Dissecting the “Difficulty Paradox”: An Expert Analysis of Faller, Unseat, and Refusal Rates by Course

Executive Summary

This report provides a comprehensive statistical analysis of obstacle difficulty at all National Hunt racecourses in the United Kingdom and Ireland, benchmarked against the requested metric of non-completion percentages (fallers, unseats, and refusals). The analysis distinguishes between courses that are difficult due to the intrinsic severity of their fences and those that are statistically hazardous due to other factors, such as track layout, race pace, and the average quality of equine participants.

The key findings of this analysis are as follows:

1. Aintree as an Anomaly: The Aintree Grand National Course is a profound statistical outlier. With a combined Fell/Unseated Rider (% Fell/UR) rate of 21.13%, its obstacles are demonstrably the most difficult in the sport, a fact attributable to their unique construction, scale, and the specific challenges of fences like Becher’s Brook, The Chair, and the Canal Turn.   
2. Irish Steeplechase Dominance: Beyond the Aintree National course, Irish tracks dominate the top tier of statistical difficulty. Leopardstown (12.91%) and Punchestown (11.07%) lead a cohort of seven Irish courses that rank in the top 10 for casualty rates. This is attributed to a combination of genuinely stiff fences at premier tracks , a higher prevalence of soft or heavy ground, and larger average field sizes.   
3. The “Difficulty Paradox” in the UK: The data reveals a significant “Difficulty Paradox,” where several minor UK tracks post casualty rates higher than renowned championship courses. Fakenham (9.46%), Plumpton (9.35%), and Ludlow (9.10%) all rank in the top 10. This is not due to fence stiffness, but rather to quirky, sharp track layouts, high race pace, and a higher proportion of low-grade or novice races.   
4. Hurdle Racing: A Separate Discipline: The analysis of hurdle courses reveals a fundamentally different and safer challenge. Hurdle difficulty is less about the obstacles themselves, which are increasingly standardized and safer due to the rollout of “One Fit” padded technology. The statistical difficulty of a hurdle course is now almost entirely defined by course topography (e.g., the downhill approach to the second-last at Taunton ) and stamina requirements (e.g., the stiff finish at Kelso ).   

This report will proceed by establishing a clear methodological framework, presenting the definitive statistical rankings for steeplechase courses, and providing an in-depth analytical dissection of each “Tier of Difficulty.” A final, comparative analysis of hurdle course difficulty is included to provide full context for the National Hunt code.

Section 1: Introduction and Methodological Framework

1.1 Defining the Analytical Challenge

This report directly addresses the query for an identification of National Hunt courses with the “most difficult fences,” as judged by the percentage of fallers, unseats, and refusals. To provide an expert-level answer, the components of this query must be precisely defined.

National Hunt Racing: This term encompasses two distinct racing codes:

• Steeplechasing (Chases): Races run over “fences,” which are larger, more varied, and more severe obstacles. This report’s primary focus is on steeplechase courses.
• Hurdling: Races run over “hurdles,” which are smaller, more uniform, and less severe obstacles.

These two codes are not directly comparable in terms of risk. Academic analysis has long established a significant disparity in faller rates. A 1999 study identified a fall risk of 6.0 per 100 starts in steeplechasing, compared to just 2.1 per 100 starts in hurdling. More recent data confirms this gap, with steeplechase races showing higher rates of falls, injuries, and fatalities than hurdle races. Therefore, this analysis will bifurcate the two codes, with a primary focus on fences.   

“Difficult Fences”: The Difficulty Paradox The central thesis of this report is the “Difficulty Paradox”: a high casualty rate is a symptom, not a singular diagnosis. A high percentage of non-completions can be caused by at least four distinct, and often overlapping, factors.

