Twenty20 July 14, 2010

An in-depth look at Twenty20 results

A detailed study of Twenty20 results indicates the matches haven't been as close as you'd expect

When one views T20 matches, there seems to be a feeling of continuous activity, not because there is a contest between bat and ball but because of the boundaries being hit, the stadium noise and the IPL hangover. At the end of the match Ravi Shastri, irrespective of how the match finished, would say that it was a "humdinger of a match". Alternately some other anchor would mouth similar "words of wisdom". But I have always felt that the matches are not as close as they are made out to be. The excitement seems to be a "manufactured" one. How does one prove or disprove this seemingly subjective observation? I propose to do that by delving into the scorecards and coming out with a suitable analysis.

First let us eliminate some of the matches. Needless to say that only T20 internationals will be considered. IPL matches are not true international matches. Also if the IPL is to be included, then all other club leagues should be included. All matches which finished through the D/L route are discarded. It is clear to most people that, Queen's honours notwithstanding, the learned duo, M/s Duckworth and Lewis, have made a pig's breakfast of the D/L calculations when it comes to T20 matches. More about it in a later article. Finally matches which finished in a tie and decided through the single-over-eliminator will be discarded. After all when the 40 overs were bowled, the teams have finished dead level.

That leaves us with 168 matches (out of the 185 we started with). Now we will separate the wins defending totals (first batting team wins) from the wins chasing the target totals since the two wins are as different as chalk and cheese. One is a bowler-driven defensive win and the other is a batsman-driven attacking win.

First let us take the matches won by teams batting first and defending their totals. There are 83 such matches, just below the 50% mark. There is only one objective in front of the defending team: restrict the opposite team to a total below their own total. Whether this is done by dismissing the other team or restricting them to a total below the total is immaterial. The win is stated as "by x runs" and this is the only measure necessary to measure the type of win. The only factor to be taken into consideration is that a match score of 200/190 is a less emphatic win than a match score of 100/90. This is achieved by dividing the run differential by the first innings total and the Win Index arrived at.

It is a fact that T20 wickets are cheaper to get than ODI wickets (a Balls-per-wicket value of 18.2 against 42.6). This makes the wickets valuation somewhat difficult. I tried adding the wickets captured component to the Win Index. It did not work out, especially for very close matches. Take a match such as 150/148 a.o. By all criteria this is a very close match and should have a very low Win Index value. However once I give credit to the winning team for capturing wickets, the Win Index moves way up and goes into a comfortable win zone (because of the 10 wickets), which is wrong.

A few statistical highlights of this group of matches.

1. The average Win Index is 20.5. This can be compared to the average for the other group later.
2. The average first innings score is 169 for 6.5 wickets.
3. The average second innings score is 133 for 8.5 wickets.
4. The average winning margin is 36 runs, which makes the wins quite comfortable.
5. Out of the 83 matches, the losing team has lost 8 or more wickets in 59 matches (71%).

Before we look at the tables, let me emphasise that absolute values cannot be used in these exercises. An over represents 5% of a team's balls-resource unlike ODIs in which an over represents 2% of the resource. There is less of a difference in terms of runs since T20 scoring rates are higher. Even then, 10 runs in T20 represents around 6% of the average T20 total while the same 10 runs represents around 4% in ODIs. What is normal in T20s is difficult in ODIs. Hence all comparisons are only in relative % values.

Now for the tables.

