The Principles of Pair Training (bonus chapter from Solo Training)

Welcome! If you’ve come here directly from the  pages of The Principles and Practices of Solo Training, then you should be pretty well oriented already. But if you’re new here, this is the resources page from the Footwork chapter of my book The Windsor Method: The Principles of Solo Training.

The Principles of Pair Training

Every martial art I have ever come across uses set drills of some kind, even if only at the beginning (and in some cases, set drills were the whole art). It is relatively easy to teach set drills to a student or class. She does this, you do that. And it is relatively easy to set up a freeplay (sparring, fencing) environment that is reasonably safe. I have seen many groups and schools that have nice set drills, and freeplay quite a bit, but there is no real relationship between the two kinds of training. The things done in freeplay bear scant resemblance to the actions in the drills. The point of this bonus chapter is to show you how to build a bridge between set drills and freeplay. This is especially important for historical swordsmanship, as the manuals tend to show short, simple sequences (an attack and a defence, usually) which are easy to turn into drills, but very hard to pull off in friendly freeplay.

Freeplay, sparring, fencing, whatever you want to call it, is a key component of most martial arts. How you should prepare for it depends on what your freeplay is supposed to accomplish, and where freeplay-like activities fit into your system.

Schools that emphasise winning tournaments tend to introduce it very early, using heavily modified weapon simulators and lots of protection (my earliest formal weapons training was sport fencing; I seem to remember that we did freeplay at the end of the first class) ; schools that emphasise other things such as historical accuracy, or self-defence, tend to modify the weapons less, use less protection, and so naturally introduce it much later.

The principle difficulty in using pair drills effectively is bridging the gap between being able to act out the choreography, and actually developing skills that will work in less co-operative environments. As we saw in the previous chapters, the way to improve is to deliberately practice, which means finding the weakest link, and fixing it. 

The pattern goes like this:

Run a diagnostic: train at the level where something goes wrong, then figure out what that something is. Is it technical, or tactical? Tactical problems are all about knowing the right thing to do, and technical problems are all about doing the right thing well enough for it to work. Most of the more complex training set-ups can go wrong at either point, so it’s a good idea to have a clear idea of what kind of problem you are working on, and you should be able to identify the type of problem before you rush off to fix it.

Let’s say you got hit when you were not supposed to. Here’s what must have happened:

  1. The actions were familiar, but too hard or fast
  2. You came across an unfamiliar action
  3. You have no idea what just happened.

The answer in every case is to repeat what happened, but slowly. This is a key skill, perhaps the key skill, in martial arts training; at least one partner must know what just happened. If you can’t repeat it, you can’t train it, correct it, recognise it next time, or defeat it. We call this “fencing memory”, and we have a simple drill for training this skill. You will need it for every stage of building the bridge.

Developing Fencing Memory:

This drill works best with three fencers, an attacker, a defender and an observer. Switch roles after each phrase, to develop your ability to remember phrases you have both done and seen.

  1. Designate an attacker and a defender. 
  2. Allow free choice of attack and defence, but no continuations (attacker can’t counter). 
  3. Attacker attacks as he likes, defender tries to defend. Notice who gets hit. 
  4. First one, then the other, then the observer, describes in clear fencing language, in detail exactly what occurred.

For example (using Fiore longsword jargon, sorry you non-longsword readers!): “Mary was in coda longa, I was in posta di donna. Mary attacked with a thrust to my face. I tried to exchange the thrust, but my sword caught on the back of my mask and I missed my parry. Mary’s thrust landed in my face”. Then Mary describes what she thought happened “well, I started in tutta porta di ferro, and attacked with a mandritto fendente… (you’ll be amazed how rarely you’ll agree with each other to start with). Lastly, the observer states what they thought happened. If the observer doesn’t have a reliable fencing memory (and even if they do!), use a video camera too.

When one attack and one defence can be reliably described and repeated, add the attacker’s counter. When that is easily recalled, then the defender can counter that, and so on. Once you have built it up so that you can accurately reproduce a phrase of at least six actions (three from each side), your memory is ready for useful freeplay.

When you are reconstructing what happened, it’s usually best to start with the blow that landed, and work backwards from there. 

