Shane Watson turns one around the corner

Sydney 2014: Australia pip South Africa to the post in a scrappy, entertaining affair

© Getty Images
4

Stats feature

The heart-stoppers

Which are the most thrilling T20 internationals ever played? We look at the stats

Anantha Narayanan |

West Indies and India played a pulsating T20I in Florida last year. West Indies put up a mammoth 245, leaving their opponents requiring a record chase. India started in the only way such a total can be chased: they were 67 for 2 after six overs and 116 for 2 at the halfway stage. The match then went through a riveting series of ups and downs, and ended in a one-run win for West Indies.

Was this the most exciting T20I ever? This article attempts to identify the most exciting among all the T20 international matches played until the end of March. A boundary needed off the last ball is not enough; there have to be changes of status all through the game.

I will look for two types of games.

- Matches in which the winning chances of, say, the chasing team, looks like a sine curve. In other words, games where the momentum swung like a yo-yo between the two teams. The excitement in some of the matches may not be obvious at the outset but is present throughout, just below the surface.

The MSI explained

The Match Status Index is a percentage value out of 100 that indicates the winning chances of the team batting second. It is based on:

-The required rate compared with the par rate. The par rate is derived by analysing overs 11-20 of all innings played so far

-The current scoring rate compared with the required rate

-Wickets lost

-The quality of the batsmen at the crease and those yet to bat

-The momentum factor: what happened in the last three balls?

-The active bowler's performance in this match

-The length of the ongoing partnership

The weightage given to wickets lost and the quality of remaining batsmen are reduced in the 18th and 19th overs, in favour of scoring-rate related parameters, since these become more relevant. However, this tweak is made only when the chasing team reaches the 18th over losing three or fewer wickets.

This method of determining MSI works until the end of the 19th over. For the last six balls, each delivery is analysed and numerical inferences drawn. For instance, assume the target is 200. A team will have a chance of a positive result if they are 194 for 9 before the last ball is bowled. However, if they are 192 for 1, it is game over. As we can see, the wickets remaining lose their value drastically as the final over progresses.

- Matches that featured remarkable recoveries, in which teams came back from the abyss. These matches are won despite teams having as low as a 20% chance of winning at some stage.

The basis for this analysis is ball-by-ball data, which is available for almost all T20I matches. The key measure is the MSI (Match Status Index), which will be determined before each ball after the tenth over of the second innings (see sidebar for explanation). If the MSI value is between 1 and 10, it indicates an almost certain win for the bowling team. If the MSI value is between 45 and 55, the match is too close to call. If the MSI value is between 90 and 99, the batting team is almost sure to win.

Any ball in which the MSI moves from above 50 to below 50 or vice versa is classed as a Winning Team Change, or WTC. An MSI sequence of 52~47, for instance, represents one WTC, and a sequence of 52~47~53 represents two WTCs. A sequence of 52~47~45~53 is still two WTCs. For any match to be considered a cliffhanger in this analysis there have to be a minimum of ten WTCs.

The graph for the remarkable game in Florida is below. Incidentally, this match has the joint-highest number of WTCs in T20I history.

T20I No. 562 West Indies v India, Florida, 2016
West Indies: 245 for 6; India: 244 for 4. West Indies won by one run MSI distribution 28 (WI): 36 (Ind)
WTCs: 17

At the start of the 11th over, with India on 116 for 2, West Indies were marginally ahead. The four off the first ball pushed India in front. Then the series of singles pulled it back, before the six off the last ball tilted the game back in India's favour. The fluctuations in this over were mainly caused by the runs scored.

In the first four balls of the 12th over, bowled by Kieron Pollard, India still held the upper hand, but a wicket in the fifth ball set them back. Even when Carlos Brathwaite conceded 12 runs in the 13th over, West Indies were ahead. The partnership between KL Rahul and MS Dhoni had yet to make its mark. India inched ahead in the 14th over, bowled by Sunil Narine, thanks to a six and two fours from Rahul. In the next over, bowled by Brathwaite, India hit a six and a four but were kept quiet in between those boundaries.

In these five overs, the MSI had several shifts up and down but finally settled at 47.6, almost the same as it was at the beginning of the 11th over. This is what made this phase so exciting for those watching.

In the 16th over, Pollard conceded a three and a four off two balls, but just four singles off the remaining deliveries. The first ball of the 17th over was a dot, but then came a wide and a six that pushed the game into India's zone. The two further sixes in the over consolidated this advantage. The partnership, now in the 30-ball zone, also contributed to the MSI.

The 18th over, bowled by Dwayne Bravo, was a terrific one. The four off the first ball pushed the MSI value past the 60 mark and allowed India to motor ahead. Then the dot ball, two, and three singles took effect and dropped the MSI to around 55 at the end of the over. It was a fine over from Bravo, considering that two well-set batsmen were at the crease, but with 24 needed off 12 and a flourishing partnership in progress, India still had a clear edge.

