21 August 2003

Mark of a Winner

After losing a series in South Africa in 1999-2000, England decided to dispense with the services of underachieving batsman Mark Butcher. It was only a spate of injuries in the last Ashes series in England in 2001 which brought Butcher’s return to the side. Many observers have noted the improvement in his average which has risen from 25 to 34.7, and in doing so Butcher is averaging 45 since his recall.

What has also become clear to us, but is less well documented, is that when England win, Butcher seems to play an important part. Butcher has got a century in three of England's last four Test victories, (versus Australia 5th Test, Zimbabwe 1st Test and South Africa 3rd Test—he got just 47 in 2nd Test against Zimbabwe), and has 5 centuries and a 94 in the last nine victories. Butcher has effectively become the team 'barometer', if he plays well so do England.

It was little more than a year ago that some of the TMS commentators were criticising Vaughan for not turning his 40s into 100s, and the same accusations are still hanging in the air around Mark Butcher. Loose shots at Lords when he looked secure in both innings and another dropped catch at Trent Bridge brought up the old comments of poor concentration not allowing him to fulfil his considerable talent. After all, an average of 45 is a considerable improvement, but it is still less than any of the top 5 Australian batsmen for example. So just how important is it that Mark Butcher gets a big score for England?

He's yet to get a century in a lost game. Judging from past history, if Butcher gets to three figures you know England won't lose and have a 75% chance of winning; but if he gets a duck, England have a 62.5% chance of losing. In the drawn games he has a lot of 50s that weren't converted into match-winning hundreds, whereas in the won games he has a 75% success rate at turning 50s into 100s. He has more runs in 15 drawn games than in 23 lost games, and almost half as many again in the 17 won games. It's as if the performance of Butcher determines the balance of the game. A half-century from him may be about par therefore a draw is likely, a century and England look favourites while anything less tips the balance against England.

What is more staggering is that he has only become the team barometer in his second incarnation. His first eight victories, up to and including that game with Cronje’s 'unexpected' declaration in South Africa (January 2000), only saw one century and one half-century from him at an average of 31.14. Below we see the breakdown of figures in his second England career. There is a definite symmetry to his performance: his average since 2001 in games won by England is all but 75, in drawn games around 50 and in lost games, you've guessed it, 25.

You may think it likely that batsmen will score more heavily in wins than in defeats, but that is not necessarily so. We chose Vaughan as a counterpoint because he has played a similar number of games as Butcher since his recall. His stats below show he scores more runs in drawn games and defeats than in victories.

So does Butcher make the difference between all sides—good and bad? Against Zimbabwe, we have seen England won despite a contribution of just 47 from Butcher, so we would conclude that England would have won against a poor Zimbabwe side with or without Butcher's runs. There is also a gap in class between England and Australia, so you would think that the world champions could surely cope with a few extra runs from England's no 3, given that they often beat the old country by an innings. But if we look at the two recent series against Australia that is not the case. Butcher has scored two centuries against Australia since his recall, and by coincidence or otherwise, England have won both these games. In other games, he has looked completely at ease with the likes of McGrath and he has contrived to give up his wicket with a run out or a careless shot, Australia have won. This merely emphasises how important it is for England that Butcher scores runs now and repeats that most unexpected match-winning innings of 173 not out at Headingley.

13 August 2003

Stewart: Batsman, Wicket-keeper, Captain

Stewart was first tried as keeper in Adelaide in 90/91 in the Fourth Test of that Ashes series. The Chris Read experiment of 1999 marked the last time Stewart played as a batsman. For 8 years in between his role was never fully resolved—add that to any admiration of Stewart's achievements in a fairly poor side. There were a lot of differing opinions as to the value of him as an opener (and perfect attacking foil for Atherton's defence), or as a keeper allowing an extra batsman (although it usually allowed an extra bowler). Selectors fluctuated and he was moved up and down the order to nobody's benefit, Stewart (a good bloke and good team member) said he preferred to open but would be happy to play wherever he was needed. One thing they didn't try was vary his position within a match—i.e., open with him in one innings and push him down the order if the other side bat for a very long time. It became one of those generally agreed (but unproven) 'truths' amongst the nodding pundits that he couldn't both keep wicket and open. Indian keepers open quite a lot, and there's nowhere harder than India to keep wicket, with spin bowlers, low bounce and hot weather.

The captaincy doesn't seem to have affected his average, whether he kept wicket or not, but his keeping clearly has. He was actually better as captain and keeper than as keeper alone.

Batting as the wicketkeeper, his average rose briefly above 30 after his second test, but then it took him till his 19th game, and 101 not out v Zim, to average over 30, since when his average has fluctuated fairly gently never dipping below 32 and never touching 36. As batsman it took him 11 games to average 30 and 12 to average 40.