How drills are structured:

Every drill should have a clearly defined tactical context, and a clearly defined technical solution. They usually look something like this:

  1. Attacker and Defender are ready to go. 
  2. Attacker attacks with action A
  3. Defender defends with action B
  4. Attacker continues with action C
  5. Defender counters C with action D

and so on. Let’s call this drill ABCD. Action A is the tactical context for defence B. So let’s begin with a kind of drill that establishes where your technical gaps are. 

Work the combinations:

Set up the drill ABCD. Then, from the same starting point, whatever that is in your system, the Attacker varies their attack; actions F, J, N, and so on. Be that “thrust from the left”, “beat attack and lunge in quarte”, “go for the hip throw”, or whatever. In each case, the Defender should make the proper defence. 

The next stage is for the Attacker to use whichever attack they have chosen to set up their counter; A draws B, so he can do C.

Then of course, the Attacker does A-C so the Defender can practice D. 

In most systems, attack A has more than one defence; so once you have worked through all the main attacks, go back to one of them, and work through all the main defences against it. (My longsword students know this as the Four Corners Drill.)

Then do the same for all the attacks.

Then do the same for the Attacker’s counter to each of the defences generated by each of their attacks. 

Then do the same for the Defender’s counter to each of those…

This generates a combinatorial explosion. Very quickly, you could end up with a ridiculously complicated set of useless drills. So your goal is to survey the system quickly, establish where you are weakest, and set about chunking the actions together. You should not have to remember more than three or four basic defence ideas (eg block, counterattack, avoid) or three or four basic attacking ideas (eg strike, go around, feint). As you can see, I only know three of each. But each one can be usefully expanded on for hours. 

Please note, we are still in the basic form of the drill. 

This is where many schools get stuck; they can get this far easily enough, and it generates all sorts of cool new sequences for students to remember, but doesn’t really help them develop workable skills. So let’s move on. I have a set of “multipliers” that add complexity in a systematic and approachable way.

  • Add a step
  • Who moves first?
  • Degrees of freedom
  • The Rule of C’s.
  • Flow drills
  • Slow fencing
  • Pressure drills

Add a Step

A simple means to make a drill more useful is to allow the “loser” to counter the last step if they can. So for instance, you set up the drill, and, as the attacker counters, the defender may, if they see it coming and can think up something useful to do, counter the attacker’s action. There is in theory no end to this drill, as every action can be countered. Add one step at a time, and stop when it becomes difficult to remember how you got to where you are. This uses a set drill to set up a kind of slow freeplay.

Who Moves First?

In any set drill, you can 

  1. start the drill with both players standing still in guard (this is the usual set-up for beginners to start with)
  2. change who attacks
  3. draw the attack by some prior movement or invitation. 

In any set drill, you can add complexity by changing who moves first. In the basic set-up, you moved first. But your partner could instead be coming towards you. In my salle we spend a lot of time doing basic drills but starting from way out of measure. The trick is to arrive at the right time, in the right place, to do the initial actions of a particular drill, without exposing yourself. This is very hard, at the beginning.

Play with this idea; begin co-operatively, so starting from way out of measure see if you can both come smoothly forwards, and make the drill look perfect. Once you can do that, see what happens if you have opposed intentions. xxx

This will quickly devolve into a mess, but so long as it’s a safe and useful mess, there’s no harm done. Just start again…

Degrees of Freedom

This is the beginning of the bridge between set basic drills and advanced fencing skill. At any stage in any drill, a set of decisions have been made. Systematically allowing a different choice to be made by one player, on the fly, introduces an element of unpredictability for the other player. For example, we might allow the attacker to choose their counter to the defence at random. This can be either to develop the attacker’s decision-making skills (so the defender is helping them), or to develop the defender’s ability to adapt (so the attacker is helping them). When there is a choice like that to be made, we say there is a degree of freedom—the attacker in this case has one degree of freedom—one point in the drill where they get to make a choice. The other has to respond appropriately in real time. By adding degrees of freedom one at a time, we can get all the way from ABCD to freeplay. Common places to add a degree of freedom are:

  • The defender doesn’t have to wait for the attack, but can pre-emptively attack.
  • The attacker can vary the type of attack.
  • The defender can vary the type of defence (the most common change is from parry-riposte (two motions, one to defend, one to strike) to counterattack (defence and strike in one motion).
  • The attacker can vary their reaction to the defence (e.g. feint, or parry the riposte, or enter on the parry, etc.)