1. Obstacle Severity: The physical height, width, and stiffness of the fences themselves. This is the traditional definition of “difficulty” and applies to courses like Aintree and Punchestown.   
2. Track Topography and Layout: Quirky, sharp, or undulating tracks. Fences placed immediately after tight bends, on downhill gradients, or in quick succession can create high casualty rates irrespective of fence size.   
3. Race Pace: Flat, galloping tracks with perceptibly “easy” fences can encourage horses to race at a higher velocity. At speed, even minor errors at “easy” obstacles are magnified, leading to more falls.   
4. Participant Quality: Courses that host a high volume of low-grade handicap chases or novice races will invariably feature more inexperienced or less capable jumpers. This can elevate the faller rate significantly, regardless of how “easy” the fences are.   

An expert analysis must differentiate between these causal factors to provide a truly nuanced answer.

1.2 Terminology and Data Metrics

To ensure analytical clarity, this report uses specific industry-standard definitions for non-completion events :   

• Fell (F): The horse falls at an obstacle in a manner that prevents it from completing the race.
• Unseated Rider (UR): The horse successfully negotiates the obstacle without falling, but the jockey is removed from the saddle. This is often caused by the horse “making a mistake,” “jinking,” or jumping awkwardly, causing the rider to lose balance.   
• Refused (R): The horse stops at the obstacle and “refuses” to jump, resulting in non-completion.   
• Brought Down (BD): A horse and rider fall, but the incident is precipitated by another horse falling in their path.

The primary analytical metric for this report will be the combined percentage of Fell or Unseated Rider (% Fell/UR). This is the most robustly and consistently tracked statistical set available for this analysis. “Brought Down” incidents are excluded from this primary metric, as the event is not a failure of the horse’s own jumping ability or a direct consequence of the obstacle’s difficulty. While “Refusals” are part of the initial query, they are statistically less frequent and are not aggregated in the primary dataset; they will be discussed qualitatively where relevant.   

Section 2: The Definitive Statistical Ranking (Steeplechase Courses)

2.1 The Primary Dataset: UK and Irish Chase Courses by % Fell/UR

The foundational data for this analysis is derived from Proform Racing, capturing all steeplechase races in Great Britain and Ireland from 2017 to the present. The definitive ranking of all 60 courses by their % Fell/UR is presented in Table 2.1.   

This table provides the direct statistical answer to the user’s query. The data immediately reveals three critical trends that form the basis of our analysis:

1. The Outlier: The Aintree Grand National course (21.13%) is an extreme anomaly, with a casualty rate nearly double its closest competitor, Leopardstown (12.91%).
2. Irish Dominance: The top of the list is dominated by Irish racecourses. Seven of the top ten non-Aintree courses are in Ireland, including Leopardstown, Punchestown, Listowel, Sligo, Thurles, Naas, and Killarney.
3. The UK Surprises: The presence of minor UK tracks like Fakenham (9.46%), Plumpton (9.35%), and Ludlow (9.10%) in the top 10—ranking statistically “more difficult” than renowned tests like Cheltenham (7.31%)—is the most prominent example of the “Difficulty Paradox.”

Table 2.1: Definitive Ranking of UK & Irish Steeplechase Courses by % Fell/UR (2017–Present) Data sourced from Proform Racing, analyzing all chase races since 2017.   

