The Grand National: a look at last year’s BHA Review Group report. Do we need a speed limit?
“The Review process has been balanced, thorough and comprehensive. The Review Group, and the BHA Board, are confident that the 30 recommendations will help achieve the objectives of not only enhancing the safety and welfare of participants but also maintaining public confidence in both the sport, and the Grand National as a race.”
So said Tim Morris, Director of Equine Science and Welfare at the BHA after that organisation’s review of the 2011 Grand National in which 2 horses died and 19 others failed to finish.
Last Saturday 2 horses died and 23 others failed to finish signifying that the enhancement of safety and welfare target was not reached. A victory could be claimed on the ‘maintaining public confidence’ aspect as attendances, TV viewing and betting turnover was up.
But given that the review was so thorough and comprehensive, what else can be done? What, if anything, did the review committee miss or get wrong last year?
In this article, I’ve listed selective extracts from the Review Group’s report, concentrating on the aspects most commonly discussed since Saturday:
Speed over the first few fences
The drops on the landing sides
At the end I offer some potential solutions to the speed issue by way of stirring up some creative thinking on the part of reformers.
NB, from here on in this document, any text not in italics is extracted directly from the review document; italicised text represents my comments.
Extracts from the Grand National: A Review of safety and welfare, published November 2011
Since 2000, the race averages 28.39% fallers, compared to 21.48% for the other four races staged on the Grand National circuit.
The fence-by-fence Grand National faller data since 1990 highlighted that, the first 1 minute 35 secs up to and including jumping Becher’s Brook (Fence 6), accounts for over 53% of all falls in the race and 28% of unseated riders.
Fence 1 appears to exhibit a particular trait inasmuch as when it is jumped as the very first fence in the race its rates of 21.6% of all falls and 8.1% of all unseats compare with 0% for both categories when it is jumped on the second circuit (Fence 17).
Clearly, a significant number of runners will not set out on the second circuit having already fallen or pulled up but the Review Group
believes it is still a striking comparison and feels that it can at least in part be explained by the fact that most of the runners will never have seen an obstacle like a Grand National fence before.
On that basis, it supports a proposal made by the Aintree Executive that they seek to construct an Aintree-style fence at each of the major training centres and encourage trainers to school their runners over it. This approach was previously adopted after the last major regulatory review of the Grand National in 1998. But there is a need to re-invigorate this practice.
In view of the unique fence design of the Grand National fences, the Aintree Executive shall again liaise with all major Jump training centres to develop the construction and encourage the use of a well maintained Aintree-style schooling fence for trainers to use at each centre.
Was this recommendation adopted post-1998? If not, who was responsible for policing it? Have the post-2011 recommendations been put in place at training centres? You will see later in this piece that the review group were not happy to make decisions based on assumptions and yet they willingly do so here in regard to the ‘surprise’ to some horses of seeing a GN fence for the first time.
This ‘lack of experience factor’ was a view I shared until I read a post on TRF by the forumite known as Venture to Cognac. His research showed that a long list of horses with experience of the fences failed to get round on other occasions. Some, of course, failed at their first attempt, but many found their previous experience to be of little value. That list includes 16 winners of the Grand National. VTC makes the point too that his research highlights the fact that raising the standard of horses by way of ratings, won’t necessarily make much difference to the number of finishers.
Back to the report’s findings . . .
Recurring fall types
It was apparent that there was a recurring type of fall at two particular fences. At Fence 1, where in very recent times there have actually been few Grand National fallers (three in the past five years), those horses that fell had a tendency to overjump the obstacle and crumple on landing some distance further away from where horses would usually be expected to land. The same manner of
landing was not apparent when the runners jumped the fence on the second circuit, as the seventeeth fence of the race.
Reinforcing the possibility of a “first fence jumped” trend is the fact that the 1990 – 2011 Topham races (run on day two of the three-day Grand National Meeting over a distance of 2 miles 5 1/2 furlongs) has produced eighteen fallers at the first in the Topham (i.e. Fence 13 of the Grand National course) out of 112 in total and yet Fence 13 is not at all a higher risk fence when jumped in the Grand National.
Similarly, Fence 1 on the Grand National course – which is jumped as the fifth fence in the Topham – has had no falls or unseated riders whatsoever in the Topham since 1990.
Of further interest to the Review Group when looking at the Topham faller/unseated data is that the Grand National Fence 4 and Becher’s (in particular) again demonstrate faller and unseat percentages that are higher than all but the first in The Topham, i.e. Fence 13 in the Grand National. This is despite the fact that they are jumped as the 8th and 10th Fences respectively in the Topham.
Jockey feedback from the consultation sessions essentially stated that all the Grand National fences looked and rode well, and that very little, if anything, needed to be changed.
