HSI - A new and exciting measure for ODIs

A 100 as the top score does not provide enough information by itself. It could be out of a team score of 200 or 300. It could be supported by an innings close to 100, by a 50 or by a 25. It could be part of 200 for 1 or 200 all out. Most of the alternatives have been given above and the expectations have been given on the right. Everything is self-explanatory. I have explained the final methodology below.

The HSI is a measure of two components for the innings top score. The batsman stands alone at the top and his contribution gets enhanced depending on the support received. On the other hand the second-placed scorer has had the support of a higher-scoring batsman. So it is sufficient to take his and other lower scoring batsmen's contributions based on the team score. With this background let me show the working.

Top batsman HSI = Hs1/Team score x HSI/Hs2. This incorporates both components.
Other batsmen HSI = Batsman score/Team score.

I worked out that there is no need to multiply the lower scores by Score/Hs1. That would lower the values too much. An Hs1 of 100 and Hs2 of 90 (out of 200) would end up with Hs1 well over 25% higher than Hs2, which is incorrect.

Now let us see all the values and check whether the expectations are met.

Case 1: 200-100-30. Work: 100/200=0.5, 100/30=3.33
        HSI for Hs1: 0.5*3.33=1.667
        HSI for Hs2: 30/200=0.15
Case 2: 200-100-50. Work: 100/200=0.5, 100/50=2.0
        HSI for Hs1: 0.5*2.0=1.0
        HSI for Hs2: 50/200=0.25
Case 3: 200-100-90. Work: 100/200=0.5, 100/90=1.11
        HSI for Hs1: 0.5*1.11=0.55
        HSI for Hs2: 90/200=0.45
Case 4: 200-50-50.  Work:  50/200=0.25, 50/50=1.0
        HSI for Hs1: 0.25*1.0=0.25
        HSI for Hs2: 50/200=0.25
Case 5: 300-100-50. Work: 100/300=0.333, 100/50=2.0
        HSI for Hs1: 0.333*2.00=0.667
        HSI for Hs2: 50/300=0.1667
Case 6: 300-100-90. Work: 100/300=0.333, 100/90=1.11
        HSI for Hs1: 0.333*1.11=0.367
        HSI for Hs2: 90/300=0.300
Case 7: 300-50-40.  Work:  50/300=0.167, 50/40=1.25
        HSI for Hs1: 0.167*1.25=0.208
        HSI for Hs2: 40/300=0.1667

I have checked each HSI value and confirmed that it meets the expectations. Hence I will not go any further in depth. Readers can verify these numbers themselves. I also do a minor tweak for result matches where lower than 100 runs were chased down. Of course any unfinished innings below 100, in a no result match, is not considered.

Now that the HSI for every innings has been determined, let us move into the many tables I have created. The first is the basic table of the HSI value itself. I have shown the top 25 HSI values. There is a downloadable Excel file that contains the 24,000 innings which have HSI values greater than or equal to 0.1. Please download and peruse it before asking about specific innings or player.

Readers should remember that these calculations are scorecard-based, non-contextual and within a team. MS Dhoni's 65 out of 188, with Hs2 being 31, fetches an HSI of 0.724. Quinton de Kock's 135 out of 351, with Hs2 being 77, gets an HSI of 0.674. It does not mean that Dhoni's innings was better or match-winning. It only means that Dhoni contributed more to his team's cause. The result is immaterial. The key word is "contribution". All comparisons, within a match, should only be within a specific team innings. What is important is that Dhoni gets 0.724 and Kohli, 0.164. de Kock gets 0.674 and AB de Villiers, 0.218. Please make sure that this point is clearly understood.

An unambiguous note on the cut-off. I have selected 3000 ODI runs as the cut-off for the main table. There are 127 batsmen that qualify. Only one of these batsmen, Wasim Akram, has an average below 20.0 and I have decided not to exclude him. This cut-off has been determined on the assumption that a very good batsman would need around 100 matches to cross 3000 runs. In fact only 60 batsmen have reached this landmark in 100 matches. David Gower, Mahela Jayawardene, Kumar Sangakkara, Stephen Fleming, Richie Richardson et al needed more. Once this stiff cut-off is set, all players are considered equal. Afterwards I am not going to say one batsman played in only so many matches and another played in many more matches and so on. The players have met the criterion set and that is it.