Rank	Racecourse	Country	% Fell/UR
1	Aintree National	UK	21.13
2	Leopardstown	IRE	12.91
3	Punchestown	IRE	11.07
4	Listowel	IRE	10.70
5	Sligo	IRE	10.69
6	Thurles	IRE	10.23
7	Naas	IRE	9.78
8	Killarney	IRE	9.56
9	Fakenham	UK	9.46
10	Plumpton	UK	9.35
11	Ludlow	UK	9.10
12	Newcastle	UK	8.58
13	Haydock	UK	8.52
14	Fairyhouse	IRE	8.52
15	Limerick	IRE	8.40
16	Catterick	UK	8.35
17	Galway	IRE	8.31
18	Tipperary	IRE	8.16
19	Wetherby	UK	7.90
20	Tramore	IRE	7.86
21	Gowran Park	IRE	7.40
22	Cheltenham	UK	7.31
23	Clonmel	IRE	7.23
24	Cork	IRE	7.21
25	Taunton	UK	7.18
26	Ayr	UK	7.11
27	Wincanton	UK	6.95
28	Navan	IRE	6.92
29	Wexford	IRE	6.66
30	Towcester	UK	6.58
31	Ascot	UK	6.55
32	Chepstow	UK	6.37
33	Downpatrick	IRE	6.31
34	Perth	UK	6.31
35	Kempton	UK	6.28
36	Sandown	UK	6.22
37	Hereford	UK	6.12
38	Newbury	UK	6.05
39	Musselburgh	UK	5.96
40	Kelso	UK	5.94
41	Ffos Las	UK	5.88
42	Lingfield	UK	5.86
43	Hexham	UK	5.86
44	Roscommon	IRE	5.69
45	Doncaster	UK	5.53
46	Market Rasen	UK	5.48
47	Exeter	UK	5.42
48	Kilbeggan	IRE	5.39
49	Southwell	UK	5.35
50	Bangor-On-Dee	UK	5.31
51	Stratford	UK	5.29
52	Sedgefield	UK	5.21
53	Down Royal	IRE	5.18
54	Leicester	UK	5.05
55	Huntingdon	UK	5.01
56	Ballinrobe	IRE	4.94
57	Newton Abbot	UK	4.68
58	Aintree Mildmay	UK	4.58
59	Uttoxeter	UK	4.53
60	Carlisle	UK	4.24
61	Fontwell	UK	3.96
62	Worcester	UK	3.87
63	Warwick	UK	3.81
64	Cartmel	UK	3.40

2.2 The UK vs. Ireland Divide: A Deeper Analysis

The systemic difference between UK and Irish casualty rates shown in Table 2.1 is a critical finding. An analysis covering 2015-2022 explicitly concluded that “far more horses have either fallen or been unseated in Irish chases in percentage terms compared to UK ones”. Older datasets confirm this, with a 2003-2015 study showing extremely high faller rates at Irish tracks like Down Royal (15.2%), Listowel (12.7%), and Cork (12.6%).   

This persistent divide is not coincidental and is attributed to several key factors:

1. Ground Conditions: Softer ground is correlated with higher casualty rates, as it is more tiring for horses and can make landings more problematic. Analysis shows that 40% of Irish chases are run on ground described as soft or heavy, compared to 34.8% in the UK.   
2. Average Field Size: Irish races, on average, feature larger fields. This can lead to “more crowding or less space at fences,” increasing the likelihood of errors, interference, and bad jumps.   
3. Regulation and Fence Density: Regulatory differences in course design play a role. The minimum number of fences required per two miles is lower in Ireland than in Great Britain. For example, a two-mile chase at Naas in Ireland requires horses to jump 10 fences, whereas a shorter 1m 7f chase at Wetherby in the UK requires 13. This can alter the rhythm of a race, potentially leading to faster speeds between obstacles and a different type of jumping test.   

Section 3: Tiers of Difficulty – A Deep Analysis of Key Steeplechase Courses

This section provides the essential qualitative analysis of the quantitative data, grouping courses into “Tiers of Difficulty” to explain why they rank as they do.

3.1 Tier 1: The Outlier – The Aintree Grand National Course (% Fell/UR: 21.13%)

The Aintree Grand National Course is not just the most difficult course; it is a different category of challenge. Its % Fell/UR rate of 21.13%  is a reflection of a test that is unique in its accumulation of obstacles, drops, turns, and distance. The difficulty is an explicit part of the race’s identity.   