Adding a degree of freedom immediately changes a drill from choreographical to complex. At this stage, we must also cover the ‘Rule of C’s’.

The “Rule of Cs”

The “Rule of Cs” defines the three ways to train any drill.

1. Co-operating in creating correct choreography.

2. Once the choreography is smooth, increase the difficulty by increasing intensity, or introducing a degree of freedom (e.g. is the attack a cut or thrust?), with one player adjusting the difficulty for the other to learn at their most efficient rate—if it works all the time, ramp it up—if it fails more than twice in ten reps, ease off a bit.

This is called: Coaching correct actions.

3. Finally, the players each try within reason to make the drill work for them. In First Drill for instance, the degree of freedom that has just been coached might be that the attacker is either continuing with a pommel strike, or angulating around the parry. When coaching, they try to make sure the defender can usually counter them; when competing, they just try to make their action work. This can be dangerous if it gets out of hand, so be careful, and wear full protection just in case. In practice, the more experienced fencer should get most of the hits, without departing from the drill. This is fine, and gives a good indication of whether your training regime is working. So,


There is a significant risk of this getting out of hand; be mindful as you play the drill competitively that you must stick to the constraints of the drill that you have both agreed on. Otherwise you lose track of the rationale behind what you are doing, and mistakes creep in that are difficult to spot and to trace back to their source.

Coaching as a practice

Coaching is hard. It is a very high level skill, so you had better start learning it now. Perhaps the easiest place to start is the Buckler game, from the Striking chapter.

When playing the game, notice how easy it is to distinguish between success and failure. For the student, success equals hitting the buckler. That’s it. For the coach, success equals the student getting visibly better at hitting the buckler. It’s that simple. Complexity is added by playing with timing and measure, and the student rises to the challenge.

The student will only get better if the coach challenges her. You can tell whether the student is being properly challenged by their rate of failure. If they are hitting the buckler every time, they are not being challenged. If they are missing all the time, they are too challenged. The optimal rate of failure is about 2 out of 10. Too many failures, slow down. Too few, make it harder.

Now run the buckler game again, and be absolutely focussed on the rate of failure. 

So what is the buckler game good for? When would you prescribe it to a student or yourself? 

I’d say for the following ailments: poor point control, poor timing, poor footwork, and poor judgement of measure.

Obviously, it is completely useless for teaching solid parries.

This is true for every drill: it will be good at some things, and useless at others. 

So now let’s run the game again, modifying it for those three things.

1. Version One: point control. Mark the centre of the buckler, or use a smaller buckler, or make the buckler move a bit more. The skill you are working on is the ability to put your point exactly where you want it. Not timing, not measure. How can you make it more difficult? Make the target smaller, or reduce the available time.

2. Version Two: timing. Be really careful about how long the buckler is exposed for, and the rhythm of the exposures. This is not about reactivity, exactly. It’s more about teaching the student to move as the target is being presented, not when it’s already there.

3. Version Three: footwork. The goal of footwork is to get you to the right place at the right time with the right structure. Very often, having taken a step or two, you will have your weight in the wrong place for lunging (or for doing whichever other striking action you’re working on). So make the student move, and try to expose the buckler when they are off-balance (tip: they should never be off-balance if their footwork is any good!). If their weight is on the front foot, they can pass; if on the back foot, they can lunge…

4. Version Four: measure. You can expose the buckler when the student is out of measure and they should ignore it. Too close, and they’ve let you get too close. Play with challenging them to recognise the measures in which they can strike.

Versions two, three, and four are obviously related, but they are distinct. 

The coach is providing a feedback mechanism. If the student is performing the desired action at the desired level, then they will be reinforced by immediate success; if they do anything else, or do the desired action at an insufficient level, the student fails to hit, and/or gets hit. 

This is why a good coach can get preternaturally fast results, because they can create and control a perfect learning environment, in real time.