Matches won by teams defending totals

No  Win  MtId Cty  First Inns  Vs Second Inns  Vs Team Result
Index          <--Score-->     <--Score-->

1. 66.2 0027 Slk 260 6 20.0 Ken 88 10 19.3 lost by 172 runs 2. 61.6 0094 Saf 211 5 20.0 Sco 81 10 15.4 lost by 130 runs 3. 59.2 0075 Zim 184 5 20.0 Can 75 10 19.2 lost by 109 runs 4. 55.9 0002 Eng 179 8 20.0 Aus 79 10 14.3 lost by 100 runs 5. 50.7 0152 Win 138 9 20.0 Ire 68 10 16.4 lost by 70 runs 6. 50.2 0055 Pak 203 5 20.0 Bng 101 10 16.0 lost by 102 runs ... ... ... 70. 4.7 0114 Pak 149 4 20.0 Saf 142 5 20.0 lost by 7 runs 71. 4.6 0123 Pak 153 5 20.0 Nzl 146 5 20.0 lost by 7 runs 72. 3.2 0046 Ind 157 5 20.0 Pak 152 10 19.3 lost by 5 runs 73. 3.0 0036 Nzl 164 9 20.0 Eng 159 8 20.0 lost by 5 runs 74. 2.3 0130 Can 176 3 20.0 Ire 172 8 20.0 lost by 4 runs 75. 2.1 0120 Nzl 141 8 20.0 Slk 138 9 20.0 lost by 3 runs 76. 2.0 0109 Eng 153 7 20.0 Ind 150 5 20.0 lost by 3 runs 77. 1.6 0134 Aus 127 10 18.4 Pak 125 9 20.0 lost by 2 runs 78. 1.2 0007 Slk 163 10 20.0 Eng 161 5 20.0 lost by 2 runs 79. 1.0 0006 Saf 201 4 20.0 Aus 199 7 20.0 lost by 2 runs 80. 0.8 0179 Saf 120 7 20.0 Win 119 7 20.0 lost by 1 run 81. 0.8 0167 Nzl 133 7 20.0 Pak 132 7 20.0 lost by 1 run 82. 0.8 0099 Saf 128 7 20.0 Nzl 127 5 20.0 lost by 1 run 83. 0.7 0083 Aus 150 7 20.0 Nzl 149 5 20.0 lost by 1 run

It can be seen that 5 of the 83 matches have been won with a very high Win Index of 50+. However more importantly, only 14 matches (around one in six matches) could be classified as close matches. The winning margin in the other matches has been 10 runs or more which is quite comfortably a full-over score. This puts paid, at least for these types of wins, to the general feeling that the T20 matches are close matches. Five out of 6 are not.

Now for wins by the second batting teams. There are 85 such matches, which is just over 50%. These are batsmen-driven chasing wins. The chasing team works with two clearly identified resources. The first, and the more important one, explained later, is the number of balls, normally 120. The other one is the number of wickets, 10. The win is normally stated in the lesser of the two resources, wickets. This is less of a resource restriction since the overall balls-per-wicket figure for all 185 matches is 18.2, meaning that the average number of wickets lost would be 6.4 in a 120-ball innings.

The balls left and the wickets left form the basis for determining the Win Index. The proportion of balls left to the maximum balls carries a 66.7% weight. The wickets remaining carries a 33.3% weight. This is not a linear scale since the top order wickets are more valuable. The wicket values are 0.12, 0.12, 0.12, 0.12, 0.12, 0.10, 0.10, 0.08, 0.06 and 0.06 for wickets 1-10. For instance if a team has lost only 1 wicket, their valuation for this component is 0.293 (0.333*0.88). On the other hand if they have lost 7 wickets the valuation for this components 0.067 (0.333*0.20) and so on.

A few statistical highlights of this group of matches.

1. The average Win Index is 27.5. This is a 25% increase over the first group of matches indicating that the chasing wins are a little more easy and the index values are higher.
2. The average first innings score is 129 for 8.0 wickets.
3. The average second innings score is 131.6 for 3.8 wickets. This confirms the view that the wins are relatively easier.
4. The average winning margin is 6.2 wickets and 18 balls. which makes the wins very comfortable.
5. Out of the 83 matches, the losing team has lost 5 or fewer wickets in 68 matches (80%).

