Gymnogenous Hazards

“You can only fight the way you practice.”
-Miyamoto Musashi, A Book of Five Rings

The context for which many historical fencers train, even if they know they will never end up in it, is lethal combat with sharp swords. Or indeed combat in, say, early 15th century Italy. We obviously cannot train in that context. The best we can do is approach that context from different directions, using different training methods. Overemphasizing any single method, or failing to adopt useful methods, introduces faults in our fencing. The relative merits of partaking in, as well as the potential hazards of overemphasizing, various methods like cutting practice and fencing competitions have been argued extensively.

Other than gloves, safety gear has received relatively little attention of this kind recently. This article will therefore look at a few of the problems we historical fencers run into when we neglect training with protective gear that would allow us to perform a full range of attacks with the sword, at proper speeds and force levels, at the correct distance. I approach this primarily from the viewpoint of unarmored longsword fencing, but it probably applies to most (unarmored) fencing, especially with larger weapons.

Sparring about Gloves

In several groups it is common to fence quite freely, with most of the body protected such that landing forceful strikes is not an issue, yet without decent protective gloves. This approach is often defended by arguing that a skillful fencer can easily protect his own hands, and that lacking protective gear on the hands in no way distorts the way people fence. Back in March 2012 a group of Polish fencers traveled to Barcelona, taking a few extra sparring gloves with them, to find out whether these arguments had merit. Videos of the sparring bouts that ensued resurface regularly, largely because of the great stylistic differences between the Polish and the Spanish fencers, both highly skilled at what they do. However, I’d like to draw attention to a particular feature of these videos.

You’ll have to imagine the thudding noise of steel on rubber.

Lingering in extended guards and parrying with extended arms, without projecting an immediate threat, leads to getting your fingers smashed a whole lot. This is typical of fencers who have little-to-no experience of fencing with proper protection for the hands and with strikes to the hands being allowed. Fixing this issue is fortunately straight-forward: Get proper gloves and fence with people who do target the hands.

The point is we have to practice in this manner, or else the way we fight is likely to leave us wide open to strikes to the hands. The same is largely true of protection for (and intentionally striking at) any particular body part. It just happens to be especially true of the hands; for reasons why see e.g. the recent article by Ilkka Hartikainen on targeting the hands, as well as “Attacking the Hands in Sparring”, by Keith Farrell.

Problems with Nudity: Faking It

The problems become much greater when training is primarily done with little or no protection. When protective equipment is severely limited, we can no longer practice attacking with proper speed and force and with intent to hit our opponents — equally, we can no longer practice defending ourselves from such attacks. A common feature of training in such a context is that large categories of techniques are completely avoided, particularly thrusts and any strikes that might hit the hands. The problems with this, as well as the solutions, are similar to those regarding lack of protective gloves.

There are at least 4 other major ways people modify their training because of lack of protective gear, although usually any one group or practitioner ends up adopting training methods that combine some or all of the four.

First of all, some people deal with the situation by intentionally missing their partners and pulling their blows, effectively turning their fencing practice into stage combat. There are two massive problems with this. First of all, the attacker is practicing to miss with their attacks, which is obviously a bad idea. Secondly, the defender cannot truly know whether they succeeded in their defense or they were simply saved by the attacker missing on purpose.

Safely planting the point 1 foot above their head

This may even form into a habit where we aren’t even quite aware of missing intentionally, allowing technique that wouldn’t work against a properly aimed attack to proliferate. The defender certainly doesn’t get any experience in dealing with properly aimed attacks, so that even if the technique is theoretically suitable, execution is likely to be lacking. In addition, since an attack that is going to miss should usually be met with different technique than one that is going to hit, intentionally missing teaches the defender to wrongly identify and respond to incoming attacks.

Pulling our blows might not be quite as bad is intentionally missing, but it causes similar problems. Going from a pulled blow to one that’s intended to hit is easier than doing so after training to miss. But the attacker now primarily practices attacking without any commitment. The crossings tend to be at least slightly off, the pressure different from what you’d get following an attack that wasn’t pulled. Thus we practice the wrong reactions to pressure and visual stimuli. And the defender is still often clueless as to whether the reason for why they weren’t hit is their successful application of the proper technique or the attacker pulling their blows.

The Slow Mo Guys

Others try to ensure safety when fencing with little protective gear by slowing down. Often this takes the form of consciously fencing slowly, which is then offset by separately practicing similar techniques and engagements at high speed. This may not be problematic, and can be a valuable addition in a balanced training program. Problems arise when training is overwhelmingly done in slow motion, and when no offsetting occurs. This usually goes hand-in-hand with coming up with excuses as to why it makes no sense to fence at high speed, as well as losing sight of what “high speed” actually is.