When presented by Review Group members with a) the faller statistics for Fences 1, 4 and Becher’s (Fence 6) and b) options for change, the jockeys acknowledged the logic of exploring a possible reduction in the effective drop of these obstacles as they were clearly amongst the fences with the highest faller rates.
Going too fast from the off?
The uniquely long run of 420 yards to the first fence – coupled with its higher than normal percentage of fallers (albeit less in recent years), many of which fell by over-jumping the obstacle – appears to indicate that speed is a risk factor in the early stages of the Grand National.
2000 – 2011 split timings data to each of the first ten fences was compiled for the Review Group with a view to establishing whether there was any clear correlation between the Going, early pace of the race and the number of early fallers/injuries.
However no such clear correlation appears to exist across the relatively small sample size of twelve races.
For instance, the fastest run to the first fence in the data set was 27.44secs in 2000 on Good Going. This resulted in five fallers. Yet the third slowest run to the first (in 2002: 29.00secs, also on Good Going) resulted in eight fallers and one unseated rider. Similarly, the 2000 Grand National was the fastest (of the twelve assessed) to Becher’s Brook and by the time that obstacle had been jumped there had been ten fallers; the 2002 running remained the third slowest to Becher’s but it too had seen ten fallers and two unseats after that fence.
Of the twelve races, the 2011 race holds a middling position of being the fifth slowest to the first and the fifth fastest to Becher’s Brook. In the 2000-2011 period the two renewals (2003 and 2005) with the least fallers/unseats up to and including Becher’s Brook were, respectively, the seventh and fourth fastest to reach the fence. Clear correlations between early speed and the Going and/or fallers are therefore not apparent.
Perhaps the group were seeking too many correlations here and relying heavily on accurate going descriptions for a specific section of the course – the first six fences which, it is worth repeating, have claimed 53% of total race fallers and 28% of total race unseats since 1990. From a fence-count viewpoint, 20% of the fences here have accounted for a large % of falls/unseats. On a time basis – duration of the race at standard time – 17.6% of the duration resulted in 53%/28% group of falls/unseats.
Back to the findings . . .
However, the Review Group supports the Aintree Executive’s plan to investigate the introduction of even more irrigation capability along the section of the Grand National course from the Melling Road to Becher’s Brook. The flexibility of being able to apply extra targeted irrigation to soften or slow down the ground, can only be a positive measure.
The Aintree Executive should investigate the feasibility of introducing additional irrigation capability to the section of the Grand National course running from the start along to Becher’s Brook. As long as irrigation is applied judiciously, with a view to providing Going just on the softer side of Good, there is no downside to seeking to implement an even more flexible watering capability along the part of the track where the majority of falls occur.
How would horses react from going from softish ground to good ground after fence 6 and, possibly more importantly, meeting that ground again on circuit 2? Also, were the weather to take a sudden late turn for the worse, what would the effect be on that section of track?
Notwithstanding the lack of clear statistical correlation between early speed and number of early fallers, the Review Group is still of the opinion – having reviewed the TV footage of all Grand Nationals from 2000 and listened to participant feedback – that the pace over the initial fences in the race is certainly faster than in any routine long-distance Steeplechase over traditional birch fences.
This pace appeared to be maintained up to and including the jumping of Becher’s Brook (Fence 6).
The Review Group and Aintree Executive concluded that more specific sectional timing research would be helpful in this area to fully understand the effects of early speed on the number of finishers in the race. The Group supports Aintree’s plan to investigate the possible use of speed and positioning technology (i.e. sectional timing equipment carried in the number cloth of every runner) to track the speed of all runners in future. This would improve statistical analysis of the pace of the race so that any correlations can be
drawn from the data.
The race is run just once a year so this seems to me a slightly daft proposal. How long would it take to build reliable data? Given the furore raised by the last two runnings, we simply do not have time to wait ten years or more.
Currently, the Group can only make a subjective judgement on the basis of a) fairly basic split time data and b) TV footage – that the over-jumping falls at the first fence and high faller rate up to and including Becher’s Brook are due solely to the faster early pace of the Grand National in general when compared to more “routine” staying Steeplechases on other British licensed racecourses.
A blog post by Matt Bisogno featured this observation:
But I think there is a bigger issue that has not yet been adequately addressed, and I have a radical proposal to help address it. The issue is that of speed in the early part of the race. It has long been held that the way to win the Grand National is to be prominent early through a mad gallop, and to cling on late when stamina is running out.
Consider this: Neptune Collonges was last from the start and not prominent until Bechers second time (as the above image shows), so there is no necessity to be close up early.
More importantly, consider this: the first two furlongs of the Listed Further Flight Stakes, a 1m6f flat race, were run in around 27.5 seconds (hand timed) last week.
The approximately two furlong run* from the start of the Grand National to the first fence was completed this year in 26.5 seconds. Last year, it was a slightly more measured 27.6 seconds (all hand timed).
