
RGD Willis            not out                                   3
Extras                (b 10, lb 11, w 1, nb 19)                41
Total                 (9 wickets, 183 overs)                  432
FoW: 1-32, 2-102, 3-107, 4-176, 5-217, 6-258, 7-271, 8-343, 9-392.
Now a modern classic by Saeed Anwar. During 1999, in Calcutta, Pakistan started their second innings, 38 in arrears. Anwar opened the innings, remained unbeaten on 188 and guided Pakistan to good total of 316, with the Pakistani bowlers dismissing India for 232. Anwar's innings has the third highest SI index value ever of 4.48, based on a runs-cutoff value of 28.7 (316/11, multiplied by a factor of 1.667 (no 1-7) and adjusted downwards by 10% for being the second innings).
Pakistan 2nd innings                                            R   M   B  4 6
Saeed Anwar           not out                                 188 452 259 23 1 4
Wajahatullah Wasti    c Mongia           b Srinath              9  54  33  2 0
Saqlain Mushtaq       c Mongia           b Harbhajan Singh     21 108  86  1 0
Ijaz Ahmed            c Mongia           b Srinath             11  55  47  1 0
Yousuf Youhana        c Dravid           b Srinath             56 139 123  7 1 2
Shahid Afridi         c Laxman           b Srinath              0   1   1  0 0
Saleem Malik          lbw                b Srinath              9  34  16  1 0
+Moin Khan            c Mongia           b Prasad               8  22  13  1 0
Azhar Mahmood         lbw                b Srinath              0   9   9  0 0
*Wasim Akram          c Mongia           b Srinath              1   7   3  0 0
Shoaib Akhtar                            b Srinath              1  14   8  0 0
Extras                (lb 3, w 5, nb 4)                        12
Total                 (all out, 99 overs)                     316
FoW: 1-26 (Wajahatullah Wasti, 10.5 ov), 2-94 (Saqlain Mushtaq, 35.3 ov), 3-147
(Ijaz Ahmed, 49.1 ov), 4-262 (Yousuf Youhana, 82.3 ov), 5-262 (Shahid Afridi, 82.4 ov),
6-284 (Saleem Malik, 88.4 ov), 7-301 (Moin Khan, 93.1 ov), 8-302 (Azhar Mahmood, 94.6 ov),
9-304 (Wasim Akram, 96.2 ov), 10-316 (Shoaib Akhtar, 98.6 ov).
To view/down-load the complete player table, ordered by the % of SIs played, please click/right-click here.

To view/down-load the complete player table, ordered by the average values of SI Index, please click/right-click here.

I have also made available the complete list of significant performances for all the 159 qualifying batsmen.

To view/down-load the table for all the first 1960 tests, please click/right-click here.

Finally the grand-daddy of all tables. Let me warn you these tables are huge, 500kb each. These are the lists of all significant innings, all 14782 of them, covering all 1960 tests played.

To view/down-load the complete table for tests 1-999, please click/right-click here.

To view/down-load the complete table for tests 1000-1960, please click/right-click here.

A few readers have asked for some summarized figures based on criteria. I have given these, and more below. I have not done the %. I leave it for the readers.

Summary information
===================

TotInns:68988 TotInnsSel: 64964 Perfs: 14782 100+runs: 3374 50+runs: 6581 <50runs:4827 200+balls: 1233 100+balls: 2545 <100balls: 2062 BPos 1-7: 12835 BPos 8-11: 1947 Both: 3053 Rpw: 11141 Bpw: 588 1Inns: 8678 2Inns: 6104 Wins: 4433 Losses: 5678 Draws: 4671 SI1: 5.41 SI2: 4.52 SI3:4.48

I will attempt to do a significant innings analysis for ODIs later as also, possibly more complex, a significant innspell analysis for Tests.
List of selected players ordered by the average SI index value

Batsman Cty Mats Runs Inns SIs SII Avge Pts SII

Headley G.A Win 22 2190 39 18 46.15 2.214 Pollock R.G Saf 23 2256 41 15 36.59 2.033 Nurse S.M Win 29 2523 54 16 29.63 2.026 Turner G.M Nzl 41 2991 72 27 37.50 1.936 Hazare V.S Ind 30 2192 52 18 34.62 1.912 Ponsford W.H Aus 29 2122 47 11 23.40 1.896 Nourse A.D Saf 34 2960 62 29 46.77 1.842 Gambhir G Ind 31 2798 53 18 33.96 1.812 Mankad M.H Ind 44 2109 70 16 22.86 1.812 Macartney C.G Aus 35 2131 53 16 30.19 1.781 Taylor H.W Saf 42 2936 76 34 44.74 1.774 McCabe S.J Aus 39 2748 61 20 32.79 1.760 Rowe L.G Win 30 2047 48 11 22.92 1.747 Richardson M.H Nzl 38 2776 64 31 48.44 1.672 Rowan E.A.B Saf 26 1965 49 21 42.86 1.665 O'Neill N.C Aus 42 2779 67 23 34.33 1.565 Dhoni M.S Ind 43 2428 66 18 27.27 1.415

This is a selected set of players whose career runs are between 1965 and 3000. This list has been requested for by John Clark. I have selected a few players including Mark Richardson, in view of Gabriel's recent article.

