All of our safety research is community funded and we need your help to make it happen. If you find this report helpful, please consider supporting us through Patreon.
Introduction
About a month ago, I put out a call to subscribers, Kickstarter backers and through my instructor Facebook page asking for people who has been injured, or had their equipment fail, by being punctured from a sword thrust. This wasn’t simply morbid curiosity: the resulting report is a key part in understanding the results of the Safety Tips Project. Safety Tips Phase 1 is nearly completed, with the remaining work largely consisting of publishing our background research to show how the experimental results relate to real-world safety. One of the remaining reports concerns the actual injuries that happen with the risks we’ve tested.
We hit a roadblock with that report when it came to describing puncture injuries – they just aren’t that common. Puncture injuries are rare enough that in the three relatively comprehensive sets of injury reports we have, they don’t appear at all. Collectively these cover 5 major events in 2024 (Three Wessex League 2024 events, Albion Cup 2024 and Bohema First Harvest) and Sean Wauters survey of 268 Dutch and Belgian HEMA in 2016. By contrast, the other risk we studied (Mask Injury risk, including mechanical injury/bruising and concussions) appears in all 3 sources.
The relative rarity of puncture injuries is great from a safety perspective – these are often serious injuries requiring medical attention, and all four deaths in Olympic Fencing are from a puncture injury with a broken blade. But the lack of data made discussing these injuries challenging, leading us to compile this report: 10 case studies in Puncture injury.
Case Studies are useful because they show what is possible to happen but they should be approached with caution because it is much harder to draw conclusions about what is likely to happen. This is particularly true when considering specific equipment, which may feature prominently purely because it is very widely used in HEMA rather than because it has any specific increase in risk. Nevertheless they provide a useful insight into the risks, how these injuries occur, and how they might be prevented. Several common themes occur in these examples that provide valuable lessons for HEMA practitioners and equipment manufactures to consider.
The case studies have been written with information provided by the injured party, and their consent to publish. All case studies are anonymised as far as possible. If readers are familiar with the case and happen to know the name of the fencers involved, we ask that they do not make public statements about it except in response to the fencer having identified themself.
If you just want to read the case studies and draw your own conclusions the report below contains only a factual description and opinions as given to me by the injured parties – with occasional context when required for equipment standards and construction. Be warned: the report contains images of the injuries, some of which are distinctly unpleasant to look at. What follows in this post is my summary and analysis of what we can learn from these injuries.
I would like to say a huge (anonymous) thank you to all the fencers who contributed their stories to help make this happen. The information they provided will hopefully help others learn from their experience and adapt their own practice to improve safety outcomes.
Key findings and recommendations:
The most shocking thing to me about these case studies is just how normal most of them are. I was expecting to see more unusual situations (e.g. unsafe behaviour, broken swords, small tips, strange custom equipment) and while some of these things do appear, for the most part the case studies describe perfectly normal HEMA equipment used in ways that would be unremarkable if not for the resulting injury or equipment damage.
- Untipped swords have the potential to cause serious injury if they strike unprotected or lightly protected locations. This includes larger tips that have been rolled or flared at ~80-100mm2 (at least up to a Regenyei Standard Rolled Tip)
- Tipped swords that still have a relatively small total surface area also have the potential to cause injury
- Sufficiently narrow or broken swords can puncture 350N jackets, although the jacket resistance may still reduce the severity of injury even when it fails
- Non-puncture resistant gloves do not provide sufficient protection to prevent puncture injuries to the hand, including many leather gloves in common use
- Seam strength is also a potential failure point even if rated cloth is used but is rarely tested for in HEMA equipment
- Thrusts can strike to even extremely narrow gaps in protection. Any gap in protection poses a potential risk.
- Unsafe behaviour and particular fencing actions can increase both the likelihood and potential severity of injuries
- More severe cases tend to have several factors involved. Any risk management process must have consider controlling each factor individually and not consider any one control in isolation
These findings agree with the experimental results of the Safety Tips Project – both untipped (one rapier and four longswords) and bullet cases punctured our target at relatively low energy levels, whereas very large tips would not puncture the same target. They also align to similar cases of puncture risk that have been heavily discussed recently, including two discussed by Martin Fabian (a mask being punctured by a longsword without an expanded point, and a 350N jacket being punctured by a rolled tip longsword) and the recent incident at Black Horns cup where a rapier punctured a 350N jacket and penetrated the bicep of the fencer. This last incident will be the subject of a future, more detailed report which will be produced when we hit 20 paid members on Patreon.
Recommendations
- Organisers and practitioners should review their sword tipping practices. It is likely that all but one of these injuries would not have resulted in a puncture with a well fitting, sufficiently large tip. Tipping swords is by far the cheapest and simplest way of reducing these risks. However it should be noted that in many cases while a tip would prevent a puncture, other injuries may still have occurred from the percussive force involved.
- Safety standards are only valuable if they are enforced and regularly checked. Practitioners should regular inspect both their and opponent’s equipment, and organisers should consider how to ensure their safety rules are followed (e.g. if tips are required).