High speed footage of a high speed fencing action

There are plenty of reasons why this is bad. For more on the subject, see my earlier article Presteza e Forteza and the links therein.

Morse Code

How could we enable the attacker to strike fast and forcefully, at the proper target, while still assuring the safety of the unprotected defender? One answer is to help the defender cover themselves securely. A simple way to achieve this is telegraphing our attacks.

Making sure they see you’re coming

Again, when done consciously, and combined with plenty of practice in attacking without telegraphing, this may not be problematic. Some technical training even requires the defender to be prescient about incoming attacks, although in these cases you have to make sure the telegraphing is something relatively realistic. Counterattacks in particular generally only work if you know in advance when an attack is coming and which type of attack it is. For more on this, see “When to Counterattack” by Kristian Ruokonen.

Winding up and loading the feet to ensure the defender has plenty of time to safely perform their technique

Yet, like missing on purpose, telegraphing forms into an unconscious habit too easily. We cease to notice how we wind up, bounce, lean, turn our hips, adjust our grips, or otherwise give off a tell as to when and how we are going to attack. At that point we are practicing to make it easier for our opponents. This is rather counterproductive.

As defenders, telegraphing teaches us to keep looking for the same cues, which may no longer be there when faced with an opponent who is unconcerned with our safety. It allows us to believe we have a completely unrealistic amount of time in which to perform our own actions. When attacked by someone who doesn’t telegraph their actions, we’re suddenly blindsided, confused by where the attack came from and why our “safe” technique failed.


One last method of ensuring the safety of an unprotected training partner is to strike from far enough away that the attack cannot hit.

Too far is the perfect distance for safety

This is rarely done as a conscious choice in regular practice. Rather, the inability to actually hit our partner changes our motivations in preparing and executing our attack. If our partner were sufficiently protected, we’d be more likely to attack from the proper distance, since our failure to hit when striking from too far away would be a big red flag. But without protective gear we can’t hit our partner anyway, so we err on the side of caution. This often results in the distance between the fencers increasing. Additionally, since distance translates to time, by initiating our attacks from further away we help our partner to safely perform their technique, much like we would by moving slower or introducing telegraphing.


Watch the above closely, paying particular attention to the footwork of the two fencers throughout the sequence, and where the point of the attacker’s sword ends up in the final attack relative to where the defender’s head was at the beginning of the attack. This is typical of training situations where there is little or no intent to actually hit our  partners. As we’re performing the techniques we fool ourselves into thinking we aren’t actually too far away, by taking extra steps, leaning a bit further, or doing other similar corrections. This allows us to blissfully ignore the fact that we are practicing at distances where the techniques are unlikely to work, were our opponent intent on actually hitting us while avoiding being hit in turn.

Practicing like this combines the downsides of telegraphing, moving slowly, and missing intentionally. We’re practicing to give the opponent extra time and space in which to safely perform techniques against us, and even if they did nothing our attacks would often still fail. As a bonus, we’re extending our hands towards the opponent without seriously threatening them, leading to easy counterattacks to the hands. As defenders, this teaches us to perform unsuitable techniques against incoming attacks. On the whole it screws up our ability to read the tactical situation while fencing, because we’re expecting the wrong things to happen at any given range.

In Conclusion

It is theoretically possible to practice primarily without protective gear and not be exposed to any of the above issues. It is extremely difficult, however. The hazards are insidious. We are easily blinded by how much our practice is being distorted. Meanwhile training in proper protective equipment, at high speed, with intent to hit with our attacks, can be done easily and is the fastest way to remove the above distortions from our practice. There’s nothing quite like being hit hard and repeatedly to bring failures in our technique into sharp focus.


P.S. All of the gifs in this article were lifted, usually out of context, from YouTube clips. I am not implying that any fencers appearing in them they are practicing “wrong”. They just happened to demonstrate (possibly intentionally) some of the issues I brought up in the article.

The Exchange of Thrusts

“This play is named “The Exchange of Points”, and it is done like this: when your opponent thrusts at you, quickly advance your front foot off the line, and with the other foot step to the side, also moving off the line, crossing his sword with your hands low and with your point high into his face, or chest, as you see drawn here.”


-Fiore de’i Liberi, MS Ludvig XV 13, 26 verso. Translated by Colin Hatcher. From the Wiktenauer.