In a perfect world, every martial artist would have access to a high-level coach and spend much of their time one-to-one with her. In the real world, that’s never going to happen, so you and your partner must learn the basics of coaching so you can help each other develop.

As with every other skill, you will get better with specific practice.

From now on in almost all pair drill practice, you or your partner should be coaching. Once the choreography is there, coach. When you compete and someone gets hit, coach them past whatever weakness got them hit.  

My course How to Teach, goes into depth and detail about coaching.

Flow drills

A flow drill is any drill which has no natural end- it just keeps going, round and round. My favourite is probably my dagger disarm flowdrill, but I have them for every weapon I teach: they are that useful. This video from my Medieval Dagger course will give you the idea:

The Dagger Disarm Flowdrill

So how do we use flowdrills?

Set up the Flow

To start with, set up the flowdrill in a friendly and co-operative way.

Break the Flow

Once the flow is working nicely, you can break it with any agreed action. In the dagger flowdrill, that could be an arm break, a takedown, or a lock, usually.

Counter the Break

Once you can set up the flow, and break the flow, you can practise countering the break. So whatever action your partner breaks the flow with, you are waiting for it, and it fails against your effortless, beautiful counter. At this stage the flowdrill is now an unstable starting point from which to practice any drill.

And of course, you should decide between you whether you’re doing this co-operatively, or with one coaching the other, or competitively.

Slow Fencing

This is a really useful drill that gets you away from choreography but limits the chaos. Simply start from out of measure, and approach each other with the intention of striking. But you have to both move at half speed. Get hit, you lose a point, speed up, you also lose a point. You can use any and every action you know, and just play with it. If you find that you’re laughing, that’s good.

This gets you away from the formality of set drills and coaching, and allows you to be creative, spontaneous, and to figure things out. You go slowly so that you have time to try things that you’re not able to do at full speed yet, and to make it easier to figure out what happened after a point is scored (through the touch or through violating the speed rule).

Spend some time on this at some point in every class from now on.

Pressure Drills

You may have noticed that the core of my approach is to isolate variables. Technical complexity, for instance. Degrees of freedom, for another. Pressure drills are exercises whose specific function is to get you used to performing under, you guessed it, pressure.

The most basic set-up has you with two partners, all of you in full freeplay gear, whatever that is for your school or system. Each will give you the same attack, alternating, 10 times each. You will defend against each attack as best you can. In theory, you do the same technique perfectly 20 times. Your partners’ job is to keep you moving, keep you under pressure. They do not wait for you to sort yourself out after each action; as soon as your defence is done, the next attack comes in. That’s why you have two partners. As fatigue sets in, you will tend to make mistakes and get sloppy. Ideally, you will find it really, really hard to defend yourself. 

The next level has the attackers varying their attacks; one from the left with a cut and the other with a thrust, say; and you have to do two different defences correctly.

Then the attackers can attack as they please, their only job to keep you working under pressure. 

Of course, in whichever set-up, immediately after you were in the middle, you take one of the attacking roles, then the other, so on the fourth round, you’re back in the middle again. Good luck. You know you’ve really done this when you’ve felt like puking into your mask. 

It’s an age-old secret of martial training that acute fatigue is a good mimic of combat stress. When your heart rate is up at 180 and your legs and arms feel like they are falling off, and you can still get your actions right, then you have truly learned something.

Setting up Freeplay:

Think of a drill that allows either person to attack or defend, to respond with any action, and adds as many steps as necessary so that the play will continue until a technique is successfully concluded, and is conducted in the  ‘competition’ mindset (not choreography or coaching).

We call it freeplay!

Let’s start easy; if getting beginners to freeplay quickly is your goal, then all you have to do is have: 

  • Safer weapons and adequate protection
  • Rules, such as a list of what is allowed and what is not
  • Supervision: make sure that everybody is wearing the right protection and following the rules. 

As they develop their skills, your school might graduate them to more realistic weapons, or less protection, or allow a wider variety of techniques. Some students will be put off if the freeplay appears too unstructured, so be ready to introduce beginners gently, such as by letting a senior student who can be relied on to go easy on them at first be their first opponent.

