There are two things that puzzle me about the latest ICC Test rankings. First, that Australia are down to the fourth place, and, second, that Sri Lanka are placed second.
Let's discuss the Sri Lanka ranking first.
Now there's no doubt that they are playing great Test cricket at the moment. Mahela Jayawardene, in particular, is batting beautifully. But are they truly second best?
We must first realize that they are ahead of India by only a miniscule fraction of a decimal point: I did my arithmetic and I note that Sri Lanka are 118.83 and India are 118.82 at the moment! So, for all practical purposes, Sri Lanka and India are currently joint-second.
[And, in passing, just a thought: when ICC's David Kendix has all the numbers on the continuum to play with while devising a ranking scheme, why do we have to look at what's at the second decimal place to decide who's ahead? Surely no two teams can be so equal?]
But to return to Sri Lanka's ranking, let's see who they have played in the recent past and where. Sri Lanka are right now playing New Zealand at home. They are leading the two-Test series 1-0 and are very unlikely to lose the second Test, especially now that they are batting first.
Let's suppose that Sri Lanka are held to a draw in the second Test. The record books will show a 1-0 series victory, but the ICC Test ranking formula would effectively see this as a 2-0 win, because Sri Lanka would've received a bonus point for taking the series!
Now ICC's idea of rewarding a team with an additional bonus point for winning the series is quite a good one, especially if it is a 5 or 4 Test series. It seems a shade generous for 3-Test series, and for 2-Test series it is an unacceptable bonanza.
Before this New Zealand series, Sri Lanka defeated Pakistan 2-0 at home in a three-Test series. The ICC rankings obligingly made it 3-0. The Feb-Mar 2009 series in Pakistan was shared 0-0 after the Lahore incident, and, back in January 2009, Sri Lanka defeated Bangladesh 2-0 in a two-Test away series -- and this effectively became 3-0.
So while Sri Lanka are winning practically everything -- and that's really the best that any team can do -- they are deriving a significant benefit because of three things: (a) Sri Lanka play Test series with relatively fewer matches, (b) they have lately played a lot of cricket at home and (c) they have only encountered weak away opposition recently.
Let me explain the meaning of 'recently', because this is one of the merits of ICC's ranking scheme. The essential idea is that wins in the recent past must get a greater weight than wins in the not-so-recent past. There's no need to quibble about this; it seems to make good sense.
But how recent must 'recent' be? ICC takes it to mean one year (they also have a curious fixation about the month of August, but we'll let that pass), but, at least for Test cricket, one year seems insufficient -- two years seems more reasonable, especially given the current reluctance to schedule too many Test matches.
I therefore believe (a) ICC should not give such a generous one-match bonus for Test series wins involving just 2 or 3 matches, (b) ICC should distinguish between the home-away results (it's always easier to win at home), and, (c) ICC should not scale down the weight after just one year (two years is better).
I must however acknowledge that the ICC ranking does a good job of taking into account the quality of the opposition.
Just to see how things can go awry, consider the case of Australia in the last 12 months.
(a) Australia played Test series with relatively more (3, 4 or 5) matches, so a bonus point didn't offer a big advantage (and they won only one series anyway!)
(b) they played most of their cricket away, and against quality opposition (India, South Africa and England; even their home series was against South Africa!)
(c) while they have had 12 relatively poor months of Test cricket, the preceding 12 months were very good. It seems unfair to scale down the value of a great series win just because it happened 13 months ago, when a win 11 months ago gets full weight.
Like elephants, rankings too must remember a little more.