On the face of it, Fiore’s exchange of thrusts, ‘scambiar de punta’, seems fairly simple. Enough so that the majority of interpretations of it look very similar. Here are two representative examples, by Guy Windsor (top) and by La Sala delle Armi (below).

The directions of the steps vary between interpretations, which is understandable since Fiore does not say which side we should be stepping towards. More interesting to me, however, is something that all interpretations I’ve seen appear to agree on: that the technique consists of two distinct movements. First a strong crossing during the step of the front foot, then a thrust during the passing step. I’m sure most people teaching this technique would say that, optimally, the two movements flow into each other, minimizing the time elapsed between the parry and the riposte. Nevertheless, tactically that remains a parry-riposte rather than a counterattack, with the riposte landing after the attacker’s thrust has already been completed.

The case for the exchange of thrusts as a counterattack

The description of the exchange of thrusts in the Getty seems to imply that the crossing of the swords occurs during the passing step and that the student’s thrust lands at the same time as the crossing is made. The drawings of the play suggest the passing step is, regardless of which side it is diagonally going towards, closing the distance to the attacker (the player). This appears to support the interpretation of the exchange of thrusts as a counterattack.

Unfortunately the description is by no means explicit in this, and one might argue that the sentence structure was not meant to imply any such thing about the timing of the technique. Since I don’t understand the original Italian, I could hardly argue against. The same footwork appears in 3 other techniques for the unarmored sword in two hands in the Getty: the yielding parry in the play of the colpo di villano, and the beating parries in the play of the breaking of the thrust and of the master defending from dente di zenghiaro. All 3 only include a parry during the off-line passing step, with a separate offensive action to follow after, and thus don’t really shed any light on the exchange of thrusts.

Luckily we also have this from the short description of the exchange of thrusts from the Pisani Dossi (translated by Michael Chidester): “In the art, a more deceptive thrust than this cannot be made.” Would Fiore have described parrying a thrust followed by your own thrust as a riposte as the most deceptive thrust one can perform? The attacker can reasonably expect that their opponent will be attempting some manner of offensive maneuver after the thrust has been completed, especially following a parry. But a counterattack, landing the exchange of thrusts while the attacker is still preoccupied with their own thrust, can easily take the attacker by surprise.

There’s also the issue of keeping the hands low, as per the text and the images. When setting aside a fully developed thrust and attempting to riposte from the resulting crossing, it can be very difficult to retain sufficient control in the crossing with one’s hands that low. What tends to happen is that the hands creep up in an attempt to wind the forte of one’s blade onto the foible of the opponent. This is less of an issue in a counterattack, because the crossing occurs before the opponent’s structure is fully behind the thrust, and the opponent doesn’t have time to adjust to the crossing or to wind against you. When counterattacking into an expected low thrust it makes sense to keep the hands low, in order to securely close the line against the incoming attack.

A counterattack has many other practical advantages. The attacker has a drastically smaller time window in which to deal with a counterattack than a parry-riposte, no matter how smooth and fluid the latter is. The counterattack also restricts the types of counters available to the attacker. A feint followed by a thrust from the other side, a disengage underneath the defender’s sword from the crossing followed by a thrust from the other side, and a parry & counter-riposte (or even counter-attack into the riposte) after the completion of the original thrust are all effective techniques against an exchange of thrusts performed as a parry-riposte — and all would be suicidal against one performed as a counterattack.

This is not to say that a parry-riposte exchange of thrusts is “wrong”. To be able to perform a successful counterattack, you require some manner of time advantage on the opponent. If you are forced to react to a sudden, fast attack at a proper measure, then a parry-riposte will generally be the best you can manage. The parry-riposte combined with voiding is also a safer bet when you are unsure how your opponent is going to attack. But, considering the advantages of a counterattack in those cases where it is a possibility, it is a bit odd that nearly all online material on the exchange of thrusts describes it as a parry-riposte.

The best demonstration I could find of what I consider perfect timing for an exchange of thrusts, even though the footwork lacks some Fiorean flair, comes from Axel Pettersson at Swordfish 2015. Note the step forwards into the developing attack, the direct path of the point to the target, and the low hands.

Presteza e Forteza

At what speeds, what levels of force, what kind of intent should we train historical swordsmanship? These questions have caused frequent, fervent debates in the HEMAverse. For a quick intro into the larger discussion, see:

My own view, after 2½ years of training at one extreme of the spectrum and 2 years of slowly coming into grips with the other end, is roughly this. The sources tell us to be fast and strong, and to use those attributes to our advantage when fencing. The advantages are quite obvious in practice. To fence with speed and strength, we must train with speed and strength. But this is not nearly always reasonable, such as when learning an unfamiliar technique or an unfamiliar application of a technique, or when there are safety concerns.