Once you have agreed to fence with someone, it is important to agree on rules of engagement. This is partly to ensure safety, and partly to create an environment in which you can learn. The two most simple rules are these: confine your moves to the safety limits of your protective gear and to the technical range of the least trained combatant. In other words, do not allow face-thrusts when wearing open helms, or throws when one of you is not trained to fall safely. The rules can be adapted further to develop specific aspects of technique: for instance, you may not allow any close quarters work at all, or even restrict allowable hits to one small target. The idea is to come to a clear, common -sense agreement before facing off. You are only ready for no-holds-barred, totally “authentic” fight simulation, when you can enter such a fight with your judgement unimpaired. Following the rules of engagement will not make you soft, nor will it dull your edge if it comes to the real thing; rather it will develop self-control.

Here are my rules:

  1. Agree on a mutually acceptable level of safety. 
  2. Wear at least the minimum amount of safety gear commensurate
    with rule 1. 
  3. Confine allowable technique to those within the limits of your
  4. Confine allowable technique to the technical ability of the least trained
  5. Appoint either an experienced student or one of the combatants to
    preside over the bout. 
  6. Agree on allowable targets. 
  7. Agree on what constitutes a “hit”.
  8. Agree on priority in the event of simultaneous hits. Usually it is better to allow a fatal blow before a minor wound, but simultaneous hits
    should be avoided whenever possible. 
  9. Agree on the duration of the bout either in terms of hits, such as first to achieve five, or in real time. 
  10. Acknowledge all hits against yourself.
  11. Maintain self-command at all times. If you find yourself getting angry or frustrated, stop.

The rules can be adapted further to develop specific aspects of technique: for instance, you may not allow any close quarters work at all, or even restrict allowable hits to one small target. The idea is to come to a clear, common-sense agreement before facing off.

One of the hallmarks of a good fencer is courtesy. It does not matter what you think of your opponent (though it is usually safer to overestimate their skill) but it is essential to your development as a martial artist that you cultivate a respect for the weapons and the art. All bouts, however informal, should begin with the salute. After the bout you should always shake hands. I often acknowledge particularly good hits against me with a quick salute, rather than just raising my arm.

Freeplay as a Diagnostic Drill

Now that you have your rules, set up these drills, in a friendly but competitive way:

  1. You approach, your partner waits on guard. Your partner can respond as they like. Stop when one of you gets hit, or the play gets messy.
  2. Change the roles; let your partner approach, and continue as before.
  3. Begin out of measure, and you may both approach; continue as before.

Notice how and why you’re getting hit, or your partner is. After these three rounds you should have a very clear idea of at least one area in which you are weak, and one in which your partner is. Spend the next ten minutes or more working on your and your partner’s weakest link, then run the three rounds of freeplay again. The possible outcomes are:

  1. You have the same weakest link, and it hasn’t improved. So try a different approach for correcting it.
  2. You have the same weakest link, but it is getting better. So continue with the same training solution.
  3. You have a different weakest link. Fix that. Then run the diagnostic again.

Many clubs run freeplay at the end of class- letting off steam after the formalities are over. I think that’s sub-optimal. It’s better to incorporate freeplay as just another part of the training programme, to be approached in the same ways as every other drill. For diagnostic purposes, it’s best to freeplay first, then use the insights from the bouts to decide the topics of the class.

Choosing What to Train: Breadth and Depth

At any moment you should be adding either breadth or depth to your art. Adding breadth means adding new techniques or new concepts. Adding depth means taking what you already know and making it actually work. Coaching is the most efficient way to add depth.

Most diagnostic drills are specific to either breadth or depth. The drill ‘do I know the whole of the Footwork Form?’ tests for breadth. The drill ‘does my attack by disengage work?’ tests for depth. 

I tend to switch back and forth. Start with a little breadth (teaching the beginners a couple of things), then let them work on those things until they are a bit more solid, then adding the next new shiny thing. Generally speaking, adding breadth is more fun for beginners because they can clearly see that they are learning something. But at the advanced levels of the art, it is all depth training. Once you know all the techniques (and once you break it down there really aren’t very many), the rest of your training life is about getting those techniques to work. 

The cycle looks like this:

Probably the best way to see how I apply these principles in action is through one or other of my many online courses. You can get access to all of them here:

For more resources related to Solo Training, see here: Solo Training Resources