After 14 Tests he never dipped below 40 and after 15 never below 44. He last played as a batsman in 1999. There may have been no real deterioration over the last four years other than its coinciding with him being wicketkeeper.

One wonders whether it is his batting position rather than the wicket-keeping which affects his performance. Batting first as keeper, therefore unaffected by duties behind the stumps, his average is down by a huge amount (30 percent) when compared to his non-keeping average. This must surely be due to his batting position in the team. His second innings variation is nothing like as extreme when he would have kept wicket, even though it would be subject to the same change in batting position. Strange. His average batting third innings after keeping wicket, is up on his overall average. By the fourth innings though he is clearly tired out by having kept wicket twice.

If you could bat first with him as a batsman, he'd get you over 100 runs a game - instead of which he's asked to keep wicket and get about 76 (also batting first).

Note the no 7 position is seriously over-inflated by its high proportion of not outs.

The Tall Guys

During the last Test between England and South Africa, one TV commentator dare to make some comparisons based on his experience watching and playing about some bowlers. At the time, it struck me as verging on the ridiculous, so I decided to examine what the stats showed. Here’s what I came up with:

[R/SO and O/10W are statistics that we have invented. Find out more about them in the Glossary.]

Each of these bowlers has played seven or eight tests at the point where these figures are compiled. Two of them are arguably among the ten best bowlers of all time. Our third Mystery Bowler is a young fellow, who is beginning to draw serious attention to his potential. Our commentator suggested that the Mystery Bowler could be the equal of the other two.

On reflection, our commentator does make some sense. While in almost every category above there is a clear difference in quality between the leader and the other two, there is sufficient commonality between them to make the comparison valid. The one area of wide disagreement, however, occurs in the key stat for bowlers, the Average.

Player A is the best choice for bottling up your opponent while still being able to take wickets. The economy is good, the runs per scoring over low, but it is going to take a 110 overs for him to take all ten wickets.

Player B is the best strike bowler among the trio. He’ll bowl out a side in 75 overs. He also doesn’t give away so many runs per wicket taken, but his economy could be improved.

Player C is not quite in the same league as Mystery Bowlers A and C. His average lets him down badly. He is hit hard when he doesn’t bowl a maiden. He definitely needs to work on his game. However, at this stage of his career he is a year younger than Player A, who is two years older than Player B.

Well, having read the headline, you probably already know whom the commentator was talking about. Player A is Curtley Ambrose, Player B Joel Garner, and Player C is Steve Harmison.

Ambrose finished with an average of 20.99, Garner with 20.97.

If Harmison can improve as much as Ambrose did, he’d finish with an average of 23.07, the lowest for an England bowler since the likes of Laker and Trueman. If he only improves as much as Garner, he’ll boast an average of 30.65.

Harmison’s career is going to be an interesting one to watch. He seems to have a problem with his rhythm sometimes. Has anyone thought of sending him to dance classes?

England v South Africa, Lord’s Analysis, Trent Bridge Prognosis

We could deploy a whole lot of our ‘sabermetric’ statistical analyses to illustrate the difference between a draw at Edgbaston and a win at Lord’s, but funnily enough one simple stat illustrates exactly where England fell short:

 This stat is Batting Overs per Inning (See Glossary.). It measures how many overs a batsman plays based on the balls he faced, and indicates how long he clings to his wicket. As is readily apparent, the openers and the middle order played worse at Lord’s, while 3 and 4 improved a little, and the tail wagged slightly. But those who got better did not get better enough to compensate for how badly the worse ones worsened.

Given that the South African batsmen managed to hang in there for the most part for far longer, one has to believe that it was the England players who, in their first game post-Nasser, reverted to the middle-order-collapse type so characteristic of 1990s England. How long before we hear cries of ‘Come back, Nasser’?

South Africa batted as well as at Edgbaston, so how did their bowlers fare? Let’s check their Match Scores:

Well, there you have it. The South Africans effectively added a second top-line performance to the single one they had at Birmingham, and got more out of their third best bowler. Ntini picked up the slack from a poorer game by Pollock, Pollock gave more support than Ntini managed at Edgbaston, and Hall proved the best bowler in spite of Ntini’s attention-getting ten-wicket, bat-breaking performance. Pity poor Peterson, who was dropped for an Adams who bowled worse.

How does Trent Bridge play?

There was no Test at Nottingham in 1999, so we are only working with three years’ data.

The second innings figure includes England’s massive 617 score there last summer. The recent heat wave in England will probably make scores like that more likely than less likely. The bowlers look likely to face another five days of toil, and we’d suggest that gives an edge to South Africa. This crop of England bowlers seems a little lost if they can’t get any swing. Hmm, maybe there is a song there. It’s not worth a thing, if it ain’t got swing? Can anybody get some music to go with that?