One can form a difficulty scale for the training of any particular fencing skill. At the easiest end, the skill is practiced in isolation, as a response to a known and obvious stimulus, from a static position, close range, slowly, and weakly. Difficulty increases when there is choice involved, the stimulus is less obvious and even unforeseen, movement is included, there are preparatory or follow-up actions, ranges increase and vary, speed is increased, and more force is applied. When a beginner is first introduced to a beating parry from Dente di Zenghiaro, they should start at the very easiest end of the spectrum. Then difficulty should be increased as the skill in question improves, such that training remains challenging but not impossible. Difficulty should be varied, using different modifiers, so that training is never monotone. If technique breaks down, you can always isolate the problem and/or reduce the difficulty.

I suspect most practitioners of historical fencing, especially those involved in coaching, agree with the above in theory. However, many have extremely strict definitions for “speed” and “force”. Many also believe that difficulty should be increased only very, very slowly as skill improves, to preserve idealized “perfect” technique. This causes problems.

Another argument for training with speed, force and intent

As Vincent Le Chevalier points out in Speed Matters, when we perform a fencing action slowly, we are not in fact performing the same action as we would at high speed. Simple physics dictate that there will be differences.

Moreover, if you closely watch the actions of people demonstrating an action slowly and compare that to how they perform those actions at full speed, the differences tend to be much greater than just those simple physics related ones. I’ve watched dozens of technique demonstrations, live and filmed, where people strike fendente mandritto/right Oberhau slowly and lightly, as well as at speed and forcefully. The two attacks look very different, and the difference is greater the less the demonstrators are used to practicing at high speed and intent.

There are a number of reasons why this matters. But I want to focus on a specific one. As Kristian Ruokonen wrote in Characteristics of fencing as an activity & how to train:

Longsword fencing is an open motor skill. You have to take into account all the external stimuli. Small changes in your opponent means you need to perform different actions.

For each fencing action, there is a time, a distance, an action by the opponent that calls for that action. I can, in a controlled drill, perform a number of different techniques against an incoming strike. But to actually pull them off “for real”, I should be able to distinguish between an incoming strike that calls for a counter-attack to the arm without opposition, a strike that calls for a counter-thrust with opposition, a strike I can void and chase after, a strike I should simply back away from, a strike I should parry and riposte in this line or that… If I fail to identify the situation correctly, it doesn’t matter how “perfect” my technique otherwise is, the outcome is likely to be bad for me.

Now, since slow, weak attacks are substantially different actions than fast, forceful attacks, reacting to them as though they were fast and forceful is teaching you to react wrong. You are practicing to parry when you should be voiding, counter-attacking when you should be parrying, etc.

This is in addition to the fact that at lower speeds you do not need to react as quickly as at higher speeds. This is often mentioned as a positive of slow practice by its proponents — and it may be when specifically trying to widen your repertoire or working on your analytical thinking skills. But it’s no good if you’re trying to improve time-sensitive motor responses.

And to top it all off, the actions you perform as a reaction to the (wrong) stimulus are often such that they would not be successful outside of slow-motion fencing.

So when you are training, say,  a parry-riposte against an unrealistic slow and weak strike, you are learning to:

  1. Take your time before deciding on whether and how to deal with the strike.
  2. Choose the wrong action against that strike.
  3. Perform the parry such that it would probably fail against a fast and powerful strike.

Now I repeat that such training methods are generally necessary in the beginning. But since the distortions caused by these methods are so great, it is a good idea to ramp up intent, speed, and force in a fairly early stage of learning new skills. Enough at least that new students get a personal feel for the extent of the distortion. Experienced practitioners should spend a lot of their training time working at speed and force levels where success requires fast reactions, right choices, and properly executed actions, in order to make sure they don’t get sloppy. To wait until a student “perfects” a wide repertoire of skills in simplistic, unrealistic settings before beginning to teach them how to apply any of those skills… Is counter-productive to say the least.

I have had to deal with the outcomes of such distorted training quite a bit over the last 2 years. Primarily in my own fencing, which was and is very painful. Yet I remain extremely grateful to the people who made me face these issues. I only hope the proliferation of improved training methods for historical fencing means fewer and fewer practitioners have to face the same in the future.

(Note: I primarily practice Fiore longsword. I suspect the above is relevant to most historical swordsmanship traditions, but I make no real claims regarding, say, smallsword.)