Info on new ICC Test C'ship

Wednesday, May 28 2003

How the ratings work
What is a rating?
A rating is worked out by dividing the points scored by the match/series total, with the answer given to the nearest whole number. It can be compared with a batting average, but with points instead of total runs scored and a match/series total instead of number of times dismissed.
After every Test Match, the two teams each receive a certain number of points, based on a mathematical formula. Each team’s new points total is then divided by its new match/series total to give an updated rating.

With batting averages, if you are dismissed in your next innings for more than your average, your average will increase. Conversely, scoring less than your average will cause it to fall. Similarly, under the ICC Test Championship method, the points awarded for a win will always be more than the rating the team had at the start of the match. Equivalently, a team losing a Test Match will always score fewer points than its rating. So a win will always boost a team’s rating and a defeat will harm it.

A draw between a higher and lower rated team will slightly benefit the rating of the lower rated team at the expense of the higher rated team. A draw between two similarly rated teams will leave both their ratings unchanged. A tied match is treated the same as a draw for rating purposes.

What does a particular rating signify?

A team that over the period being rated wins as often as it loses while playing an average mix of strong and weak opponents will have a rating of close to 100.

A rating of 100 could also correspond to a side that wins more often than it loses but who has generally played more matches against weak teams. Similarly, if the majority of its matches are against strong teams, then a rating of 100 could be achieved despite having more defeats than victories.

The recent pattern in international cricket has been for there to be a number of teams in the 90-110 range. These teams are of broadly similar standard. A rating above 120 suggests consistently strong performances. Above 130 is rarely achieved and suggests a high degree of dominance over all other teams.

In every match the total rating points available equals the sum of the initial rating of the two teams, so ratings can be thought of as being redistributed rather than created. There is therefore no ‘inflation’ in this rating system, so a rating of 120 suggests the same degree of superiority over opponents now as in the past or future, and a team can meaningfully compare its rating movements over time.

How quickly do ratings change?

The amount by which a rating improves after winning a Test will depend on the rating of the opponent. A win over a much stronger team (i.e. one with a much higher rating) boosts the rating more than beating a much weaker opponent. Conversely, losing to a much stronger team will not cause the rating to drop too far, but losing to a weaker side would.

The predictor table described in Section 2 is the most convenient way to monitor the sensitivity of teams’ ratings to forthcoming results.

It is possible for a team to win a series yet for its rating to fall. This will happen if a stronger team wins a series but by a smaller margin than the respective ratings suggest should be the case. For example, when Australia played England in a five Test series in 2002-03, Australia needed to win by a margin of at least three Tests just to maintain its very high rating.

How is the series result incorporated?

At the end of any series comprising two or more Tests, a series bonus will be awarded. Like an individual Test match, a series can be won, drawn (tied) or lost. For rating purposes, the series result is equivalent to the result of one further Test.

Suppose a team has just won a Test series. The series bonus can be regarded, for rating purposes, as if one extra Test has been played and won by the team that has just won the series. If a series finishes level, the series bonus is equivalent to the two teams playing an extra drawn Test.
Period covered

The table reflects all Test Matches played since 1st August 1999, i.e. nearly four years of fixtures. All Tests played until the start of August 2003 will be added to this table, so by then the ratings will be based on a full four years of results. Then, in August 2003, the first year of results will be dropped, so the table will then cover the last three years of results starting in August 2000.

This pattern will be repeated each August, with the oldest of the four years of results removed to be gradually replaced with results of matches played over the following twelve months. This means that once a year the positions could change overnight without any new Test Matches being played.

Weighting results

All matches included within the ICC Test Championship will always fall into one of two time periods:

 Period One covers the earliest two years of matches
 Period Two covers all subsequent matches, i.e. the last one to two years

Weightings are applied to these two groups of matches so that the ratings more fully reflect recent form. The weightings are as follows:

 Period One matches have a weighting of 50%
 Period Two matches have a weighting of 100%

In the current table, matches played since August 2001 receive a weighting of 100%.

