Saturday 14 July 2012

How to almost double Bangladesh's success rate.

In One Day cricket there are a few cliches. Some are general like "catches win matches." Others are more specific, like "keep wickets in hand for the final overs" and "it's a cardinal sin to get bowled out before the end of 50 overs." (not referring to the Filipino version.) However there is one factor that is regularly overlooked, run outs.

In their last 193 matches that ended in a result, Bangladesh have taken 116 run outs. In the matches where they haven't taken a run out they have won just over a quarter, 25.9%. In the matches where they have taken at least one run out they have won just under half, 48.2%.

A similar (but not so dramatic) difference is true for most teams. Here are the winning percentages from the last 200 matches for each team (I've excluded Zimbabwe as they have not played sufficient matches against reasonable opposition recently):

TeamNo run outsWith run outsImprovement
New Zealand44.162.742
South Africa65.963.8-3.1
Sri Lanka52.65912.2
West Indies35.44732.7

It is fairly clear that run outs make a significant difference to a teams winning percentage. Only Bangladesh and West Indies have losing records if they manage a run out, and then it's only by a couple of matches.

The surprising exceptions here are South Africa and Australia. South Africa are the team most likely to run someone out, taking .78 run outs per match, but they actually do better when they don't manage to run anyone out than when they do. There are a couple of possible reasons for this. Run-outs are often a result of pressure, and so when a team is under pressure they are more likely to get run out. However, if a team has a reputation as being very good at fielding, then teams take less risks. Taking less risks results in less runs. It may be that the times that South Africa don't get run outs is when their opponents are not taking risks, and as a result they are scoring less anyway.

Breaking the numbers down further, it is interesting to see how they work in more detail.

Here are the winning percentages by number of run outs. This covers 1634 completed innings, but there were very few innings with 3 or 4 run outs, so the numbers are less reliable for those two categories (4 run outs have occurred only 6 times in this time period.)

Team0 run outs1 run out2 run outs3 run outs4 run outs
New Zealand44.154.565.2100-
South Africa65.96166.772.7-
Sri Lanka52.656.161.980-
West Indies35.443.65066.7-

If we look at the overall rate, we can get an idea as to what a run out is worth. Every run out roughly adds 8% to the winning percentage.

However this does not adequately explain the data, as there is a big fluctuation between the better teams and the worse teams. A better option is to look at the losing percentage. It turns out that every run-out reduces the chance of losing by 19%. For example Pakistan lose 51% of their matches when they have no run outs. If they have 1 run out that is reduced by 19% to 41.3%, so we expect them to win 58.7%. If we look at the table above they actually win 58.1% which is remarkably close.

This is nice, but what does it actually mean in a game situation?

Simply it means that creating, and then taking, run out opportunities is a very effective way for a side to win games. It may be an idea for teams to actually think about how they are going to achieve run outs, rather than just hope that the batsmen make a mistake. An example of this is perhaps Martin Guptill running out Marlon Samuels in Wednesday's game at St Kitts. Guptill was sharp enough to be able to run out Samuels without Samuels even setting off for a run. It was amazing skill on the part of the fielder, but also awareness of the situation and good field placement by Williamson.

This sort of skill is as likely to turn a game as bowling a wicket maiden or hitting a quickfire cameo is, and as such it is something that captains and fielders need to think about, and plan for.


  1. Gold! When NZ were a complete fielding unit the value of the run out was often evident.

    1. One clear example of the importance of run outs was the NZ tour to Bangladesh. The Deshi's had just finished a 60 day camp where they did fielding practice for a couple of hours a day. They managed a couple of run outs in most of the matches, and really that was the difference between the sides as they won 4 close matches.

  2. One very surprising thing when doing this analysis was how good India has become at getting run outs. They managed 0.766 run-outs per match. Second only to South Africa. That certainly is not the stereotype.

  3. What a gem of a discovery! I can think of one explanation that reverses the direction of causation that you suggest (i.e. rather than run outs leading to success, success creates run outs.) If teams are struggling in a run chase, or struggling to set a competitive target, there may be more run outs as risks are taken to boost their score. Just speculation, but could it be that a product of having the upper hand in a game is that the opposition, striving to get back into the game, run riskily or carelessly?


    1. I think both are true. Often when teams are under pressure, they are more likely to try silly runs. Likewise when a fielding side is under pressure they are often less likely to hit the stumps for a direct hit.

      I could possibly go through all the scorecards and look at them to see which is the case, but I'd assume that it's about 50-50. The classic example for me was the match in 1990 when Australia needed 2 runs off the final over, and Chris Pringle bowled a maiden to win the game. He managed a run out off the last ball to seal the win, and that wouldn't have happened unless the Australians were under pressure, but New Zealand wouldn't have won unless he hit the stumps with a direct hit (with his foot).