For the other 12 tables, there are varying cut-off points. In general, 50 innings has been used as the minimum for qualification. However, readers should note that to qualify for the later tables only the appropriate cut-off is needed. In other words, Sunil Gavaskar would qualify for the "BatPos 1-2-3" table even though he has scored only 2651 career runs. Brad Haddin would qualify for the "Wins" table even though he has scored only 2692 career runs. Clive Lloyd would qualify for "BatPos 4-5-6-7" table even though he has scored only 1977 career runs. And so on.

After getting the HSI values I evaluated on the need to do an alternate mean-evaluation. I decided that it is not necessary to use GM and used AM itself since the distribution pattern revealed a few important facts. The top entry is at 9.9, the next one at 6.7, the third one at 5.2, the fourth one at 5.02 and the fifth one is at 4.2. See how steeply the values drop: less that 50% value for the fifth entry. So there is really a single outlier: Brendon McCullum's innings, which is from the Twilight Zone. I did not want to be influenced by this single performance. Andrew Jones could also be considered as an outlier. After this, normalcy returns. Also the values, 5.24 to 0.0, are already in some form of logarithmic representation, representing 400 to 1.

A few important facts on HSI.

1. The highest HSI value is 9.983 for McCullum's 80 out of 87(+8) for 0, with the next highest score being How's 7.
2. Two great 100s: Viv Richards' 189* and Kapil Dev's 175* are in third and fourth position in the HSI table. The key numbers for Richards are 189 out of 262, followed by 26. For Kapil, 175 out of 254, followed by 24. Two almost identical innings. And Shane Watson's 185 is placed sixth. Great credentials for this table indeed.
3. The lowest HSI value for an Hs1 innings is in match #257. New Zealand scored 116. Richard Hadlee and Derek Stirling were joint top scorers with 13 runs each. Their HSI value was a mere 0.126.
4. The highest HSI value for an Hs2 innings is for Jesse Ryder. In match #2677, New Zealand were chasing 158 and finished at 165 for 0. Ryder scored 79 and McCullum, 80. Ryder's HSI was 0.496 and McCullum's, 0.509.
5. The lowest HSI value for an Hs2 innings is for Martin Crowe's 5 runs in match #629. Crowe's HSI value is 0.075.
6. The highest HSI value for a non-Hs1-Hs2 innings was for Herschelle Gibbs in match 1760. This was a funny innings. South Africa was chasing 229 and scored 230 for 1. Lance Klusener top-scored with 75 (HSI 0.347), Boeta Dippenaar followed with 74 (0.337) and Gibbs' scored 70. His HSI was 0.319.
7. 969 HSI values are 1.0 and above. This represents 1.6% of the total.
8. 3329 HSI values are 0.5 and above. This represents 5.6% of the total.
9. 35548 HSI values are below 0.10. This represents 59.6% of the total.
10.The average Hs1 for 6736 team innings is 71.3. The average Hs2 for these innings is 46.3. The ratio is 1.57. However, the average of ratios taken at innings level is 1.65 which is the more relevant figure.
11.The average HSI value for the 59655 innings is 0.144. This average will also let us take a stand on career averages of HSI. Maybe 0.22 would an excellent career average.

Now for the multiple HSI tables based on various selection criteria. This was one of the main objectives of this exercise. For most tables I have shown the top 30/25 players. Needless to say (or more appropriately, needs to be said) that the complete set of entries is available in the downloadable file with 13 tables. Please make an attempt to answer your question by downloading that file before asking me. Since this is one of the longest articles I have ever penned, I will only provide minimal comments.

It is safe to say that McCullum's innings is a true outlier. Chasing 95, scoring 80 (in 28 balls) out of 95, reaching the target in about six overs, allowing Jamie How to score 7: well, it does happen, but once every 40 years. Only the 100-run tweak kept this to below 10.0. Jones' innings was a more acceptable instance of domination. He scored 47 out of 74 with Crowe's 5 being the next highest score. Not as much of an outlier as McCullum's, but out of the ordinary.

