RagingLamb

01-29-2012, 06:38 PM

This thread confirms what most of us have noticed:

It seems that as the years go by, more and more matches end with players retiring due to an injury (e.g. 12 at the USO last year).

I tried to look at the rate at which this happens at the slams.

The table below contains the number of retirements (and walk overs) at RG, USO and Wimbledon since 1978, and at AO since 1988 (corresponding to when AO's surface changed from grass to hard courts). There were very few walkovers, so most of these numbers represent retirements during the match:

Year AO RG Wim USO Average

1978 4 2 2 3

1979 1 1 7 3

1980 2 0 3 2

1981 3 0 2 2

1982 2 0 3 2

1983 3 4 1 3

1984 1 3 1 2

1985 0 0 3 1

1986 2 1 0 1

1987 1 0 1 1

1988 3 1 1 2 2

1989 2 0 0 3 1

1990 4 2 0 4 3

1991 1 6 1 5 3

1992 3 1 3 7 4

1993 2 0 3 7 3

1994 7 6 4 4 5

1995 3 2 3 4 3

1996 6 4 1 4 4

1997 5 5 2 4 4

1998 9 5 2 6 6

1999 3 3 2 7 4

2000 3 6 4 3 4

2001 2 6 6 3 4

2002 5 5 3 10 6

2003 9 7 4 5 6

2004 7 6 1 7 5

2005 6 5 4 4 5

2006 4 9 3 2 5

2007 6 8 4 7 6

2008 4 1 10 6 5

2009 4 6 7 4 5

2010 6 5 5 8 6

2011 6 3 6 12 7

I computed the correlation between years and number of retirements at each major, and got the following r-coefficients:

AO: 0.43 (p = 0.04)

RG: 0.59 (p = 0.0003)

Wim: 0.69 (p = 0.000054)

USO: 0.57 (p = 0.0005)

*See note below for more details about the stats.

This means that as the years increase, so do the number of retirements at every major.

I'd be interested to know why you guys think this is happening. Are the athletes playing harder? Is the schedule too busy?

Or do you think it has to do with the slowing of the surfaces (e.g. look at the numbers for Wimbledon)?

I don't think it's one factor by itself. For example, if it was just the slowing of surfaces, then why do we see the trend at RG as well? The consensus seems to be that RG was sped up over the years.

My guess is that the homogenization of the surfaces does have something to do with it, but so does the accompanying homogenization of playing styles. There seem to be an abundance of grinders these days. Tennis matches have become increasingly more physical, and the lengths of rallies seem to have increased.

It would be interesting to take a sample of wimbledon matches from the early 1990s and compare their lengths with recent wimbledons to see how they compare. But the ATPs duration values from the 90s don't seem reliable.

Anyhow, I'd like to hear why people think the number of retirements are increasing at the slams.

__________________________________________________ _______________

*Ideally I would have used an ANCOVA or a multiple regression, but I did not have the necessary software to do so.

Also, I computed for r-coefficients without correcting alpha. But even applying a bonferroni correction to alpha = 0.01 still leaves the RG, Wim, and USO correlations significant. The overall trend in the number of injuries across years is very clear.

I also compared the correlation coefficients using a z test and found no significant difference between these values.

It seems that as the years go by, more and more matches end with players retiring due to an injury (e.g. 12 at the USO last year).

I tried to look at the rate at which this happens at the slams.

The table below contains the number of retirements (and walk overs) at RG, USO and Wimbledon since 1978, and at AO since 1988 (corresponding to when AO's surface changed from grass to hard courts). There were very few walkovers, so most of these numbers represent retirements during the match:

Year AO RG Wim USO Average

1978 4 2 2 3

1979 1 1 7 3

1980 2 0 3 2

1981 3 0 2 2

1982 2 0 3 2

1983 3 4 1 3

1984 1 3 1 2

1985 0 0 3 1

1986 2 1 0 1

1987 1 0 1 1

1988 3 1 1 2 2

1989 2 0 0 3 1

1990 4 2 0 4 3

1991 1 6 1 5 3

1992 3 1 3 7 4

1993 2 0 3 7 3

1994 7 6 4 4 5

1995 3 2 3 4 3

1996 6 4 1 4 4

1997 5 5 2 4 4

1998 9 5 2 6 6

1999 3 3 2 7 4

2000 3 6 4 3 4

2001 2 6 6 3 4

2002 5 5 3 10 6

2003 9 7 4 5 6

2004 7 6 1 7 5

2005 6 5 4 4 5

2006 4 9 3 2 5

2007 6 8 4 7 6

2008 4 1 10 6 5

2009 4 6 7 4 5

2010 6 5 5 8 6

2011 6 3 6 12 7

I computed the correlation between years and number of retirements at each major, and got the following r-coefficients:

AO: 0.43 (p = 0.04)

RG: 0.59 (p = 0.0003)

Wim: 0.69 (p = 0.000054)

USO: 0.57 (p = 0.0005)

*See note below for more details about the stats.

This means that as the years increase, so do the number of retirements at every major.

I'd be interested to know why you guys think this is happening. Are the athletes playing harder? Is the schedule too busy?

Or do you think it has to do with the slowing of the surfaces (e.g. look at the numbers for Wimbledon)?

I don't think it's one factor by itself. For example, if it was just the slowing of surfaces, then why do we see the trend at RG as well? The consensus seems to be that RG was sped up over the years.

My guess is that the homogenization of the surfaces does have something to do with it, but so does the accompanying homogenization of playing styles. There seem to be an abundance of grinders these days. Tennis matches have become increasingly more physical, and the lengths of rallies seem to have increased.

It would be interesting to take a sample of wimbledon matches from the early 1990s and compare their lengths with recent wimbledons to see how they compare. But the ATPs duration values from the 90s don't seem reliable.

Anyhow, I'd like to hear why people think the number of retirements are increasing at the slams.

__________________________________________________ _______________

*Ideally I would have used an ANCOVA or a multiple regression, but I did not have the necessary software to do so.

Also, I computed for r-coefficients without correcting alpha. But even applying a bonferroni correction to alpha = 0.01 still leaves the RG, Wim, and USO correlations significant. The overall trend in the number of injuries across years is very clear.

I also compared the correlation coefficients using a z test and found no significant difference between these values.