The West Indies v India T20I in Florida had the joint-highest number of fluctuations of fortunes (according to this analysis)

The West Indies v India T20I in Florida had the joint-highest number of fluctuations of fortunes (according to this analysis) © Getty Images

Things got worse for West Indies in the 19th over. A six, a four and a wide left India as firm favourites with the MSI at 66.5. Now the equation was eight required off six.

The 20th over was bowled by Bravo, the death-bowler extraordinaire. The game appeared lost for West Indies when Marlon Samuels fluffed the easiest of catches off the first ball. However, India took only a single and the MSI remained around 65. If Samuels had taken the catch, the MSI would have dropped to around 50. Bravo then bowled a bewildering mix of low full tosses, yorkers and slower deliveries to keep India to five runs in the next four balls. The MSI gradually dropped, and when the last ball was bowled, it was firmly at 50, anybody's game, since two runs were needed to win.

Another slower delivery followed and Dhoni sliced to short third-man, only to be caught by Samuels. India, who had looked like winning for about three-fifths of the latter 60 balls of their innings, had lost by one run.

Now, a look at the other seesaw battles.

1. T20I No. 197: Australia v England, Adelaide, 2011
Australia: 157 for 4; England: 158 for 9. England won by one wicket
MSI distribution 28 (Aus): 32 (Eng)
WTCs: 14

Shane Watson's innings helped Australia reach a competitive total, given how Australian pitches often aid bowlers in T20 matches. England's chase was patchy. They were almost always scoring at above eight runs per over, but they lost wickets regularly, keeping Australia in the game. The MSI climbed above 60 only in the 19th over. A single was needed off the last ball with the last wicket in hand and England just about managed to score the winning run.

2. T20I No. 257: Pakistan v Australia, Dubai, 2012
Pakistan: 151 for 4; Australia 151 for 8. Match tied
MSI distribution 44 (Pak): 16 (Aus)
WTCs: 12

Pakistan's first innings was a below-par performance. There was no explosive finish despite the availability of wickets. Chasing 152 for a win, Australia were comfortably placed at 79 for 1 in ten overs. Then the loss of two quick wickets turned the tide. Australia were pegged back through accurate bowling and the required rate mounted. Two wickets in three balls effectively scuttled the Australian chase and they were lucky to level the scores thanks to a Pat Cummins six. One run was required off the last ball but Pakistan dismissed Cummins, and went on to win the eliminator.

3. T20I No. 294 India v England, Mumbai, 2012
India: 177 for 8; England: 181 for 4. England won by six wickets
MSI distribution 28 (Eng): 32 (Ind) WTCs: 13

India finished with an above-par 177. England were comfortably placed at 90 for 1 at the halfway mark. They managed to keep the required rate below 10 until the 15th over. Yuvraj Singh's excellent spell pushed them back and the required rate steadily went up. The last three overs produced 40 and England won with a six off the last ball. Before the last ball, India were still the favoured team.

4. T20I No. 409: Australia v South Africa, Sydney, 2014
South Africa: 145 for 6; Australia 146 for 8 in 19.5 overs. Australia won by two wickets
MSI distribution 16 (SA): 45 (Aus)
WTCs: 17

After a good start, South Africa scored only 33 runs in the last six overs, to finish with a below-par 145. Australia struggled through the rather easy chase and lost wickets regularly. A terrific 18th over by David Wiese yielded only two runs and produced a wicket. Then a four meant that five runs were needed in nine balls. Two dots and another wicket kept South Africa in the hunt until the penultimate ball but Australia eventually sneaked through for the win. A very exciting but scrappy match, with a record 17 WTCs in the last ten overs.

The following three matches were also in the final reckoning:

1. In T20I No. 23, Australia scored 138 for 9 and Zimbabwe overhauled this total off the penultimate ball. The MSI distribution, however favoured Australia, who had 40 chances of winning against Zimbabwe's 19. There were 13 WTCs.

2. In T20I No. 275, it was all Sri Lanka up to the 15th over in the second innings. Slowly New Zealand fought back. In the last over, Sri Lanka needed five off two balls. They scored a four but the run-out of Lahiru Thirimanne meant that the match ended in a pulsating tie. Sri Lanka, however, won the eliminator. The MSI distribution heavily favoured Sri Lanka (51-10) and the WTC count was 10.

In T20I No. 418, Zimbabwe scored 175. Pakistan were on pace for the win but the required rate of nine and the fall of wickets at the end kept Zimbabwe in the hunt. Twelve were needed off the last over but a six off the first ball clinched the match for Pakistan. The MSI distribution was 31-28 in favour of Pakistan and there were 16 WTCs.