Now we get to two all-time classics: in my opinion, seconded by many, the two greatest ODI innings ever played. No comparisons can ever be made of the higher scores on flat-belters with these classics in bowler-friendly conditions. I place the 189* higher only because of the quality of England bowling attack: Bob Willis, Ian Botham, Neil Foster and Derek Pringle. Richards came in at 5 for 1, saw the score slump to 102 for 7, scored 189 out of 272. The next highest score was Eldine Baptiste's 26, and Richards added 106 for the last wicket with Michael Holding, who scored 12. This was not in front of a super-charged crowd nor was it a television spectacle. It was total domination by an undisputed colossus. If there is an innings better than this, I am waiting to hear of the same, and will probably wait forever. This carries an HSI value of 5.243. I think HSI values of 5.0 are for once-in-a-lifetime performances.

Kapil's 175* (HSI 5.023) stands second only because there is a slightly better 5.xx innings ahead. Kapil came in at 9 for 4, saw the score at 17 for 5, scored 175 out of 266, including an unbroken stand of 126 with Syed Kirmani, who had the next best score of 24, Kapil played the innings of his life and those of million other lives. The similarities between the two innings are startling. That these two innings are third and fourth confirms to me the validity of HSI. If either of these innings had gone out of the top five, I might have had frowns on my forehead.

This table of 30 hosts a number of all-time classics: Richards' 189, Kapil's 175, Watson's explosive 185 (out of 232), the lesser-known masterpieces of Charles Coventry and Paul Stirling, Saeed Anwar's 194, David Gower's majestic 158, Scott Styris' all-time-classic 141 and so on. Barring minor personal preferences this is a table of many ODI classics.

This is the most important table since it measures the HSI across the career. It can be seen that the numbers are at a different magnitude from the article on Sachin Tendulkar's ODI career, since I now measure all innings and have incorporated the team scores. Richards is comfortably ahead with a career HSI average of 0.311. This is confirmed by an average of 47+ and a win percentage around 70. Gordon Greenidge follows with 0.302, the only other batsman with an HSI value exceeding 0.3. Hashim Amla follows next close behind with 0.297. Even if Amla maintains 80% of his 80-match form in the next 80 matches, he may be in a similar position. Now comes Tendulkar: 446 innings, just over 50% wins, an average of 44.83 and an HSI of 0.293. Numbers depicting a magnificent career, embellished by the huge number of matches played and spread over 23 years. This top quintet is rounded off by Desmond Haynes. Look at the top 11 players in this table. The other six are Virat Kohli, Chris Gayle, Crowe, Dean Jones, Brian Lara and Javed Miandad. These, plus Ricky Ponting, are arguably the best 12 ODI batsmen ever, the Dazzling Dozen. Thus the importance of this measure is established once and for all. The relative positioning is immaterial.

Richards batted at the No. 3 spot just enough to qualify: 51 innings, an amazing 80+% wins, an RpI value of 47+ and an HSI value of 0.356. Lara is next having spent well over 50% of his career in these pivotal positions. Tendulkar follows just behind, but with a lot more matches. Then Greenidge and, surprisingly, Shivnarine Chanderpaul. Four of the top five players are West Indians of two different generations. In the next two positions are the two current giants.

Crowe has stolen the thunder from Richards in the Nos. 4 to 7 positions table. The generally lower HSI values, all barring one below 0.3, indicate the difficulty of making significant contributions when batting in the middle order. Richards maintains his win percentage value which is around two-thirds. That master of the middle order batting, Miandad appears next. Steve Tikolo, the only world-class batsman Kenya produced, is a well-deserved presence in this table. Ross Taylor rounds off the top five.

Richards excelled in the first innings: most of his top innings - 189*, 181, 138, 153, 149, were in the first innings. Look at the wonderful win ratio of 70% when West Indies batted first. And Richards outstanding RpI figure of 46. Haynes is placed second, with marginally lower figures. In fourth position is Marsh, the quintessential opening batsmen, with impressive figures under all columns. Knight rounds off the table, but with less that 50% wins. Tendulkar's figures are very impressive but the win percentage remains just below 50%, matching his career numbers.

A smattering of modern batsmen fills up this table. Four of the top five, Gayle, Kohli, Watson and Graeme Smith are still active. If we say that this indicates a better chasing scenario currently, then it may be the correct conclusion for the wrong reason since all these HSI numbers are measures within a team. But it is possible that there more chasing wins now than ever. Kohli's outstanding RpI in chasing wins stands out. Gayle's is slightly low but I expect that he makes up for this with a higher scoring rate. Tendulkar is fine, not outstanding, with a RpI of 38. However, he has a higher win percentage batting second, than batting first.