Anatomy of Difficulty:

• The First Fence (Fence 1 & 17): Paradoxically, this 4ft 6in obstacle is one of the smallest on the course, yet it is one of the most feared. Epidemiological analysis has shown that horses are seven times more likely to fall at the first fence than at other plain fences. This is attributed to the high speed at which the large field (historically 40 runners) approaches it.   
• Becher’s Brook (Fence 6 & 22): This is arguably the most controversial and famous fence in racing. While the fence itself is 4ft 10in, its notoriety comes from the severe drop on the landing side, historically as much as 6ft 9in. Jockeys have referred to jumping it as “jumping off the edge of the world”. Between 1960 and 2019, Becher’s Brook was responsible for 186 falls.   
• The Chair (Fence 15): Jumped only once, this is the highest fence on the course at 5ft 2in. Its challenge is technical: it is preceded by a 5-foot-wide open ditch, and the landing side is six inches higher than the take-off side, creating an “opposition difficulty” that tests a horse’s scope and balance.   
• Canal Turn (Fence 8 & 24): The difficulty of this 5ft fence is not its size but its location. Upon landing, horses and jockeys must immediately navigate a sharp 90-degree left-hand turn.   

Critical Caveat: The Impact of Modern Safety Modifications An expert analysis must acknowledge that the 21.13% figure represents a historical average. This data is already being impacted by significant, data-driven safety modifications implemented ahead of the 2024 race, specifically designed to reduce the casualty rate.

Key changes include :   

• A reduction in the field size from 40 to 34 runners.
• The introduction of a standing start and moving the first fence 60 yards closer to the start, both measures designed to reduce the approach speed.
• Lowering the height of Fence 11 by two inches.
• Adding softer foam and rubber “toe boards” to the base of all fences.
• Raising the minimum handicap rating of participating horses to 130, in theory improving the minimum standard of jumping ability in the race.   

These modifications are expected to significantly lower the future faller rate. The 2024 Grand National, which saw only three falls, suggests this trend is already in effect.   

3.2 Tier 2: The Irish Championship Test – Leopardstown (12.91%) and Punchestown (11.07%)

These two premier Irish tracks represent the pinnacle of “championship” difficulty. Their high statistical ranking is not a “quirk” but a genuine reflection of a severe test of jumping.

• Leopardstown (% Fell/UR: 12.91%): As the second-most statistically difficult course, Leopardstown’s challenge is uncompromising. Modelled on Sandown in the UK, its 10 fences are part of a relatively flat track. The high casualty rate, particularly in high-stakes Grade 1 races, suggests the fences are unforgiving of errors made at high speed. The final fence, in particular, is noted for being “particularly tricky”.   
• Punchestown (% Fell/UR: 11.07%): The data for Punchestown resolves a common contradiction. While some sources have described its fences as “not… stiff” , this is strongly disputed by expert testimony and the statistical reality. Multiple track guides, including commentary from jockey Charlie Swan, confirm the chase track is a “proper Grade 1 track” and that “the fences are stiff, but very fair”. The high faller rate of 11.07% strongly supports the “stiff” fence argument. The difficulty at Punchestown is therefore the combination of stiff fence design and the high pace of its festival races, which punishes any jumping deficiencies.   

The high ranking of other Irish tracks like Listowel (10.70%), Sligo (10.69%), and Thurles (10.23%)  reinforces the systemic factors identified in Section 2.2: a higher baseline casualty rate across Ireland exists due to ground, field size, and race composition.   

3.3 Tier 3: The UK’s Statistical Hotspots (The “Difficulty Paradox”) – Fakenham (9.46%), Plumpton (9.35%), Ludlow (9.10%)

This tier provides the clearest evidence of the “Difficulty Paradox.” These three minor UK tracks rank statistically higher for casualties than premier, “stiff” tracks like Cheltenham and Ascot. Their difficulty is a product of layout and pace, not fence severity.