Terrific comebacks
1. In the history of T20Is, there is only one match in which the winning (chasing) team never had a single MSI value exceeding 50. In other words, they never looked like winning, even before the final ball was bowled. This was T20I No. 463 between Bangladesh and Zimbabwe in Mirpur in 2015. Chasing a moderate 136, Zimbabwe were struggling at 52 for 5 at the end of the tenth over. The required rate kept climbing and 18 were needed off five balls and later, six from two. Bangladesh were still favourites but a Neville Madziva six sealed Zimbabwe's sensational win.

2. In T20I No. 176, a World T20 semi-final, Pakistan put up 191 and Australia were in dire straits with 125 needed off ten overs. The required rate almost never went below 12 and it wasn't until the penultimate delivery of the match that the pendulum swung. Mike Hussey's three sixes and a four in the last over won a remarkable match where the MSI tally was 2-60.

3. In the most recent World T20 final, T20I No. 557, England were winning comfortably even at the beginning of the final over. With 19 needed of six, it was England's match. Thirteen from five was still England's match. Seven from four was a changeover. At one from three, West Indies won with the fourth successive six by Brathwaite. Out of the 62 balls bowled in the last ten overs, England had winning chances in 60 and West Indies in 2.

4. Finally, a wonderful win defending. In T20I No. 223, South Africa finished with a par score of 165. New Zealand were well placed at 103 for 2 in ten overs and 141 for 3 in 15. Twenty-five runs were needed in 30 balls with seven wickets in hand: a cakewalk indeed. Then came the extraordinary stutter. In the next five overs, New Zealand scored 21 for the loss of four wickets and ended three runs short. The last over started with seven needed. The sequence by Marchant de Lange was 1, 0, W, 0, W, 2nb, 0. The MSI count was 5-59 - arguably the greatest defending win of all time in T20Is. The graph for this match is interesting in that barring the five balls at the end, the MSI values are in the top half.

One match under serious consideration was T20I No. 459, where Afghanistan beat Zimbabwe despite being behind for most of the chase. T20I No. 78, between New Zealand and West Indies, which ended in a tie, also figured in the final selection. So did T20I No. 296, a low-scoring classic between India and Pakistan in 2012.

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

 

RELATED ARTICLES

 

LOGIN TO POST YOUR COMMENTS

  • POSTED BY aketh on | June 19, 2017, 22:59 GMT

    An impressive metric to analyze a game. I really think this is amongst those few metrics that actually require a ball by ball detail of the entire game.

    It would be interesting t know the MSI metric of the recent IPL 10 final.

    I am interested in knowing the details of the MSI formula. It would want to explore MSI metric for the IPL games.

  • POSTED BY himhig9486921 on | May 24, 2017, 4:25 GMT

    Great way to analyse the swings of fortunes in a match. I'd be interested in knowing the details of the MSI formula. I also think the MSI factor should correct the RRR using the overs remaining. After all, sustaining an RRR of 8 for 9 overs is tougher than sustaining the same for 4 overs. Also, since you must have calculated the MSI values for all games, another couple of interesting measures we could derive from that would be:Batting MSI Impact: The Sum of the MSI changes when a particular batsman in batting, averaged over all balls faced (in overs 11-20 of a chase) by that batsman. Essentially, the sum of MSI changes happening while the batsman bats. Bowling MSI Impact: Same thing averaged over all balls bowled by a particular bowler. Of course the signs of the MSI change will be positive towards the team of the player while computing this index. This could be a great set of measures to quantify the value of batsmen and bowlers at the business end of a T20 game. Another related number could be the number of WTCs effected by that player in favour of his team over a career. What do you think?On a side note, if I may, is there any way to contact you? I have a couple of works I'd like you to critique.
    [[
    I have made a note of your mailid and I will contact you directly.
    The MSI/WTCs currently are team-centric. What you are suggesting is a way to re;ate these to players. An interesting concept. But a lot of work. I do an article for TCM only once in 4/5 months and I am not sure whether I would do a T20 one in the near future. The Par Rate changes for each over.l So the effect of number of overs remaining is already taken into account.
    Ananth
    ]]

  • POSTED BY Cricinfouser on | May 24, 2017, 4:01 GMT

    why is the India vs pakistan 2007 t20 final not considered?it was a hell of a match.. a match that changed the landscape of t20 cricket for ever..
    [[
    It was a great match, no doubt. But did not meet the qualification criteria of 10 WTCs.
    Ananth
    ]]

  • POSTED BY sreedev on | May 23, 2017, 19:29 GMT

    What about the T20 between India and Sri Lanka few years back when India were 7 or 8 down with a lot of runs to get. The Pathan brothers then got India home
    [[
    Yes, a good match indeed. T20 #82. Slk: 171/4. India: 174/7 in 19.2, after being 115/7. This match was under consideration for the comebacks but other matches got ahead. In the last two overs India was ahead and finally they won quite comfortably.
    Ananth
    ]]