Headley almost touches Bradman. The other great, Greame Pollock, also crosses 2.00.

On 29 June 2010

As requested by Abhi and Alex I have expanded the Player tables with the following information.

1. Add number of fifties and % of selected inns to enable comparison with SI %.
2. Runs per innings for significant innings.
3. Total of SI Runs and % of total career runs.

I have also corrected the format of the Selected players Si report to enable proper downloading into XL files.

To view/down-load the complete revised player table, ordered by the % of SIs played, please click/right-click here.

To view/down-load the complete revised player table, ordered by the average values of SI Index, please click/right-click here.

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 K Soundararajan on (August 5, 2010, 22:30 GMT)

Dear Ananth,

I enjoy reading your columns very much. I'm a mathematician, and a statistic which has emerged recently in scientific circles is the so-called "h-index." A scientist has the h-index, k if he has k papers each of which has been cited at least k times, and k is the maximum number for which this holds.

I wonder if this is a reasonable statistic to consider for batsmen. For example, Tendulkar would have an h-index of 72 meaning that he has scored more than 72 seventy two times.

There are some appealing things about this index. It is simple to understand, always increases over the batsman's career, and doesn't privilege a century over a 99. Also consistency and longevity are rewarded. There are of course also negatives: the h-index cannot be more than the total number of innings, and so Bradman has only an index of about 40.

It may be interesting to see how different batsmen compare with this index.

Best regards,

Sound. [[ Soundar This is a most fascinating mail. However it is like facing a "doosra" "carrom" ball delivered by Paul Adams. Have moved your mail to the retention folder and will do this at leisure. Have already saved the Wikipedia entry on h-index for study at leisure. Thanks a lot for opening up a new line of thinking. Ananth: ]]

Posted by Praveen on (July 27, 2010, 19:03 GMT)

Yes i truely agree the below about Lara

His form was also very good in 2007. One can never forgive some of the players in the current team, the board members and some 'esteemed' commentators who worked together to get Lara out. One thing is certain. Tendulkar is more fortunate. The adulation he receives in India is in sharp contrast to Lara's case. Tendulkar can choose when and how he retires.

LARA WAS PUSHED OUT OF THE TEAM BECAUSE OF INTERNAL POLITICS,,

SACHIN KNOWS HOW TO PLAY IT SAFE,, he is a true politician.

Posted by Alex on (July 6, 2010, 6:14 GMT)

Ananth - re Greg Chappell. He scored 50+ runs 55 times. Yet, # SI for him is just 51 (much like 40 vs 42 for Bradman). I haven't followed his career avidly but recall him as a true all-time great. G. Armstrong in his book (100 Great Players: 2009 revised ed) has argued for Chappell as follows. Let A be the set of players that played at least 30 test innings during Chappell's career. Then, only 8 of these players (3 of which are Aussies: Walters @45 & Border @50) average more than 45 with Chappell leading them at 53+. Your analysis shows that some of the average boosting 50's possibly came in tall scoring matches.

A request - could you please do a similar analysis for an entire series? Many remember Lara in SL '01 but how many remember Border averaging 74 vs the ferocious WI in '85 when no other batsman scored even half as many runs?

Posted by craigmnz on (July 5, 2010, 20:01 GMT)

Ananth - actually more like a phrase 'was a better batsman' is what's missing from the first sentence. Guess I just got carried away in my excitement!

You're quite right about cut-offs, of course, but it must be highly tempting to set the bar at 2000 runs thus including Headley, Pollock (R.G.), Ponsford and MacCartney. This would naturally expand the field to include any number of lower-order batsman of dubious claims to batting greatness so I can see why you make these decisions. You almost need a sliding cut-off based on the amount of cricket available to the various eras.

I like to think of this as the 'SF Barnes' factor - I'm sure many of the bowling analyses I've seen begin at 200 Test wickets just to avoid the skewing effect of Barnes's career. [[ Craig I have done quite a number of analysis having 2000 as the cut-off and more often than not Headley has come second to Pollock. It is just that there are fo more batsmen, 159 vs 255. The tables become quite large. And think of poor Tom Hayward who clocks in at 1999 runs. In fact I did a couple of bowling analysis with 189 as the cut-off. However in bowling 100 wickets seems the right number. Ananth: ]]

Posted by craigmnz on (July 5, 2010, 5:44 GMT)

Ananth: Thank you for a statistical measure that supports my long-cherished belief that Glenn Turner than Martin Crowe. Yes, I saw both of them bat at various stages of their careers and in my honest moments I've suspected that I'm favouring the hero-worship of adolesence over the more jaundiced eye of adulthood, but at last something to confirm my prejudices.

