The Myth of the Nightwatchman

 

Call it tradition or habit, the use of a nightwatchman is well-entrenched as a tactic in first class and Test cricket. When a wicket falls shortly before the scheduled close of play, and a top-order batsman is listed to bat next, he will more often than not be substituted by a lower-order player, whose sole job will be to prevent further losses before stumps. The rationale seems to be that using a top-order batsman will place him in a no-win situation; he cannot achieve much in the short time available, but if he gets out the team’s fortunes are set back further.

 

Looking at Test history, it appears that, in spite of its popularity, the nightwatchman tactic has never really been evaluated or proven. Derek Lodge took a look at the question in his book Figures on the Green in 1982. He theorised that the tactic was illogical and did not work, but did not attempt to prove this quantitatively. This has now become easier to do, and we can compare the outcomes of innings in which the nightwatchman was used, against similar situations where it was not.

 

Firstly, we can look at the effect on the nightwatchmen themselves. Cricket watchers will probably be able to remember vividly examples of nightwatchmen made good; one (Tony Mann) has recorded a century, and Alex Tudor, without a first-class century to his credit, recently scored a match-winning 99 not out for England vs New Zealand. But how do nightwatchmen respond to the added responsibility? To examine this question, calculations have been made based on 113 examples since 1980 of nightwatchmen being employed, compared to 89 similar situations where they were not, selected according to the following criteria:

 

·        Fall of the 1st, 2nd or 3rd wicket, shortly (i.e. within 20 runs) before stumps.

·        The batting team was ultimately all out.

·        The nightwatchman had an average career batting position of 8 or greater.

 

These criteria encompass the great majority of nightwatchmen used from 1980-98, and non-nightwatchman cases from 1980-94.

 

The first question is how does the reponsibility of being put in as nightwatchman affect the player involved? The scores by the nightwatchmen range from 0 to 94 (RC Russell of England). The average score for the 113 nightwatchmen is 15.0. Many of the scores exceed the player’s respective career averages, and many of them do not. In fact, the two rather balance out since the average batting average for the players used is about 15.2, so overall the promotion has had  little effect on the performances of the players involved.

 

We can also assess the effect of the tactic on the team as a whole. The final scores of the teams using nightwatchmen can be compared to the teams not using them. An example, comparing the 38 cases where nightwatchmen were used at the fall of the second wicket to the 36 cases where they were not, is shown as a graph in Figure 7.4.

 

View graph here

 


 

 


Figure 7.4  Effect of Nightwatchman at Fall of Second Wicket 1980-98  The Blue line is the trend in normal team innings where the second wicket fell just before stumps, but no nightwatchman was used. Half of the 36 points are above the line, and half below. However, of the 38 examples found where nightwatchmen were used, 33 of them are below the trend line, ie the teams did not reach the expected final scores. Only 5 teams reached or exceeded the expected final  score.

 

As always, there is a lot of scatter in the outcomes of the innings. The teams which lost their second wicket before reaching 50, but did not use the nightwatchman, ended up with scores ranging from 150 to 606. But we can say without question that the nightwatchman cases are overall a much worse set of outcomes, especially when the nightwatchman was brought in at very low scores. The degree to which this has happened is very surprising. Do teams use nightwatchmen more when pitch conditions are worse? There is no real evidence for this, and unsuccessful team are no more or less likely to use nightwatchmen than successful teams, and so it can be said categorically that there is no evidence that nightwatchmen provide any advantage when used at the fall of the second wicket.

 

The effects of nightwatchmen at the fall of the first and third wickets are less dramatic, but still clearcut. The results of all the cases are summarised in Table 7.2.

 

Table 7.2     Effect of Nightwatchmen on Final team Score 1980-98

 

 

 

Teams Using Nightwatchman:

 

Wkts

Scored Higher than Expected

Scored Lower than Expected

 

1

12

22

 

2

5

33

 

3

12

23

 

Totals

29

78

Six other cases were too close to call.

 

So the nighwatchman tactic fails more than twice as often as it succeeds.

 

The average team shortfall when a nightwatchman is used is more than 25 runs per innings. The graphs do not reveal the reasons for the failure, but at least one can be suggested. It may be that the disruption to the batting order outweighs any benefits from using the nightwatchman. Certainly, there are examples in Australian Tests where Steve Waugh, the best batsman in the team, has been relegated to the bottom half of the batting order by the tactic. Waugh frequently gets left without batting partners, and as we have already seen, having leading batsmen regularly finishing not out will ultimately rob a team of scoring potential.

 

Using a nightwatchman may spare a batting team further embarrassment of another wicket falling just before stumps (it often fails to do even that), but it also limits the team’s ability to achieve an advantage the next day, because the bowling side now has a chance to start on an inferior batsman at the beginning of play. The tactic has proved popular probably because most captains are batsmen, and they know the insecurity that batsmen feel about going in to bat at the end of the day. But if they bit the bullet and were willing to take the risk, more often than not the team, and the batsmen themselves, would benefit.

 

 

Breaking Partnerships: What Happens Next

 

How often do you see a long partnership where the eventual dismissal of one player is followed closely by the other? If I had a dollar for every time I had heard this remark from commentators….. well I wouldn’t be rich, but it is a rhetorical point. For one example, the all time record partnership of 576 by Jayasuriya and Mahanama of Sri Lanka ended when both players were out with the score on 615. One could find many such anecdotal cases. After all, there have been almost 1,500 Test matches to find them in.

 

In reality a very big partnership only makes it easier to continue scoring. Table 7.3 shows how followup partnerships perform after partnerships of various sizes. In it, the partnerships immediately following those in particular size categories (greater than 50 or greater than 100 for example) have been averaged and compared. Only partnerships involving recognised batsmen have been used for the calculation.

 

 

Table 7.3  Effect of Partnership Size on Follow-Up Partnerships

 

 

Following partnership (average)

% of following partnerships under 10 runs

All Partnerships (1st - 5th Wicket only)

29.5

36.7%

50-100 Partnerships

36.3

29.9%

100-200 Partnerships

41.0

27.4%

200-300 Partnerships

41.9

25.8%

300+ Partnerships

49.8

13.2%

 

The trends here are clear. The bigger the partnership, the bigger the next partnership is likely to be, and the chances of the next partnership failing declines significantly. Players in long partnerships are not prone to losing their wickets in pairs. Case closed.

 

//