Matches won by teams chasing totals

No  Win  MtId Cty  First Inns  Vs Second Inns  Vs Team Result
Index          <--Score-->     <--Score-->

1. 72.2 0131 Bng 78 10 17.3 Nzl 79 0 8.2 won by 10 wickets 2. 70.4 0021 Ken 73 10 16.5 Nzl 74 1 7.4 won by 9 wickets 3. 65.5 0041 Slk 101 10 19.3 Aus 102 0 10.2 won by 10 wickets 4. 61.6 0014 Pak 129 8 20.0 Saf 132 0 11.3 won by 10 wickets 5. 58.3 0129 Sco 109 9 20.0 Ken 110 0 12.3 won by 10 wickets 6. 58.2 0052 Ind 74 10 17.3 Aus 75 1 11.2 won by 9 wickets 7. 57.0 0067 Ber 70 10 20.0 Can 71 2 10.3 won by 8 wickets ... ... ... 80. 9.4 0172 Nzl 149 6 20.0 Eng 153 7 19.1 won by 3 wickets 81. 8.9 0082 Slk 171 4 20.0 Ind 174 7 19.2 won by 3 wickets 82. 7.2 0176 Pak 191 6 20.0 Aus 197 7 19.5 won by 3 wickets 83. 7.2 0072 Slk 137 9 20.0 Pak 141 7 19.5 won by 3 wickets 84. 7.2 0048 Nzl 129 7 20.0 Saf 131 7 19.5 won by 3 wickets 85. 4.6 0151 Slk 135 6 20.0 Nzl 139 8 19.5 won by 2 wickets

In line with our findings, the top 7 wins have a Win Index value exceeding 55. Also only 6 of the matches could be termed close. The cut-off for determining close matches varies between the two types of wins.

Adding the five tied matches to the 168, only 25 of the 173 matches (14%) can be termed as quite close. The other 86% of the matches are relatively easier wins. This confirms my feeling that the excitement is mostly artificially created.

Those of you who would like to raise a point on the relative strengths of the teams, let me point out that in T20s there are fewer contests between the top teams and the minnows. This normally happens only in the World Cups.

Also furthermore, the lesser number of overs actually reduces the relative strength-differential between teams because there is lesser room for error. Hence, it is quite intriguing that there have been such wide margins of victory.

To view/down-load the table of defending wins, please click/right-click here.

To view/down-load the table of chasing wins, 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

Comments have now been closed for this article

  • testli5504537 on July 16, 2010, 16:23 GMT

    Your table is in apparently random order, with some close games in the middle and some one sided games towards the bottom. Your formula clearly doesn't work.

    A more accurate picture would be this: Out of 168 contests, 45 could be considered "close" - ie a victory of less than 12 runs or 6 balls, or a tied game. That is 26% - so over a quarter of T20 internationals (which are typically played between less evenly matched sides than domestic cricket) ending in a close contest! I would love to know what % of ODIs finish that closely. Less than 10% I would guess. All in all, a genuine vindication of the T20 format.

  • testli5504537 on July 16, 2010, 15:08 GMT

    I just think the analysis, albeit an interesting idea, is fundamentally flawed. If we accept that we are simply measuring "closeness" rather than excitement - the indices are unnecessary. For the defending team winning, we can simply talk about how many runs they won by. Less than 10 runs was a very close game, anything more than 20 was not.

    With the chasing team winning then the wickets in hand is the limiting factor far less often than the balls remaining, but because there are 2 factors, the judgement of "closeness" is more complicated. A victory with 1 wicket remaining is a close game, as is a victory with less than 6 balls left. Your analysis completely ignores the number of balls in hand. Look at game 62. Saf 126 for 9, Ind 127 for 4 off 19.5 overs India won by 6 wickets with 1 ball to spare. Sounds like an incredibly close and exciting game. Yet with an index of 17.9, you don't rank it as a close game at all. [[ "Your analysis completely ignores the number of balls in hand." What sort of statement is this. Have you read the article carefully. Pl don't make statements for the sake of making them. Also you don't seem to read replies to other comments. If you read the response to Goel you will understand that I have provided the tables and if you set your closeness cut-off at 20%, your match would fall into that category. How many times should I tell this. I am waiting for the next response from you on similar lines without going through the other responses or reading the article completely. Ananth: ]]

  • testli5504537 on July 16, 2010, 11:55 GMT

    I agree with Abhi's second comment. A good way to measure excitement would be to measure what percentage of the game has gone at the point when the result becomes inevitable.