After August 2003, the weighting of matches being played now will remain at 100%, while the weighting of matches played in the year to August 2002 will fall to 50%. Matches being played now will drop out of the ratings altogether in August 2006.

The ‘match/series total’ column in the ICC Test Championship table comprises a combination of individual Tests and series. This total along with the number of points earned in each period is multiplied by the weighting factor. For example, suppose a team played 20 Tests and 6 series in Period One, plus 15 Tests and 5 series in Period Two. The total matches played for rating purposes is 50% of (20+6) plus 100% of (15+5), which equals 33. (A small technical adjustment ensures that, for all teams, the total number of matches and rating points is always a whole number.)

If two teams have the same rating, the team with the higher match/series total will be placed higher.

Defining a series result

At the end of each series, the series result is determined by awarding one point for each Test win and half a point for each Test draw. In addition, for any series consisting of two or more matches, there is a bonus point available for the series winner or half a point to both teams if the series is tied. For example:

A five match rubber finishes 2-1 with 2 draws. The series result for rating purposes is 4-2. (The winning team has two wins worth one point, plus two half point draws, plus an extra point for the series win.)

The two figures in the series result (4-2 in the above example) are referred to as series points scored and series points conceded respectively

For series of at least two Tests, the total series points available will therefore be one more than the number of Tests played in a rubber. This total is the number by which the matches column will increase in the rankings table.

Rating Points Formula

The series result derived above is then converted into a certain number of rating points for each team. These points are then added to their existing total and used to generate their updated ratings.

The number of rating points a team scores for a given series result will depend on gap between the ratings of the two teams going into the match.

There are two different formulas. One applies if the gap between the two teams' ratings is less than 40 points, the other if the gap is 40 points or more.

Case 1 – if the gap between the two teams' ratings at the start of the series is less than 40 points:

The number of rating points scored by each team equals: (series points scored) multiplied by (50 points more than your opponent's rating) plus (series points conceded) multiplied by (50 points less than your opponent's rating)

Case 2 - if the gap between the two teams' ratings at the start of the series is 40 points or more: The number of rating points scored by the stronger team equals:(series points scored) multiplied by (10 points more than your own rating) plus (series points conceded) multiplied by (90 points less than your own rating)

The number of rating points scored by the weaker team equals: (series points scored) multiplied by (90 points more than your own rating) plus (series points conceded) multiplied by (10 points less than your own rating)

These figures are then added to the existing points. This is divided by the new match/series total to produce the new rating.

Actual Numerical Example

The following example illustrates how the formulae was applied in practice to a recent series.

India v West Indies, November 2002

India (rated 92 at the time) played West Indies (rated 80) in a three Test series. India won two Tests with the other drawn.

The series result for rating purposes was India 3.5 West Indies 0.5. (India earned one point for each of the Test wins plus a point for winning the series. Both sides earned half a point for the drawn Test).

The gap between the teams’ ratings was less than 40 so the formula in Case 1 applies

The rating points scored by India were: 3.5 (their series points scored) multiplied by 130 (50 more than West Indies’ rating) plus 0.5 (their series points conceded) multiplied by 30 (50 less than West Indies’ rating)which equals 470

Similarly, West Indies scored:0.5 (their series points scored) multiplied by 142 (50 more than India’s rating) plus 3.5 (their series points conceded) multiplied by 42 (50 less than India’s rating)which equals 218

Before the series, the table would have shown India with 3297 points from a match/series total of 36 and a rating of 92 and West Indies with 2549 points from a match/series total of 32 and a rating of 80.

The next step is to add India’s 470 points and West Indies 218 points from this series to the table.

India then have 3767 points from a match/series total of 40, increasing their rating to 94 while West Indies have 2767 points from a match/series total of 36, with their rating falling to 77.

Note that the match/series total has increased by 4 (India 36 to 40, West Indies 32 to 36). This equals the series points available, corresponding to the three Tests played plus the bonus point for the series result.

(Adopted from ICC's press release.)
If you want full text of ICC's press release, send me an email at ujjwol@clicknepal.com

Posted by Ujjwol