May 022011
 

This is a companion blog entry to the one where I merely published the table of fatality statistics. That article was the raw facts; this one is where I can whitter on about anything I please – ideally backed up with some analysis of the raw statistics.

One of the things that became apparent to me as I worked on the table, was that the Grand National of old was not anywhere near as fatal as one would assume. Over the years, all sorts of things have been tried to make the Grand National safer – removing stone fences (!), removing the ploughed fields, reducing the heights of fences, etc. Yet it doesn’t seem to have made that much difference.

Over the last 20 years, there have been just 7 years without fatalities. In the 35 races for which I have details ran before 1900, “just” 10 had fatalities in. So we have gone from a majority of races (in the earliest supposedly most dangerous era of the Grand National) being fatality free, to a state where the majority of races do have fatalities. So much for making things safer.

Looking more closely, we can average out the fatality rate over time. The average fatality rate over the whole period for which I have figures comes to 1.70%. This compares to an average rate of 2.05% over the last 10 years (2001-2011), and 3.2% for the 10 years before that. So I guess the rate is falling, but it is still well above the average over time. If we go back earlier in time, we have a rate of 2.8% for the years 1950-59 (including the infamous 1954), or a rate of 1.4% for the years 1930-1939.

There is a great deal more that could be done with the figures … not least of which is to chase down the figures for the missing years. However what seems to be the case is that what has happened over time is that the race has been made easier and not safer.

So How Can We Make The Grand National Safer ?

I am totally unqualified to make any sensible suggestions, but someone who just criticised without trying to be helpful is nothing but a whinger, so I will try …

Let me repeat that overall statement – the Grand National has been made easier and not safer over the years. That is of course subject to debate, but let us assume it is true. How does making a race easier, make it less safe ? Well, simply if you make a race easier you make it possible for horses to run faster. When horses have accidents they are more likely to have fatal accidents the faster they are going. By making the race easier, we have let the horses run faster and so make it more likely they will have fatal accidents.

So make the race harder – higher fences, plough some of the track, etc.

Make it easier for horses to give up and disqualify themselves – a horse who has had enough should be given the opportunity to say “No, thanks. I’ll pass on this one”.

Apr 242011
 

The following is a rather large table giving the fatality statistics for the Grand National at Aintree over time. There are a considerable number of missing years, but this is as good as I can do given the limited time to work on this. Some explanation can be found at the end :-

