The Last Shall Be First

December 2008

It has been the choice of generations of captains: if you win the toss, bat first. But does it really work? Charles Davis looks at the evidence.

It is a hallowed tradition of cricket, from time immemorial, that captains should bat first when they win the toss. A Test match captain (was it W.G. Grace?) put it succinctly:

“When you win the toss, usually you bat. If you are in doubt, you think about it, and then bat.”

Traditionally, this has been supported by the fact that teams batting first have won more often than they have lost. Yet, slowly but surely this pattern has changed, so gradually that few commentators have given it any notice. The surprising fact is that, since about 1995, Test cricket has favoured the team batting second. India’s successful pursuit of a target of 380 just a few days ago against England is an extreme example, but it is also part of a long-term trend.

The first chart, showing this historical trend, illustrates this clearly.

Graph1

Not including draws. Tests involving Bangladesh excluded.

The traditional advantage of batting first, which once gave rise to a 59:41 win:loss ratio, has vanished. For the last 400 Test matches, teams batting second have won more than they have lost. Why should this be?

It would appear that back in the days of uncovered pitches, and perhaps for some time thereafter, pitches wore out as matches wore on, and batting was usually easiest on the first day. Before 1945, teams batting first led, on average, by about 30 runs on first innings; since 2000 this difference has disappeared. The winding back of the old advantage has exposed a countervailing benefit for teams batting second that is intrinsic to the game. Put simply, teams batting first must take greater risks to win than teams batting second.

For evidence, here is a simple statistic. Tests won by teams batting first are, on average, 10% longer than those won by teams batting second (335 overs to 304, since 2000). This is because teams batting first must attempt to score an excess of runs to secure victory, which takes time. When time is constrained, most captains will delay declarations until the probability of defeat is minimal. No captain wants to declare in a Test match and lose, and the delay can give his opponents an escape hatch, bringing the sanctuary of a draw closer.

Drawn Tests tend to finish in favour of the team batting first, who (since 2000) have led on first innings 60 times, and trailed only 31 times (excluding rained-out draws). So teams batting first see more winning opportunities turn into draws.

In Sheffield Shield cricket, captains are more likely to make risky declarations, but even so, teams batting first two have won only Sheffield Shield matches this season, with ten losses; food for thought, one would think.

Another headache for a captain batting first can arise in follow-on situations. Enforcing the follow-on can stretch bowling resources to breaking point, and resulting injuries can affect performances in subsequent Tests. Since 1994, teams enforcing the follow-on have done no better than those not enforcing it, and have even lost on one memorable occasion in Calcutta. It is not surprising that not enforcing the follow-on, once rare, has become more popular in recent years. In seventeen of the last 41 follow-on situations, over 40%, it was not enforced.

Is It Better to Lose the Toss?

All this suggests that captains winning the toss should, more often than not, choose to bowl. Is this happening? Well, there is a trend in that direction. In the old days, the choice to bat was made about 90% of the time. In the last five years it has been about 63%, an all-time low. However, the old-timers at least could prove that preferring to bat was sensible (in the long run), but today this choice actually runs against the statistical grain.

A captain sensitive to criticism will certainly know what to expect if he puts the opposition in and they score 500. But how often do you hear criticism of a captain who chooses to bat, if his opponents score 500 on the second and third days?

Graph2

Tests involving Bangladesh excluded.

The second chart shows that in this decade winning the toss now confers almost no overall advantage, with a 50:50 ratio in recent years, compared to 57:43 before 1945. This occurs even though captains choosing to bowl have a strong winning record. The current figures actually represent an improvement over the 1990s (see chart), when teams winning the toss actually lost more matches than they won (46% to 54%). An increase in the frequency of choosing to bowl is probably responsible for the improved outcomes since 1999.

A final statistic should reinforce the point: since 2000, teams choosing to bat have won 85 and lost 95 Tests, while teams sending the opponents in have won 49 and lost 40 Tests (excluding Bangladesh). There is some subtlety in the detail – for instance, the West Indies were very lucky with the toss for a while, but still lost repeatedly. Even so, it is clear that there is no overall advantage in choosing to bat first.

In fairness, there is more complexity to the situation, not discussed here. Some captains are less reflexive in their choices than others. There has also been a very recent tendency for teams batting first to do better than earlier in this decade; whether this is a statistical blip or a new trend remains to be seen.