    I would be willing to be bet that this happens on average in T20 a lot later than in either of the other formats.

    One of the most exciting games I have ever seen in any format was Notts losing to Durham by 4 wickets, with Durham cruising in and needing 1 run to win off the last over with only 4 wickets down.

    Backing by a roaring crowd, Notts bowled 5 straight dot balls - including 2 direct hit run-outs, before the last ball was edged millimeters over slip's outstretched hand for the winning. You don't call that "exciting"?!?

    In the same season, I saw Warwickshire need 15 runs off 2 balls against Worcestershire and almost pull it off (thanks to a helpful 5 wides).

    10 runs may be a bigger % difference in T20 than in F50, but it's still just 2 big hits. [[ py You are percectly correct in what all you say. However you have taken one match out of context and made a statement "You don't call that "exciting"?!?". I would be a stupid guy to call that "not exciting". However that does not take away anythin away from my article. There are crazy overs such as Notts vs Durham or Pakistan vs Australia (5 wickets in the last over) or a 28 run last over or Tendulkar's last over against South africa in 1994 at Calcutta... Oh we could go on. The fact is that these happen once in 50 matches. Finally pl read my reply to Goel. It is your choice as a reader to determine what is close. Nothing is set in stone. Also do not confuse exciting and close. If the last over started with the batting team needing 45 runs and the last batsman did a Sobers/Shastri/Yuvraj. That might be exciting but not close. The last over by Aamer to Australia was exciting but did nothing to further Pakistan's chances since the total was already huge. Ananth: ]]

  • testli5504537 on July 16, 2010, 5:10 GMT

    I think that T 20 is spoiling the spirit of cricket. no technique is required by the batsmen they just come and have to beat the ball. it will die soon. it is one sided game just designed for batsmen. totally rubbish

  • testli5504537 on July 16, 2010, 3:16 GMT

    I actually think you are right in thinking T20 matches aren't as close as they are made out to be, but I think what went wrong with your analysis is that the closeness of a match is all based on perception, rather than statistics. It only becomes a close match when you see stress on the faces of the players on both teams late in the match, regardless of what the numbers say. I think in this particular case you might be trying to measure something unquantifiable, even though I agree with what seems to be a gut feeling for you that T20 games are not as close as they appear. Love your work though, this column is excellent.

  • testli5504537 on July 15, 2010, 23:24 GMT

    I am against Inter. T20. Or at least just play World Cup thats it, no need to play Country vs Country each time they visit certain country. All teams play this format domestically (i.e. IPL Champions league, Big bash, Provident t20 etc. etc.) Because internationalizing t20 will attract few players to choose to play only that format, not others. And also by canceling T20 internationally, will force people to perform better in domestically to get into their Country team. So, one way you are not killing any format, and nurturing it. I found later half of the IPL pitches extremely difficult to bat on. If IPL starts making bowler friendly pitches from beginning, I don't know then there will be lot of low scoring matches, like one of the semifinal match. It can be helpful to create good players who can play in difficult condition all the time. But it might just destroy excitement of fast scoring. There should do 50/50. 50 high scoring 50 bowler friendly. All bowler friendly will kill IPL.

  • testli5504537 on July 15, 2010, 17:31 GMT

    The list is a good one and gives us many close matches. But I dont think it is an accurate representation of close matches. The analysis is totally dependent on the final score and does not consider the path taken to reach that score. For eg.