• Fakenham (% Fell/UR: 9.46%): The primary cause of Fakenham’s high casualty rate is its unique and challenging layout. It is described as a “very sharp” , “quirky” , and “square-shaped” track of “barely a mile”. Analysis from Timeform suggests its high rank is “likely a consequence of its acute tightness with one fence coming up barely 100 yards after taking a near-90 degree turn”. Races are often “strongly-run,” and this combination of pace and sharp turns places a premium on agility, catching many horses out.   
• Plumpton (% Fell/UR: 9.35%): Plumpton’s high ranking is attributed to a combination of tricky layout and, crucially, a lower average quality of competition. The track is “tight, undulating”. Timeform analysis is explicit on this point: the “surprisingly high” casualty rate is “probably reflective of the low standard of horses rather than the stiffness of the fences”. This is compounded by a tricky layout, including a fence (the first in the back straight) that “falls away” on landing, a known challenge for novice jumpers.   
• Ludlow (% Fell/UR: 9.10%): Ludlow presents a “velocity” problem. While some observers, including jockey Mick Fitzgerald, describe the fences as “easy” , and other guides concur (“fairly easy” ), this is difficult to reconcile with a 9.10% casualty rate. The resolution is found in the track’s design. The chase course is a “flat” , “fast circuit” of only 1m 3f, and races are noted for being “so may races being well run”. Ludlow’s “easy” fences and fast-circuit layout encourage riders to attack them at high speed. This velocity magnifies any error, turning “easy” fences into high-casualty obstacles.   

3.4 Tier 4: Reputation vs. Reality – The Major UK Festival Tracks

This final tier addresses the inverse of the paradox: why are famously “stiff” tracks not in the top 10? The data for Cheltenham (7.31%), Ascot (6.55%), Kempton (6.28%), Sandown (6.22%), and Newbury (6.05%)  shows that their statistical faller rates are suppressed by the elite quality of their participants.   

• Sandown (6.22%): Widely “perceived as a tough test” due to its famous line of seven “Railway Fences” down the back straight, Sandown’s faller rate is modest because it “primarily attracts established jumpers”.   
• Ascot (6.55%): The data shows that Ascot attracts the fewest runners with known jumping issues (an ‘x’ rating in Timeform), meaning its races are contested by the most secure and reliable jumpers.   
• Cheltenham (7.31%): As the premier jumps course, Cheltenham’s fences are a stiff test. However, its faller rate is mediated by the high quality of the horses. Analysis has shown that if the metric is widened from just falls/unseats to include all jumping errors, the Cheltenham Old Course “features much more prominently… as would be expected of the premier jumps racecourse”.   

The fences at Cheltenham, Sandown, and Ascot are objectively difficult. However, they are contested by the world’s best jumpers, who are inherently less likely to fall. This demonstrates conclusively that a simple % Fell/UR ranking is not a perfect proxy for fence stiffness in isolation.

Section 4: A Comparative Analysis of Hurdle Course Difficulty

4.1 A Fundamentally Different Challenge: Hurdles vs. Fences

Providing a complete answer to the “National Hunt” query requires a separate analysis of hurdle races. The challenge presented by hurdles is fundamentally different and statistically safer than that of steeplechases. The physical and technical demands are lower, resulting in significantly lower casualty rates.

• As previously noted, academic studies show a stark difference, with a chase fall risk of 6.0% versus a hurdle risk of 2.1%.   
• A four-year study (2015-2018) encompassing 111,983 hurdle runners in the UK and Ireland found a combined Fell/UR rate of just 3.04%.   
• Fatality rates are also lower, with one study finding 5.9 fatalities per 1000 steeplechase starts compared to 4.5 per 1000 hurdle starts.   

Table 4.1: Comparative Risk Profile: Steeplechase vs. Hurdle Races Data synthesized from.   

Race Type	Faller Rate (per 100 starts)	Fatality Rate (per 1000 starts)
Steeplechase	6.0	5.2 – 5.9
Hurdle	2.1 – 3.04	2.4 – 4.5

This data confirms that the two codes are not comparable. The primary driver of this safety difference is the obstacles themselves.

4.2 The Technological Intervention: “One Fit” Padded Hurdles

A primary reason for the low and decreasing faller rate in hurdling is a deliberate, data-driven regulatory intervention. The British Horseracing Authority (BHA) has been progressively replacing traditional birch hurdles with the “One Fit” padded hurdle.   

