April 14, 2016

How much does the toss really matter?

Gaurav Sood and Derek Willis
The numbers show that the team winning the toss has a marginal advantage, and the extent of it varies depending on the format and the quality of the teams
18

Data shows that it's wiser for visiting captains to field first at the start of the English season © PA Photos

The run of good fortune enjoyed by Darren Sammy at the toss during the World T20 was noticed by commentators and journalists. On the way to the World T20 title, Sammy called ten consecutive tosses correctly.

Leave aside the conspiracy theorists - the chance of correctly calling out the side of the coin left facing the sky after the flip ten times in a row is just a shade less than one in a thousand, but over many matches, streaks do occur. Instead, focus on the more fundamental question: how much of a difference does winning the toss make? The general feeling is that it does matter. How else would we arrive at the recent changes in the English County Championship?

After analysing data from more than 44,000 cricket matches across formats, however, we find that there is generally just a small - though material - advantage of winning the toss. The benefit varies widely, across formats, conditions, and depending on how closely matched the teams are.

We find that over all those matches, the team that wins the toss has won the match 2.8% more often. That small advantage increases for one-day matches and decreases for T20 contests. For day-night ODI and List A matches, the advantage is greater still, with the side winning the toss winning nearly 6% more games.

Winning the toss convey an advantage of 2.6% in first-class and Test matches, where pitches can deteriorate, giving the team that bats last a tougher challenge. But the largest boost appears to be in one-day matches, where teams that win the toss win the match 3.3% more often.

None of this proves that winning the toss guarantees a team an increased chance of winning every time, of course. The outcome depends on many things, including whether teams are able to capitalise on the toss advantage. It probably did not hurt that Sammy had the bowlers to make an immediate impact against England in the World T20 final. A weaker team may have seen little benefit.

Our findings might not change the minds of captains who, having lost the toss, apportion it some blame in the post-match press conference. "But you have to realise it was half an hour early start, a bad toss to lose," MS Dhoni said after India's loss to West Indies in the World T20 semi-final.

Dew can be a factor during evening contests, but among the three formats T20 matches showed the smallest benefit for the toss winner (and thus the smallest disadvantage for the loser). The advantage shrinks even further in day-night T20 contests.

Playing conditions, which occupy a large portion of pre-match commentary, matter as well. It's not surprising that we found that in ODI, List A and T20 matches where the Duckworth-Lewis rule was invoked, teams that won the toss won the match even more often than when full matches were played.

These findings probably won't reassure critics of the Duckworth-Lewis method, since in theory winning the toss should convey approximately the same benefit each time, regardless of whether the match is shortened. Perhaps there is some inherent bias in Duckworth-Lewis that makes winning the toss even more important.

Other factors influence the outcome of a cricket match, not the least of which is the quality of the sides involved. Using ICC monthly rankings for international sides, we looked at whether winning the toss made a difference when teams were closely matched or at opposite ends of the rankings. When closely matched teams play, winning the toss has a larger impact on the probability of winning. As expected, the impact of winning the toss was less when a clearly better side played a weaker one.

We also looked at performance: are some countries better than others at capitalising when they win the toss? The data isn't clear on this. There are small but insignificant boosts for most sides, with the exception of New Zealand, who seem to do better in matches where they lose the toss. Sri Lanka and India seem to enjoy the greatest benefit.

To try to counter some of the advantages that home sides could enjoy in English domestic matches and to encourage spin bowling, County Championship matches this season do not have a mandatory toss. The visiting team can choose to field first, and if they do not want to, then the toss occurs. The data on the impact of winning the toss in England is similar to the overall picture - a small but significant advantage is gained by the side that wins the toss - with one glaring exception: matches played in April.

Whether due to cold weather or grassy pitches that can make batting difficult, teams that won the toss in April matches in England lost nearly 5% more often than they won. In every other month, the toss winner was more likely to win the match. Perhaps that alone will encourage visiting captains to take the field first, at least at the start of the English season.

A full write-up of the facts presented in this article, and the underlying data, are here

Gaurav Sood is an independent researcher and Derek Willis is a news applications developer at ProPublica

Comments have now been closed for this article

  • spintl on April 22, 2016, 0:58 GMT

    In this IPL, most of the matches is being decided with the team batting second. The mantra is win the toss, put the opposition to bat. How about, a rule is made that the team winning the toss bats first. It will be fun to watch how many captain will be praying that they lose the toss. So a loser of the toss can be a winner!!!