A very unlikely player at the top: Marcus Trescothick, who was very good at home, but memory tells me that many of these good innings were in lost matches. A tally of 25 wins in 57 matches confirms this. Nathan Astle follows next. Then it is Tendulkar, with a very good RpI value of nearly 44 and an HSI value of just below 0.3. It is surprising that there are not many Indian batsmen in the top-20. Maybe they all took HSI points off each other. It is also surprising that the HSI values are at a lower level: most values are below 0.3.

Richards was the best traveller. I am almost certain that the very high HSI of 0.424 was caused by his stupendous performances in England and Australia. The 138, 153*, 189*, 149 were enough to give him this extraordinary HSI. Look at the very high RpI: nearly 48. Look at the away-win percentage of West Indies teams: 55 out of 82. Then Greenidge. The West Indians seem to have very good away figures, possibly because they do not play many home matches: Lara and Chanderpaul are there in the top ten. Virender Sehwag seems to have liked the pitches away from India too. Mathew Hayden has an above-average RpI value of nearly 41. But Richards stands supreme.

Tendulkar was king in the neutral locations. He also played in as many as 145 matches. Many of these were played at Sharjah and the VB Series-type triangular tournaments. An excellent RpI value of around 44 shows the extent of his contributions in these away matches. The win percentage is around his career level of 50%. Gayle is the only other batsman to have an HSI of above 0.3, with a decent RpI value. Gary Kirsten and Miandad, the king of Sharjah, complete the top five positions. Let me hasten to add that matches in UAE are neutral matches for Pakistan.

In this and the next table, the "Wins" column is superfluous. However common programs generate the tables and I have left it at that. Chris Gayle is at the top of the Wins table. When West Indies won, he was right there outperforming his team-mates. That is what the value of 0.408 indicates. Gayle's RpI is well above average. Amla has won only 52 matches, but chips in very well. And look at his RpI which is an extraordinary 62+. Astle is a surprise entry. Let us not forget that he would have contributed with the ball too. Then comes Richards, winning 114 matches and having an excellent HSI of 0.352 and very good RpI value of 45. Lara is very close behind Richards. It is no surprise that the top-20 batsmen in this table have RpI values exceeding 40. Kohli is the only batsman other than Amla to have a 50-plus RpI value.

Javed Miandad figures at the top of the "Losses" table. Tendulkar is in fifth position. How do we explain this? The only sane explanation seems to be that they did what was expected of them and the others failed. However, let me add that the RpI of Tendulkar and Miandad in these matches is around 32. This indicates that they were not at their best but still out-performed their team mates. Let us not forget that the highest value of HSI in this table is only 0.294. It is not a table which can be easily explained. AS Milind says, A typical loss is normally due to fall of early wickets. So the top-three batsmen might not get too many HSI points.

These are the matches from the World Cups and Champions Trophies, the true world-level tournaments. The matches from quarter-finals onwards and Super-xxx matches are included. A note on the qualification: There were only two important matches, as defined here, in the first few World Cups: the semi-final and final. As such there would be batsmen who have played in these World Cups who would barely reach six matches, leave alone ten. Hence, I have lowered the cut-off for inclusion in this table to six matches.

The table has very surprising entries at the top. The evergreen Saeed Anwar leads the table with 0.422. His RpI is an outstanding 60. Sourav Ganguly is another surprise. It shows the value of his partnership with Tendulkar and confirms the fact that he contributed more than his share in key matches. A magnificent RpI of 62 corroborates this. In third place is the unfancied New Zealand all-rounder, Styris. He has performed very effectively, almost always under the radar. Then we have de Silva, no doubt helped by those three top class performances during 1996, in which he scored over 200 runs. The top five segment is rounded off by the star allrounder from Bangladesh, Shakib. What he has done with limited support is wonderful. Richards is down in 11th position and Tendulkar is further down in 19th position. Look at the win percentage values of Mark Waugh, Hayden and Adam Gilchrist.

Richards leads in four of these 12 tables. Gayle tops in two. After that comes a collection of wonderful players with one each: Tendulkar, Miandad, Kirsten, Saeed Anwar, Crowe and Trescothick.

To download/view the Excel sheet containing the 24106 HSI values (0.100 and above), please CLICK HERE.