• This new design, which features a custom-fitted foam pad, is “statistically proven to reduce faller and injury rates”.   
• A BHA-led trial at six UK racecourses (including Taunton, Kelso, and Uttoxeter) provided clear evidence of their efficacy. The faller rate over the padded hurdles during the trial was just 1.59%, a statistically significant drop of 0.5% from the 10-year average for traditional hurdles.   
• This success has led to the recommendation that all traditional birch hurdles be phased out.   

The implication of this technological standardization is critical: the “difficulty” of the obstacle itself is being engineered out of hurdle racing. The variance in obstacle difficulty between different hurdle courses is therefore much smaller than in chase courses and is shrinking every year.

4.3 Identifying Statistically “Difficult” Hurdle Courses

If the hurdles themselves are becoming standardized, what makes a hurdle course statistically difficult? The challenge reverts to factors of layout, topography, and stamina, which can lead to tired horses making mistakes at otherwise standard obstacles.

• Case Study 1: Taunton The course is noted for a “high casualty rate in terms of fallers and unseats”. For hurdlers, this is not attributed to the hurdles themselves but to the course layout, which combines with the lower quality of racing. The specific “danger” obstacle is “the second to last hurdle which [is] met on a downhill slope”. This downhill approach, often taken at speed by inexperienced novice horses, is the primary cause of Taunton’s high casualty rate for hurdlers.   
• Case Study 2: Kelso Kelso is known for a “low completion rate, especially on the hurdles course,” particularly when the ground is testing. This is not due to the hurdles but to the course’s demanding nature. Jockey testimony describes it as “a stiff course, for both disciplines” that “puts a premium on stamina”. It has a notably “stiff finish”. The difficulty at Kelso is a test of stamina; horses tire on the testing ground and in the long home straight, leading to jumping errors at the final obstacles.   

Section 5: Conclusion – A Synthesized Answer for the Expert Analyst

This report has moved from a quantitative ranking to a qualitative analysis to provide a definitive answer to the user’s query. The analysis concludes that there is no single “most difficult” course, but rather “categories of difficulty.” The statistical data is the starting point, revealing a complex interplay between obstacle design, track layout, ground conditions, race pace, and participant quality.

5.1 The Definitive Ranking

Based on the primary metric of % Fell/UR, the definitive statistical rankings are:

• Overall Most Difficult: The Aintree Grand National Course (21.13%) is a unique global anomaly, though this figure is expected to decrease following 2024 safety modifications.   
• Statistically Hardest (Non-Aintree): Leopardstown (IRE) (12.91%). It leads a dominant group of Irish tracks, including Punchestown, Listowel, and Sligo, which are statistically more hazardous than their UK counterparts.   
• Statistically Hardest (UK): Fakenham (9.46%), Plumpton (9.35%), and Ludlow (9.10%).   

5.2 Final Analytical Summary: The “Difficulty Paradox” Explained

The data demonstrates that “difficulty” must be defined by an analyst’s specific interest. A sophisticated stakeholder must differentiate between a stiff test of jumping and a statistically hazardous race.

• For the “Stiffest Test” (Largest Fences, Highest Class): An analyst looking for a true test of a horse’s jumping scope and ability should focus on Aintree (National Course), Punchestown, and Leopardstown. The statistical faller rates at Cheltenham and Sandown are suppressed by the elite quality of their competitors, but the fences themselves remain a formidable and “stiff” test.   
• For the “Trickiest Test” (Layout, Pace, Agility): An analyst looking for courses that generate a high number of casualties due to other factors should focus on Fakenham (for its acute, tight layout ), Ludlow (for its high-velocity races ), and Taunton (for its downhill obstacles ).   
• For Hurdle Races: Obstacle difficulty should be largely disregarded due to the standardization of padded hurdles. Analysis of hurdle races should focus entirely on course topography (e.g., Taunton’s downhill hurdle , Kelso’s stiff finish ) and the class of race (e.g., the high volume of inexperienced novices at certain tracks).