- Glove manufacturers should review their use of cloth components and seek to use puncture resistant cloth and suitably secure seams
- Practitioners should review the gloves they use and consider whether an underglove is required and/or whether to use a puncture resistant model
- Suitable undergloves or light puncture resistant fencing gloves are notably lacking. Olympic Fencing gloves may be the only suitable choice (note: it is not clear that coaching gloves have the same rigorous testing/construction as competition gloves)
Common Themes
The common themes below draw together patterns that occur across several case studies. The themes are grouped into three different categories of safety risk and control: protective equipment, swords (including tips) and behaviour.
In each case, I list the case study number which corresponds to the order in the document linked above. Note: the case studies are listed in order of severity, so higher numbers cited represent the more severe injury.
Protective Equipment
Failure of Protective Layer
Almost all of the cases involved direct failure of a protective layer. This happened in all but 1 of the case studies – with 7 of these cases being a glove failure.
Most of these were fabric failures where the cloth was directly punctured (1,4,5,7,8,9,10). In the only case that didn’t involve direct equipment failure (2), the thrust still punctured a cloth layer (the sock) of the practitioner, showing that normal cloth does very little to mitigate this risk. Apart from two (1,8), the fabric was unrated. Of the two with 350N rated equipment, one was partial (1) and the other was with a broken sword (8), discussed in more detail below.
With some exceptions, the protection failures happened with equipment that is perfectly normal and common by HEMA standards; none of the protective equipment would have been out of place at any event I’ve visited. While some also happened with unusually narrow swords, many others did not (see below).
Rated Fabrics
Four (1,2,6,8) of the case studies involve direct failure of “rated” equipment, but all in quite unusual ways. There are only two cases with a direct failure of rated cloth – both 350N jackets (1,8) and two where the cloth remained intact but the sword broke a seam (2,6) (see next section).
In Case 1, while there was partial penetration of a 350N jacket it is not clear if the outer layer is the rated layer on this jacket – it’s possible that the “rated” portion was not punctured but this can’t be determined.
Case 8 was also against a 350N jacket, and total penetration occurred. However, this is the only example included of a puncture injury caused by a broken blade: clearly the blade was sharper and stiffer than “normal” HEMA blades because of this failure. The strike was also close to a seam which may have reduced the fabric strength. In any case this clearly shows that a broken blade can penetrate a 350N jacket in the wrong circumstances.
Burst Seams
Two of the rated equipment failures involved burst seams. Case 2 was a seam in a Leon Paul X-Change 1600N fencing mask bib, which is governed by the EN13567 standard for Fencing Equipment. While this standard does contain a test for seam strength, it does not apply to the mask bib. Outside of the Leon Paul X-Change system, masks do not typically have exposed seams in this way – this may not be a risk with other styles of mask. So while the mask fabric would have been rated to 1600N on the puncture tests, the actual point of failure may have been completely untested.
The other seam failure was in a pair of undergloves, which were rated as EN-388 2X43DP. EN-388 is a work glove standard, and has various tests and thresholds which are described in the 6 digit code. These gloves were rated at Level 3 for puncture resistance – the 2nd highest level, or over 100N. Note that the test in En388 is different to the EN13567 measure for puncture resistance and the numbers look comparable but aren’t. The handful of protective items I have found that measure with both tests suggest that the same equipment will have between 2-10x as high a N rating in the EN13567 test as the EN388 test – it is likely these gloves exceeded 350N in rating by the fencing standard.
However much like the above, seam strength is not tested for in this standard at all. Additionally, the test area for the standard is the palm – the back of the hand (where the injury occurred) is not tested and may have no puncture resistance at all, although in this case it appears to be the same material as the palm but there is no way to determine this for the gloves in question. Note: olympic fencing gloves (EN13567) do have burst seam tests and are rated in the location in question.
Gaps in Protection
Several of the case studies involved the sword thrust hitting a gap in protection. None of these gaps were especially large, with the smallest being ~2cm across, and they all also had other layers of clothing or protection underneath (that failed). Given the nature of the different types of strike, it seems likely that thrusts are more prone to bypassing protection in this way than cuts would be due to the smaller striking surface area.
Swords, Points & Tips
There is an extremely wide variation in the types of sword used and the point or safety tip involved in the incidents, with different weapon types involved (Messer x 1, Sidesword x 1, Rapier x 2, Longsword x 6). Most injuries were caused by an untipped sword (1,2,3,4,5,6,7,10), although in two of these cases (1,4) the sword would normally have been tipped but the tip had fallen off.
Untipped Swords
The studies clearly show that untipped swords have the potential to cause injury or damage. Most of the injuries involved swords that the practitioners considered to have an unsafe sword point when untipped (1,3,4,5,7,9). However, several swords had tips at the larger end of normal and which most practitioners would consider safe in an untipped state (2,4,6,10).
