Improving the ICC Test Team Rating SystemDave Wilson | A little while ago (2008), I wrote an article in which I applied the ICC rating system back to 1877 to determine the strength of opposition for a player-ratings system, an offshoot of which was I could then also identify which was the best ever Test team. The original article was summarised for Cricket Web here. At the time, I was surprised to see that the Australian team of 1959-60 rated as the second best ever Test team behind the 2007 Aussies, but as the ICC system is objective and is the de facto standard for ranking teams, I went along with it. The original list is reproduced below:
|Points||Margin||Team||Peak Year||# of teams|
(The table shows peak points rating, margin above second-placed team, the team, year peak points was achieved and the number of Test-playing nations at that time.)
* High scores by second-ranked teams. Australia were the highest-rated second-place team ever, behind the 1962-63 West Indians (130, peaking in 1964-65 with 131); other high-ranking runners-up include the 2007-08 South Africa team (second to Australia’s 126) and the West Indies, second to the 1948 Australians (128, who peaked with 133 in 1951- 52)
Setting The Record Straight
Since researching the original article, however, I’ve come to realise the shortcomings of the ICC system, which I feel are twofold:
1. the existing system does not differentiate between a home series win and an away series win
Since Test matches began, only one country can boast an overall winning record away from home – Australia (0.559); every other country is below 0.500, the average record of the other teams being 0.365, so clearly it is more difficult to win away (incidentally, Australia’s home record is 0.700). By contrast, between 1976 and 1988-89 West Indies won no fewer than nine overseas series, compiling a win percentage of 0.725 away from home. This level of overseas dominance, being rare and as such a better indication of true dominance, should be rewarded.
2. the existing system does not take into account the degree of dominance
In 2002, Australia beat South Africa by an innings and 360 runs while losing only seven wickets – this scores exactly the same as the 2-run game between Australia and England in 2005, i.e. one point for a win. Looking from a series perspective, in 1983-84 the West Indies beat Australia 3-0 without losing a single second-innings wicket, posting runs-per-wicket (RPW) of 48.7 as against 25.0; in 1901-02 Australia beat England 4-1 yet barely outscored England and lost only one fewer wicket. Yet both of these series register as exactly the same margin in series points, i.e 5-1. This is unfair and should be accounted for in the system.
So, to take into account the above I re-ran the system ratings with the following modifications:
1. an away series win was awarded an additional series point, and
2. the degree of dominance was used to offset the total series points awarded.
Note that the degree of dominance is only factored in if the dominance (measured by ratio of runs per wicket of the two teams) is significant. The ratio is also offset on a per-era basis, as the level of home dominance has varied over the years.
The two systems, original and revised, are compared in the chart below. This chart shows the relative dominance, i.e. separation of first and second teams ratings points, plotted against time (this chart shows dominance only since 1930, since when there have been at least five Test-playing nations, as it was less difficult to dominate only two other teams prior to that):
What we can see here is that the more dominant teams, i.e. late-1940s Australians, late 1960s South Africans, 1980s West Indians and 2000s Australians, have much higher margins over second place than previously, which more realistically measures their level of dominance. Take as an example the 1948 Australians, considered by many (including their captain, Don Bradman) to have been the greatest Test team ever – in the revised system they are seen to have been the most dominant side in history, whereas in the original system they are not seen to have been as dominant as, for instance, the 1937 Australians. At other points in history, where no one team particularly dominated, the levels are about the same for both systems.
The revised system throws up some interesting aspects as compared to the current system. For example, using the original system South Africa rated as top team in 1909-10 and 1930-31, though not with the revised system, which rewards more dominant performance scores – they arguably didn’t deserve to be top ranked either of those times. Australia’s Invincibles originally rate at 128 points after their comprehensive victory in England, but at 150 points with the revised system (the other ranked teams rate at about the same). Pakistan does not rank as number one team at any point using the current system, however the revised system shows Pakistan as top side on no fewer than three separate occasions – 1990-1, 1992 and 1994-5, by a single point each time in very competitive circumstances:
|(Pak 1-1 WI)|
|(Pak 2-1 Eng)|
|(Ind 1-1 WI)|
The strongest era under the new system was the mid-nineties, when there were five teams rated at 120 or higher:
|(Aus 3-1 Eng)|
|(Wind 1-0 over Nzl)|
|(Aus 2-1 over Wind)|
Although we currently have five strong teams now, they are not all rated as highly as the five shown above.
Best of the Best
Here then are the highest-rated teams of all time using the revised method. Because of the additional away series multiplier and the dominance factor, the ratings points using this method are higher than those of you who are familiar with the ICC Test rankings are probably used to. That said, here is the revised list of top-rated teams through the ages:
|Points||Margin||Team||Peak Year||# of teams|
* England was tied on 140 points with Australia.
** South Africa is the only second-placed team which is ranked here, 19 points behind Australia’s 154.
Note: the previous article was written prior to India’s ascendency as world leader.
We now see the 1969-70 South Africans assuming the mantle of highest rated Test team, which probably merits some discussion. This team, including such luminaries as Mike Procter, Eddie Barlow, Graeme and Peter Pollock, Denis Lindsay, Trevor Goddard and Barry Richards, while generally considered to be one of the finest Test sides ever is not typically described as the very best ever. Their high rating is partly explained by how the system was designed to work – the system was conceived in the early 2000s and was based on the fact that teams were by then playing many Test series in a 3-4 year period, as many as 45-50 Tests against nine different opponents, whereas in the period covered to determine the rating of this South Africa side they had played only two winning series and nine Tests, all against Australia. Their 7-1-1 record against a good Aussie side means they score very highly, but the lack of variation of opponents must be taken into account when considering the best ever team, though they are probably deserving of a top five spot.
Australia’s 2007-08 vintage ranks just two points below the South Africans – a 3-0 win in Pakistan saw the 2000s Aussies go over 150 points for the first time, and they were able to maintain around that level or better for seven years. Considering that only five teams ever rated at higher than 150, this is an incredible level of sustained domination. After that we see what is basically the 1948 Invincibles, which peaked after their heavy defeat of South Africa following the domination of England, then the 2002-03 Aussies and the 80s Windies rounding out the top five. The 1959-60 Aussies drop down to seventh, probably a fairer reflection of their greatness than the original second place. Compared to the original ranking, this revised system, which takes into account better performances away from home and also dominating performances, gives a ranking list which concurs more closely with generally accepted opinion on the greatest ever Test teams, the top-ranked team notwithstanding.
The best ever ratings for those countries not listed above are:
Pakistan 129 (1994-95, second place to Aus on 133); note also 126 (1992, but first place by 1pt over Aus)
New Zealand 120 (1986-87, second place to WI on 140)
Sri Lanka 115 (1999-00, third place); 115 (2010, fourth place)
Zimbabwe 83 (1999-00, ninth place)
Bangladesh 12 (2009-10, ninth place)
The Strong Get Stronger
Considering that the 2000 Aussie sides dominated so may opponents for so long, I’d still have to say the conclusion of the original article was correct, i.e. the 2007-08 Aussies deserved to be the highest rated. I do feel that the revised system more accurately measures the relative strength of sides as compared to the current system, in that more dominant sides distance themselves to a greater degree, and is therefore a fairer reflection of the relative dominance of great teams. Which is, after all, the point of a rating system.