This is patently too fast, and extremely dangerous. And it creates a problem of momentum: once a rider has a horse travelling at that pace, trying to establish a position and a rhythm in the race, that rider must maintain the pace. Or at least feels he must.
*It’s 20 yards short of two furlongs
Matt goes on to suggest moving the start forward by a furlong, therefore reducing the race distance to 4m 3f. The Review group did consider moving the start . . .
Options for Managing Initial Race Speed
In the meantime, the Review Group still wished to consider whether there were options that could be implemented now to materially reduce the initial speed. These were discussed with the sport’s participants.
The possibility of reducing the run to the existing first fence by bringing forward the start position found no support whatsoever from the jockeys consulted. They believed that to have any effect the start would need to be approximately 110yds from the first fence and this would result in less time for all the runners to find room before the obstacle. They felt that this could have the unintended consequence of increasing the number of incidents at the first fence. Some of the jockeys also felt that the pace would just rise soon after jumping the first if the run to it were reduced. They also pointed out that few runners are ever being vigorously ridden or pushed along “off the bridle” as they approach the first fence.
The majority of trainers consulted believed the start position should remain unchanged. However, there was some support for reducing the distance to the first fence on the basis that this approach was adopted in the Topham Chase from 2005 when the run to the first fence from its then two miles six furlongs start was reduced by half a furlong. There have been four fallers and just one unseated rider at the first fence in the subsequent seven renewals of the Topham Chase from the new start. Albeit it is probably too early to conclude statistically that this improvement is purely due to the new start position.
The trainers also believed that the jockeys had a responsibility to ride the Grand National sensibly at a maintainable gallop and that this should be emphasised at their pre-race briefing. (Er, I think we know by now folks that this is simply not going to happen and there is plenty of evidence to back that up!)
The members of the Authority’s Course Inspectorate within the Review Group have reservations as to where a substantively shortened start position could be suitably located. Therefore, they did not support a reduced run to the first fence from 2012. Similarly, they do not believe there is real scope to significantly and safely bring forward Fence 1 towards the current start location, due mainly to the position of the Melling Road. Neither of the participant groups had supported that option when consulted.
They didn’t support what seems a very sensible idea because it presents a practical/logistical difficulty?
The concept of an additional, smaller (but still Aintree-style) fence between the current start position and first fence was also discussed with the participant groups and within the Review Group. This was considered on the basis that it could help to decrease initial speed and then be removed ahead of the runners returning on the second circuit. The idea of a “sighter” fence was not supported, however, with most consultees believing it would simply increase the fundamental level of risk by effectively creating a 31st fence to negotiate, as well as provide less time for the jockeys to find racing room. Course topography also ruled out this option.
Whilst the possibility of bringing the current first fence closer to the current start position (or vice versa) found little support amongst the participant groups and brings with it practical challenges and potentially unintended consequences, both options should remain under close consideration beyond 2012. The impact of the new changes to Fences 1, 4 and 6 (Becher’s Brook should dictate whether the start/first fence dynamic still needs to be altered in future.
The RSPCA among others, is keen to eliminate drop fences; according to the review group, that means altering almost every fence. Back to the findings . . .
Another unique aspect of the Grand National course fences is that virtually all of the obstacles have a “drop” to some degree when measuring the height difference between the ground level at the take-off area and the ground level on the (lower) landing side of the obstacle. The professional survey work carried out since this year’s race shows that fourteen of the sixteen fences have an average drop of over four inches, when measured at five metre intervals across the width of the landing area, with the biggest being at Becher’s Brook (thirteen inches).
At Becher’s Brook (i.e. Fence 6 and 22) - the obstacle with the biggest drop on the landing side – the clear reason for most jockeys and horses parting company involved the horse being angled by the rider from a position opposite the middle of the fence towards the inner at take-off and either: • making a mistake and taking a very steep or rotational landing trajectory with the jockey often landing feet first, or; • jumping the fence well but nodding on landing and falling or unseating the jockey whilst sliding to a halt along the
At the Review Group’s consultation meeting with the jockeys, they reported that the methodology for starting the Grand National was good and they did not believe there was any need to change it. However, they all agreed that the horses should be on course at the start for as short a time as possible after the official Parade had taken place.
There was no suggestion from any of the participants consulted that the physical size of the start area negatively impacted on fairness or the welfare of the runners. It was noted, however, that the proximity and nature of the grandstands at Aintree contributed to high crowd noise levels as the runners approached the starting tape or if there was any perceived delay. By extension, the position of the start was also considered in relation to whether the initial pace of the race was a contributing factor to falls or injuries.