I am particularly struck by the way this measure seems to be timeless - neither weighted in favour of the high-scoring present or of any particular previous era. Well done.

In a more tongue-in-cheek comment - you do realise that you've produced a stat that places Mark Richardson second only to Don Bradman - when you order by % of SI and include the batsmen with under 3000 runs. That's got to be a worry! [[ Craig Your first sentence seems to miss a keyword. If we did not have a cut-off, Bradman will not be first in the batting average list, but second to Ganteaume whose average is 112.00 and will be ahead, only by a fraction of run, of M.N.Nawaz of Sri Lanka whose average is 99.00. At least Mark is a competent opening batsman !!!

Ananth: ]]

Posted by Alex on (July 2, 2010, 1:09 GMT)

Boll: Richie Benaud rated (see youtube) PBH May as the best England batsman since WW2, adding a proviso that he viewed Hutton & Compton as pre-WW2. Also striking is the similarity in the numbers of these pairs that continues to hold in SI analysis as well: Hobbs & Barrington, SRT & Sobers, Lloyd & Border ... the more things change, the more they tend to remain the same. Here is hoping Lara plays in and lights up IPL4! [[ Alex Lara at IPL4 ??? Forget about it. More than that I wished he had played test cricket for 2 more years. His form was also very good in 2007. One can never forgive some of the players in the current team, the board members and some 'esteemed' commentators who worked together to get Lara out. One thing is certain. Tendulkar is more fortunate. The adulation he receives in India is in sharp contrast to Lara's case. Tendulkar can choose when and how he retires. Ananth: ]]

Posted by Boll on (July 1, 2010, 21:42 GMT)

@Alex and Ananth. yeah, very interesting comparisons there. I was also surprised to see Peter May perform so well based on these criteria. I think I`ve always placed him alongside people such as Mark Waugh - the ultimate stylist for many Aussies of my generation, with a good, if not quite great average, and prone to scoring eye-catching, fluent 60s and 70s rather than match-changing/saving hundreds. Both players seem to attract the `boys-own` type of journalism - grace, timing, how we all wished we could bat etc. very complimentary, but also implying, I think unfairly, a lack of `steel within` as you put it Ananth. May`s place on this list, alongside (well, just above) more acknowledged greats such as Sutcliffe, Hutton and Hammond, indicates rare quality and purpose, to go with his oft-quoted style.

Posted by Alex on (July 1, 2010, 2:01 GMT)

Ananth: thanks - I was deluded by Lara's high Ave SIIdx.

1. Lara's & Chander's high SI % clearly underscores the weak WI batting: Chander is very effective but has SI % of 44.29 despite only 72.9 RPI in SI. His spiritual predecessor was Gomes, an important cog in the champion WI teams. However, Gomes has SI % of only 24.14 at a much higher 85 RPI in SI.

2. Note the similarity in the number of Dravid and PBH May.

3. Also, high RPI in SI with a relatively low % of SI suggests that the batsman attacked & scored heavy when others in his team scored well too. Holds true for Sehwag and Ponting. These tables illustrate their contrast with Viv, a similar type batsman, who scored less heavily in SI but got those runs when his team-mates did not score as well. [[ Alex Very perceptive copmments. These add-ons allow interested readers to do some lateral thinking. May is another often under-rated batsmen. His "gentleman" like exterior hid a steel within, a la Dravid. Ananth: ]]

Posted by Abhi on (June 30, 2010, 13:58 GMT)

Ananth, Thanks. Read the bit for the fields in the database. So, will attempt to calculate the avg. for the SIs for myself!

Posted by Alex on (June 30, 2010, 3:55 GMT)

Ananth - thanks very much. I had requested a 40+ columns to smooth things out in case someone questions why 50+ scores are so special (why not 45+ etc.). These columns add a lot more info! Lara's SI RPI is surprisingly low at only 95.2; could you pl doublecheck? I would have guessed it to be >110. [[ Alex I am running out of fields in my Player database. Even for this I had to do a bit of jugglery. I think the 50+ should be sufficient. Re Lara, work out the total runs in the SI knocks for him. It comes out to 10090 runs. So in the other 126 innings he has scored only 1860 runs. Clearly indicating his propensity for 100 & 0. Anyhow even 100 is quite tough and has been reached only very few. If you ignore Bradman (and Sehwag) the next highest is 108. However all the 100+ batsmen have somewhat lower % of SIs. With Lara's quite high % of SIs, 110 woul;d take him beyond his career total. 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