  • Minder37 on April 21, 2016, 14:45 GMT

    How much does the toss really matter? A lot. Consider the ongoing IPL. The latest count is 13 of 14 matches were won by teams chasing! It's becoming too predictable!!! The only team that broke this trend was RCB in Match 4. Obviously, winning the toss and fielding first is the only decision that a captain can make. So far, 4 teams opting to bat first after winning the toss, lost their games - 2 of these by MSD, who is a more discerning captain!! So why have a toss! Since games are played home and away, why not let the visiting captain decide if his team will bat or field. Then, at least, the 'win-the-toss-and-field' advantage element for the home team - who also have control over preparation of pitch - will be eliminated.

  • DKGP on April 20, 2016, 12:22 GMT

    .....Continued

    Advantages 1. Rather than luck, this method relies on the competency of the teams in assessing conditions and the teams' confidence in their team composition and capabilities. 2. The system eliminates any significant advantage to the team winning the toss. 3. Will bring some real tension and excitement to the toss.

  • DKGP on April 20, 2016, 12:18 GMT

    .......Alternatively, if both teams chose different options on winning the toss (i.e. one team choosing batting with the other choosing bowling or vice versa), then they can just get on with the game (with no concession) since both have got what they wanted.

  • DKGP on April 20, 2016, 12:12 GMT

    A method to reduce the advantage of toss in cricket. May need more tweaking, but probably is a good start.

    Captain decides prior to the toss on his choice upon winning. Similarly, he also decides how many runs he is willing to concede for utilizing the advantage of winning the toss. Both these details are exchanged in a sealed envelope along with the team composition just before the toss. The choices cannot be altered once the exchange happens.

    If both the teams choose the same option on winning the toss (i.e. both choosing batting or both choosing bowling), then the team losing the toss is awarded half of the total runs that both teams together were willing to concede on winning the toss. If both teams chose different options on winning the toss (i.e. one team choosing batting with the other choosing bowling or vice versa), then the team losing the toss is awarded half of the difference between the runs that both teams were willing to concede on winning the toss

  •   Ronald Myers on April 18, 2016, 6:33 GMT

    It is more interesting if you break out the stats by whether the team winning the toss chose to field or bat. I did this analysis a long time back, so my memory is foggy, but I recall that teams who won the toss and chose to field, which was very rare, had a huge advantage in most formats, where as teams who won and chose to bat didn't gain much. The exception was Day/Night ODIs, where the team that won the toss and chose to bat, which was almost always, had a huge advantage.

    This probably makes sense, as in most games, the conditions tend to be slightly better for batting for the first team, but this is offset by the advantage of batting second and knowing exactly what total needs to be chased. When the conditions are better batting second, there is a huge edge to be gained.

    One last thing, in big mismatch games the better team ended up batting first regardless of toss more than one would expect.

  • derekpwillis on April 17, 2016, 1:39 GMT

    Thanks to all the commenters here - I'm glad to see the responses. Some specific thoughts:

    DonnieK: What a great idea. Would love to try and make this happen. JohnMitchell: Agreed - that should be doable. Diganta Sarkar: Great idea. Willsrustynuts: We should be able to break that out.

  • donniek on April 15, 2016, 14:58 GMT

    Do the authors think it's worthwhile extracting a sample from Cricinfo's vast commentary database of international matches over say, the last five years and see if there is a difference in the outcome of the statistical material advantage where both captains would have done the same thing and where they disagreed. Because, where one captain wants to bowl and one to bat, the coin toss wasn't actually required and therefore had no effect on the following cricket match. No material advantage was gained or lost as the toss was immaterial. But if they both wanted to do the same thing, the randomness of the toss could've played a much bigger role in the outcome of a very different cricket match - especially with two evenly matched teams.

    The hypothesis is would be that where they agreed, the material advantage increases in the respective match type and where they disagree, the material advantage would be in line with or less than the overall material advantage for the different match ty

  • JohnMitchell on April 15, 2016, 11:14 GMT

    That's an interesting analysis - it would also be helpful to see it broken down by choosing to bat versus choosing to field.

  • michua on April 15, 2016, 5:32 GMT

    Do not show statistics, toss was part of the game.Taking out too much traditions would make the game boring.Tomorrow they may try with two stumps and 24 yards.Too much of experiment is bad.A good captain knows how to utilise the toss for the advantage of the team.Please do not take away the most exciting part of the game.

  • No featured comments at the moment.