    69. 13.9 0023 Aus 138 9 20.0 Zim 139 5 19.5 won by 5 wickets

    I distinctly remember this match, in the first T20 WC in 2007, to be a close and exciting one. I checked out the scorecard and Zimbabwe needed 16 runs off last 9 balls with 5 wkts remaining. Anything could have happened.This for me , ought to be a close one, but is not under your analysis. [[ Goel Your comments are quite valid and, more pertinently, stick to the theme of the article. Please remember one thing. I am not sitting in a high chair and stamping a match as "close" "not close" forever. Nothing is set in stone. I set my own cut-off as 10.0 and designated some as "close" matches. There is nothing to prevent from your setting 15.0% as the cut-off, in which case this match would have been termed close. The point is that closeness is a subjective term and can be derived from this objective exercise. Once again many thanks for trying to understand the idea behind the analysis and saty away from issues not covered. Ananth: ]]

  • testli5504537 on July 15, 2010, 14:29 GMT

    Ananth, Don’t know about others but I feel 20/20 is the worst form of cricket imaginable. A couple of mediocre players, on their day, could defeat another team comprising of 11 members all of whom may actually be greater quality cricketers.

    So, couple of players have their day, bang a few sixes and boundaries- and the match is won! How weird is that?

    But anyway, I think I get the hang of some reader gripes…the thing about T20s is because of what I mentioned above. i.e even though the end result may not show a “close” match…till almost the very last the other team has the “potential” to win…for eg.18 odd runs an over are now almost common….so , I guess part of the excitement is that “you never know” …a team may just pull it off ( as someone said that in ODIs and Tests the end result is often almost a forgone conclusion well before the actual end-except for a few humdinger matches)……In T20s you normally have to wait just a little “longer” into the match. [[ Some editing done. Ananth: ]]

  • testli5504537 on July 15, 2010, 11:43 GMT

    Ananth It is the words 'manufactured' 'artificial' excitement. I fail to see what is manufactured or artificial. T20 is a sport that involves certain rules like any other sport. But whose nature makes it as exciting as it is. To compare it with Tests(especially a statistially oriented person like you) is completely ridiculous. For one, the scale of the numbers in all the areas is completely different. And you're saying it is an uneven contest between 'ball and bat'(actually the most common argument). In the business section of the newspapers you can always see the phrase 'reset economy'. Its the same thing here. in T20s 8rpo is the 3 rpo in Tests(it IS the anture of the game). I don't believe you dont understand THAT. I just wonder what makes you say it is an uneven contest. Why can't 8rpo be the PAR and bowlers with 6 and less be considered exceptional. Why should 3 be the par in T20s too? Perhaps you have a better answer than my brain could come up with! [[ Anabayan I never said that 8 is not par. Where did I say a par of 3 in T20s. I only say that the matches are not as close as they are made out to be and that there should be a tweaking in the bowlers side a bit. PERIOD. I think you are overreacting. The manufactured or artificial wordsl relate to television coverage, and I stand by that. I have never seen a T20 match live in my life. Maybe the ground atmosphere might be diifferent. Anyhow I suggest you also relate to what I am really trying to say and not put words in my article. Ananth: ]]

  • testli5504537 on July 15, 2010, 11:29 GMT

    Ananth I am amused to read a few negative comments on the T20 article. I can certainly understand what is happening since I understand both your point of view and the readers' point of view. I get the feeling that your gripe mainly is against IPL and the excessive commercialization around the same. However you are being seen to pull down T20s. The readers seem to get a great 200 minutes and any attempt to criticize what they see as entertainment does not go well with them. Your attempts to explain the article seems to fall on deaf years. The television which seems to be great for many does not seem to sit well with you. I can also understand that. [[ Vivek You seem to have hit the nail on the head. Probably I am getting my aversion for the over-commercialized IPL percolate to T20. However let me emphasize that my attempt is not to deride T20s. Immediately I want to say that the matches are not as close as they are made out to be, as television demands. In the longer run I would like to redress the balance between batsmen and bowlers a bit. Maybe it may not be a bad idea to call T20 International matches T20 and IPL and other leagues as C20 matches (Club-20). That might avoid this slight confusion and give importance to the true international matches. Anyhow, thanks. Ananth: ]]

  • No featured comments at the moment.