There is no doubt that loose horses can be a major danger to themselves, other participants or even Emergency Service personnel or spectators at any race meeting. Since 1990, three horses (16% of the total) have died during or very shortly after the Grand National from injuries sustained whilst riderless. Furthermore, it is impossible to plan exactly for what a loose horse might do next. Consequently, it was important for the Review Group to clearly understand how riderless horses are managed by the Aintree Executive during the Grand National – particularly in the context of such a large footprint of flat land.
The Review Group fully appreciates the difficulties of controlling a unique site like Aintree and trying to catch all the loose horses in a timely manner. Since 2000, on average eighteen horses part company with their jockey during the race. Many will stop immediately and be caught straight away by the jockey, fence attendant, or horse-catcher. However, some do not, and it is important that the Aintree Executive does everything it can in this vital area.
The Aintree Executive informed the Review Group that on Grand National day a team of around 30 local horsemen are allocated sectors of the course, which they patrol to catch loose horses during and after the race.
Limiting the number of runners
It was clear to the Review Group from its analysis of all the TV footage of all the professional races staged on the Grand National course since 2000 that three incidents of multiple fallers/unseats/brought downs/refusals have occurred during the period reviewed:
• Fence 8 (Canal Turn), 2001: Nine horses;
• Fence 1, 2002: Nine horses;
• Fence 6 (Becher’s Brook), 2004: Eight horses
Incidents involving that number of runners are rare at other licensed Jumps racecourses, including Aintree’s Mildmay Course, and could therefore simply be a function of the Grand National fence design. At the same time, injury rates (on the basis of five years of nationwide Jump data) do appear to show an upwards trend as the numbers of runners increase, although this has not been validated by a statistical analysis, probably because of the small sample size.
A number of points suggesting a Safety Factor reduction to between 30-34 were made by the welfare organisations in their feedback to the effect that: • it is logical that if the number of horses exposed to the risk factors of the race is reduced, so too will the number of injuries and the likelihood of loose horses causing incidents; • no other Jumps race has a Safety Factor higher than 30 and yet the
Grand National’s is 33.33% greater than that figure.
The delegations of trainers and jockeys consulted by the Review Group unanimously supported the retention of a Safety Factor of 40. (Turkeys, voting and Christmas are words the review group might have considered on hearing this.)
The Review Group found no recurring trend whatsoever of horses systematically failing to get a clear sight of the fences as they prepared to jump them. Virtually all the fallers reviewed during that period had a clear run to the fence where they fell or unseated their jockey.
Furthermore, the Review Group considered research carried out through its Inspectorate team and established that the average available “width of fence per horse” on the Grand National course was comparable to the averages for all licensed Jumps courses, including the width of fence per horse at other very high profile jumps fixtures.
(NB, from here, my comments are no longer in italics)
So where now for the BHA and for the world’s greatest race? Pressure from the public/media for a reduction in field size is the change most likely to be resisted by trainers and jockeys. If drop fences are to be altered again, they will need to consider which ones and to what degree they will change them. The RSPCA seem strongly opposed to these drops and their fairly new Chief Exec, Gavin Grant could well push for the complete elimination of all drops. Racing should not, I believe, underestimate Mr Grant’s ambition for change. In a Radio 4 phone-in on April 17th, he said “”Unless the BHA really respond here, and are seen to respond, I think the days of NH racing and the Grand National are numbered”
The Review Group’s suspicion that speed over the first six fences plays a large part in non-completions was not fully reinforced by the 2012 stats (36% of fallers, 16% of unseats) but given that the figures are based on data since 1990, the group will be under pressure now, I believe, to act ’on the balance of probabilities’ rather than trying to gather further data by way of technology. With the first 95 seconds of the National accounting for well over 50% of fallers/URs combined, that part of the race simply must be slowed. But how?
Well, they might try the selective watering mentioned in the report. Or they could opt for much more radical solutions like setting speed limits for that section, but how would you enforce any limit?
Maybe a rule could be brought in decreeing that any horse landing over the first in under 33 seconds is disqualified: touching down over Becher’s in under 1 minute 50 (15 seconds longer than the average), means disqualification. Large digital clocks could be set high above each side of the first six fences . . .
What about replacing the turf on that 420 yard run to the first with a deep all-weather type surface, consistent and resistant to temperature and rain?
Or perhaps running a lead vehicle on the inside track just after Melling Rd, travelling at a pre-agreed speed with the jockeys instructed not to pass, under penalty of disqualification, till after Becher’s? This is very practical from a logistics viewpoint: the old Grand Prix track at Aintree is still in excellent condition. Its back straight runs close enough to that line of fences for jockeys be to be able to see it easily, without it being a distraction. In practise, I think the vehicle would need to pull a long trailer – with a large board/sign at the rear (good branding opportunity for the sponsors!)
What would you do? (Please leave your ideas in the Comments section below. Perhaps Aintree will pay a nice fat fee to anyone coming up with the answer!)
The full Review Group report is here