Smaller Safety Tips
The tipped example was a rapier with the point wrapped in tape to expand the tip. The resulting tip was still relatively small at 69mm2 – this is smaller than a 9mm bullet case (~75mm2), or the standard tip on most longswords (70-110mm2 for standard Sigi and Reyenyei blades) but still significantly larger than the untipped rapier at 17.2mm2, and within the lower end of the normal range for many weapons (e.g. sideswords, sabres). Note that while this is a relatively small tip and injury occurred, many fencers will routinely use a sword with a point this size. The injury was severe but also in quite unusual circumstances that prevented the blade from flexing.
Tips falling off or not used
Two cases involved swords that were usually used with a rubber tip, but the tip had fallen off earlier and been unnoticed (1,3). Both fencers recognised the point as being unsafe without an additional tip, but were not aware that the safety tip was removed and did not stop or adapt fencing accordingly. Another case involved a sword was recognised as requiring a tip and the owner had been told to add a tip by the club instructor, but they had not followed the instructions and fenced with it anyway (5). A final case (10) was at an event where tips were required, but the sword in question had not been tipped. In this case the wielder had borrowed the sword and neither fencer had checked if it complied with safety standards.
Broken blade
Only one case (8) involved a broken blade. The blade broke during a thrust and the fencers momentum continued to carry them forward leading to the broken sword thrusting their opponent. The break was not particular sharp, and fairly typical for most longsword breaks. This is also the only instance of 350N rated cloth being punctured and injury resulting. While the circumstances may be rare, it shows that broken blades do pose a risk that is not fully mitigated by 350N fabric.
Behaviour & Actions
Behaviour in this context includes both violations (intentional or otherwise) of agreed safety standards and specific fencing actions by one or both fencers. This includes cases where both fencers were moving but the individual actions were not themselves “unsafe”. Behaviour was a contributing cause in seven cases (1,5,6,7,8,9,10) – and note that the six most severe cases all have one or more behaviour factors present. This suggests that these issues may increase the possible severity of puncture risk, as well as the likelihood.
Unsafe Fencing Actions
Several of the cases involved fencing actions by the opponent that either the injured party or at least some of the HEMA community would consider unsafe (7,8,10). The nature of these actions is quite different. In 7, the opponent deliberately targeted a lightly protected area (hands with light leather gloves) with a messer. 8 and 10 both relate to the force of the thrust. In 8, the opponent made a fleche attack as part of the thrust and was unable to arrest their momentum when their blade broke. Both the force of the initial thrust and the inability to stop it would have contributed to the injury. In 10, the opponent was described as being “frustrated” and making a particularly strong thrust towards the hands. This is the most severe injury on our record and the sword in question had a relatively large tip – it is almost certain that the increased force was a major factor in the severity of this injury.
Ignoring Safety Instructions & Issues
Several cases involved the opponent either not complying with the safety instructions and policies that had been set (5,7, 10) or failing to deal with known issues with their equipment (1). In 1, the opponent’s sword tip had fallen off multiple times in previous sessions and the tip falling off was a known risk that should have been addressed. In 5 and 10, the opponent was using a sword that should have been tipped according to the local conventions but was not (deliberately in the case of 5, unknowingly for 10). 7 was the most extreme case of defying safety conventions; the agreed conventions had been for light sparring with no targeting of the hands, which was why the fencer felt confident in using light gloves for messer fencing. Their opponent instead deliberately targeted then hands, and with more force than agreed. This is apparent from the wider exchange where they first cut the hand (breaking a finger) then thrust at it (causing puncture injury) in rapid succession. Worse, they had been told to target the hands by their instructor despite the agreed conventions. While ignoring conventions was a contributing factor in all four of these cases, it seems to be the primary factor in this one.
Simultaneous Movement & Force
Both fencers moving at the same time may have increased the injury risk in four case. This includes two cases where both fencers thrust simultaneously (6, 9) and two where the attempt to defend against the thrust was unsuccessful and the thrust struck the hand mid-defence (5,10). In all four cases, the simultaneous action would have changed where the strike landed and led to a relatively unprotected target being struck. It is likely that this may also have increase the force levels involved and/or prevented the opponent’s sword from bending in it’s usual manner.
Future Work & Research
If you’ve reached this point, I hope you have valued the insights this research gives into some of the causes of puncture injury. For my own part, I found producing this to be a strange mix of feelings: horror mixed with fascination, and both being glad to finally have some data and the realisation of how much which we simply don’t know.
This report fits within a wider set of reports that will look more broadly at the risks studied in Safety Tips, but also suggests some new projects to perhaps take forward. One area we are looking to pursue is improving reporting on injuries by providing organisers with a place to submit anonymised injury statistics, and practitioners to submit case studies from their own experience. Increasing the available data on injuries is an important first step in improving our safety practices. If you’d like to be involved in either of those goals, please get in touch by emailing us here.
All future projects are dependent on me carving out the time to do them – this will be much easier and go much faster if I have enough funding to dedicate time to these projects instead of having to squeeze it in the few hours between everything else. If you want more reports like this, please considering supporting it through joining our Patreon – or if you can’t support us financially, by sharing our work and discussing it within your own local area to make sure your decisions about risks are as well informed as possible.
Leave a Reply