Combinations to Decide Group A

The final round of Group A in the Champions Trophy is gearing up as a really interesting couple of games. New Zealand and England just played 2 very good ODI series, the first won 2-1 by England in New Zealand, the second won 2-1 by New Zealand in England. Australia and Sri Lanka have played each other a lot in the last 2 and a half years. They have played 18 ODI matches and both teams have won 8 each (with 2 no results).

So recent form has nothing to suggest who is likely to win. But both games are crucial.

There were 6 possible options as to which two teams went through at the start of the round. Five still remain. Here's a quick guide to what the different results would mean:

If you're looking for a calculator to test the possible outcomes, try this link.

Result	New Zealand	No result	England
Australia	New Zealand go through first and Australia goes through as runner up.	New Zealand top the pool, and then the runner up will be decided between England and Australia on Net Run Rate(NRR). Unless Australia win by about 100 runs this will be England	England go through first and then the runner up will be either New Zealand and Australia again determined by NRR. It will take something quite remarkable for Australia to get past New Zealand, but both teams are quite capable of being involved in one-sided matches, in either direction
No Result	New Zealand go through as first, Sri Lanka go through as runners up.	New Zealand go through first, England go through as runners up	England win the pool, the runners up is decided between New Zealand and Sri Lanka on NRR. New Zealand would have to lose significantly to England for Sri Lanka to go through
Sri Lanka	New Zealand go through as winners and Sri Lanka go through as runners up	New Zealand go through as winners and Sri Lanka go through as runners up	Sri Lanka and England go through. The winner of the group will be decided by NRR. England are likely to go through first, unless Sri Lanka have a very convincing win over Australia

Every combination except for Australia and Sri Lanka is possible. The question now is which of the 5 remaining combinations will go through.

More Net Run Rate issues

In my last post I looked at the problem with using net run rate in games where both teams lose a number of wickets. Only 2 days later the tournament threw up possibly the best counter-example to the net run rate system yet. New Zealand won an absolute cliff hanger over Sri Lanka. Even on the last ball there was a question of if the game was a tie or a New Zealand win. However, on the points table New Zealand were the most dominant of any team.

Because the game ended in the 37th over, New Zealand are recorded as winning with a net run rate of +1.048. The most comprehensive victory of the round, (England over Australia) only got +0.96. This is clearly not right.

Net Run Rate strikes again

Again in a big tournament, where net run rate is quite likely to be called on to separate teams, it has been exposed as an insufficient way to look at difference in performance. The West Indies vs Pakistan match was very close. When Mohammed Irfan dismissed Sunil Narine (only a couple of balls before the match was over) the game could have gone either way. West Indies were certainly favourites at that point, but they were not in a commanding position.

However they ended up winning with a commanding difference on net run rate. The match goes down as Pakistan scoring 170 in 50 overs and West Indies scoring 171 in 40.4 overs, giving West Indies a net run rate of +0.68.

+/- 0.68 is the same as a team scoring 200 and then restricting their opposition to 166. I'm not sure that these two results deserve to be weighed similarly. It is effectively a team batting first and scoring 34 runs more than their opponents. Instead I'd suggest that a better system needs to be used.

One possible suggestion is to use a modified version of Duckworth-Lewis. Duckworth-Lewis tells us how many resources a team had left. I don't have access to the professional version of Duckworth_Lewis, but using my modified version of their amateur system I found that West Indies were on track for 193. If we were to give West Indies +23 and Pakistan -23 it would make more sense to me.

There is still an issue with a blow-out, where (for example) a team can win by a huge margin, and therefore be uncatchable, but this encourages teams to go for it, and means that the games have something riding on them right down to the end. There are probably other problems, particularly with rain affected matches, but I think taking into account wickets and overs is better than just looking at overs used.

I'm sure that Pakistan fans will agree with me at the moment, particularly if they miss out on the semi-finals on net run rate by a very small margin.