Selecting a fourth bowler
Russell Degnan

Stuart MacGill's supporters are a disbelieving lot. Not only can't they believe he isn't playing, they can't believe the selectors might have good reasons for it. Personally, I approve; MacGill isn't accurate enough; goes for too many runs, and is a one-dimensional cricketer with nothing to offer if he isn't getting wickets.

It is often said that this draws unfair comparisons with Warne, and that leg-spinners shouldn't be expected to bowl with such unerring accuracy. It is precisely for this reason that leg-spinners were on the way out -- expect for a few one-off tests -- until Warne revived the art a decade ago. An innacurate and expensive bowler is a liability. But just how big a liability is worth considering.

There are two aspects to this. The first is to consider the affect on the "team" bowling performance. Whether it better to keep things tight and force the batsmen to play bad shots, or to bowl a heady mixture of unplayable deliveries and complete rubbish [1]. Measuring how each approach effects the bowling at the other end is difficult though, and will have to wait until I can see a nice (inexpensive) way of doing it.

The second measure is to consider it as a selection problem between two bowlers, all other factors being equal. For this, a simple though experiment will suffice.

Suppose we have two bowlers, each of whom bowl 30 overs. One takes 5/150 (MacGill), the other 3/90 (Bracken) [2]. Who is the better option to select?

At first glance you might take MacGill for his five wickets and superior strike-rate. But it really depends on how the bowling at the other end performed. In the case of Australia, the other end is likely to be Warne and McGrath, each of whom average in the low 20s. If those two bowlers combined to take the two remaining wickets at a cost of 50 runs, then the combined figures would be 5/150 for MacGill versus 5/140 for Bracken and co. Given those initial conditions, MacGill would need to average 28 to Bracken's 30 to be as good an option.

If you have the situation that occured during the Ashes, when one player (Warne) is getting wickets at close to 20, then a player who can hold up one end, regardless of his average, is very valuable indeed. This is why MacGill is a worse bet than Bracken [3]. His strike-rate might be better, but their averages are close enough that Bracken's ability to slow the scoring effectively gives Warne and McGrath more runs to play with.

[1] As we always said in the juniors, and for that matter, the fours: "shit gets wickets"
[2] I've marginally exaggerated the difference between what you'd expect from MacGill and Bracken, but the general principle still holds.
[3] Except in Sydney, where MacGill's average is substantially lower and he is one of the strike bowlers.

Cricket - Analysis 3rd November, 2005 19:39:06   [#] [4 comments] 

Ratings - November 2005
Russell Degnan

Zimbabwe v India
Opening Ratings: Zim: 682.83 Ind: 1142.64
1st Test: India by an innings and 90 runs
2nd Test: India by 10 wickets
Closing Ratings: Zim: 672.64 Ind: 1151.25

Zimbabwe would be pleased by two things in this series. They survived the opening day's play in the first test, and they showed admirable discipline and skill in bowling India out for 366 (from 1/195) on the second day of the second test. Once again Taibu batted well and there were good contributions from Blignaut, Streak and Mahwire. For India, Pathan showed he loves a minnow, and there were plenty of runs even if Ganguly's run in with Greg Chappell is ill-advised.

Sri Lanka v Bangladesh
Opening Ratings: Sri: 1086.46 Ban: 614.29
1st Test: Sri Lanka by an innings and 96 runs
2nd Test: Sri Lanka by an innings and 69 runs
Closing Ratings: Sri: 1089.47 Ban: 610.86

About the only interesting aspect of this series was whether the second test would finish before the similarly dull contest between Zimbabwe and India. It did, but only because of the timezone. It is impossible to tell if Bangladesh are improving. They continue to get thrashed by the same margins and their rating continues to fall. As much as I hate two-test series, at least these are only two-test series.

Forthcoming Series:

Pakistan (1046.49) v England (1252.93) - 3 Tests.

Coming off their Ashes victory, England are clear favourites for this series. Nevertheless, Pakistan are playing good cricket in patches, as is their way, and can be expected to challenge. Expect a few English players to be found wanting on sub-continental pitches, but for their disciplined attack to hold them in good stead.

Australia (1342.10) v West Indies (816.39) - 3 Tests.

On paper this is a no-contest. In reality Australia are looking to reshape their side to avoid the sort of catastrophic decline the West Indies have suffered as players retire. The Windies can be a bogey side for Australia, expecially Lara, but anything other than an easy series victory will precipitate a clean-out of the Australian side.

Cricket - Ratings - Test 1st November, 2005 00:53:31   [#] [0 comments]