Year Runners Finished Fatalies Source Fatality Rate
1839 17 10 1 W 5.88%
1840 13 4 0 W 0.00%
1841 11 10 0 W 0.00%
1842 15 10 0 W 0.00%
1843 16 9 0 W 0.00%
1844 16 9 0 W 0.00%
1845 15 4 1 W 6.67%
1846 22 5 0 W 0.00%
1847 26 7 0 W 0.00%
1848 29 5 3 W 10.34%
1849 24 6 3 W 12.50%
1860 31 #N/A #N/A T 0.00%
1861 9 4 #N/A T 0.00%
1864 25 9 #N/A T 0.00%
1866 30 6 #N/A T 0.00%
1868 21 12 1 T 4.76%
1869 22 10 #N/A T 0.00%
1870 23 12 #N/A T 0.00%
1871 25 #N/A #N/A T 0.00%
1872 25 10 1 T 4.00%
1873 28 14 1 T 3.57%
1874 22 12 #N/A T 0.00%
1875 19 10 1 T 5.26%
1876 19 11 #N/A T 0.00%
1877 16 10 #N/A T 0.00%
1879 18 9 #N/A T 0.00%
1880 14 10 #N/A T 0.00%
1889 20 12 #N/A T 0.00%
1890 16 5 #N/A T 0.00%
1891 21 6 1 T 4.76%
1892 25 16 #N/A T 0.00%
1893 15 8 #N/A T 0.00%
1894 14 9 1 T 7.14%
1895 19 11 #N/A T 0.00%
1896 28 7 #N/A T 0.00%
1900 16 11 #N/A T 0.00%
1902 21 11 #N/A T 0.00%
1903 23 7 1 T 4.35%
1904 26 9 1 T 3.85%
1907 23 8 1 T 4.35%
1909 32 18 #N/A T 0.00%
1910 25 5 #N/A T 0.00%
1911 26 4 #N/A T 0.00%
1913 22 3 #N/A T 0.00%
1914 20 8 #N/A T 0.00%
1915 20 8 #N/A T 0.00%
1916 21 12 #N/A T 0.00%
1919 22 11 #N/A T 0.00%
1920 24 4 #N/A T 0.00%
1921 35 4 #N/A T 0.00%
1922 32 5 2 T 6.25%
1923 28 7 #N/A T 0.00%
1924 30 8 #N/A T 0.00%
1925 35 9 #N/A T 0.00%
1926 30 13 #N/A T 0.00%
1927 37 7 #N/A T != T 0.00%
1928 42 2 0 T 0.00%
1929 66 6 #N/A T 0.00%
1930 41 6 #N/A T 0.00%
1931 43 12 2 T 4.65%
1932 36 9 1 T, or T (5 finishes) 2.78%
1933 34 #N/A #N/A T 0.00%
1934 30 #N/A #N/A T 0.00%
1935 27 6 #N/A T 0.00%
1936 35 7 1 T 2.86%
1937 35 5 #N/A T 0.00%
1938 36 13 1 T 2.78%
1939 37 11 #N/A T 0.00%
1940 30 11 #N/A T 0.00%
1946 34 6 1 T 2.94%
1947 57 18 1 T 1.75%
1948 43 15 #N/A T 0.00%
1949 43 11 1 T 2.33%
1950 49 5 1 T 2.04%
1951 36 3 #N/A T 0.00%
1952 47 10 1 T 2.13%
1953 31 5 2 T 6.45%
1954 29 9 4 T, T, [0] 13.79%
1956 29 9 #N/A T 0.00%
1957 35 11 #N/A T 0.00%
1958 31 7 #N/A T 0.00%
1959 34 4 1 T 2.94%
1960 26 8 1 T 3.85%
1961 35 14 #N/A T 0.00%
1962 32 17 #N/A T 0.00%
1965 47 14 #N/A T 0.00%
1967 44 18 0 T 0.00%
1970 28 7 1 T 3.57%
1971 38 13 #N/A T 0.00%
1972 42 9 #N/A T 0.00%
1976 32 #N/A #N/A T 0.00%
1983 41 10 0 W 0.00%
1984 40 23 1 W 2.25%
1985 8 6 0 W 0.00%
1986 12 10 0 W 0.00%
1987 44 22 1 W 2.27%
1988 40 9 0 W 0.00%
1989 40 14 2 W 5.00%
1990 38 20 2 W 5.26%
1991 40 17 1 W 2.50%
1992 40 22 0 W 0.00%
1993 39 7 0 W 0.00%
1994 36 6 0 W 0.00%
1995 35 15 0 W 0.00%
1996 27 17 1 W 3.70%
1997 36 17 2 W 5.56%
1998 37 6 3 W 8.11%
1999 32 18 1 W 3.13%
2000 40 17 0 W 0.00%
2001 40 4 0 W 0.00%
2002 40 11 2 W 5.00%
2003 40 14 1 W 2.50%
2004 39 11 0 W 0.00%
2005 40 21 0 W 0.00%
2006 40 9 1 W 2.50%
2007 40 12 1 W 2.50%
2008 40 15 1 W 2.50%
2009 40 17 1 W 2.50%
2010 40 14 0 W 0.00%
2011 40 19 2 W 5.00%
2012 40 15 2 BBC 5.00%
2013 40 17 0 BBC 5.00%

First of all, the source columns has the values BBC (for the BBC News website), W (short for WikiPedia) and T (for The Times newspaper). Some of the sources columns have multiple sources (far too few) with an indication of whether the sources agree with each other or not. The “[0]” in the sources column for the 1954 race refers to a quote that is worth duplicating here :-

“Worst day anyone can remember for fatalities.”

Some of the columns have values in that look like “#N/A” … as you might expect, this is the value for “not available”. This is counted as zero when performing calculations on the “fatalities” column. There are three reasons for assuming “n/a” can be assumed to be usually zero :-

  1. It is clear when reading the race reports from The Times, that fatalities were unusual and there is every indication that the reporter made a point of mentioning them.
  2. It would be unusual to say the least to make a point of indicating that a race had no fatalities – when was the last time that a rugby match report pointed out there were no fatalities ?
  3. The report on the 1954 Grand National made it clear that this was the worst Grand National for fatalities “as long as anyone can remember”, which although does not indicate that the previous races had no fatalities, does indicate that 1954 was the worst year for a long time (actually from what I can find, probably the worst ever) and that very few involved in horse racing takes fatalities lightly.

That is not to say that there are not some additional fatalities that I have missed. I am only human and could easily have missed something, and it is certainly possible for reports of fatalities to be missing. However it is noticeable that even the earliest races where you could expect a certain more casual attitude towards the death of horses, that fatalities were very clearly pointed out.