The Direct Attack Drill is Special: Self-Scaling Games

The direct attack drill is a drill that I was first introduced to by Adrien Pommellet, who learned it from a Giovanni Rapisardi video. I later read about it in the book Make the Cut by John Chow, in reference to modern Sabre fencing. Over the last couple of years, we have used this game to great effect at Bucks, and as I understand it, it has become a staple of the repertoire of the other GD4H coaches as well. The drill goes like this:

Fencers take the roles of attacker (A) and defender (B). A starts in a point high cutting guard, B starts in a low guard. Starting from a fixed distance, A throws a direct cut to B’s head, which B attempts to parry. If the cut lands, A takes a small step back, if the cut is parried, A takes a small step forward, and the drill repeats. 

It is a very simple competitive game with only one path to success for each player: land a hit or parry. The attacker gets to choose the timing and rhythm of their hit, but that’s about it, only direct attacks are allowed, so they may not feint or attack an unexpected target. The purpose of the drill is for the attacker to make their attacks as efficiently as possible, and to find the distance at which they can expect to reliably land a direct attack (Chow refers to this as the “PONE,” or “point of no escape” in Make the Cut). Secondary purposes are perception and parrying ability for the defender, and working on the “ballistic passing step” motion for the attacker. 

The thing that makes this drill special is that it is what I consider to be “self-scaling.” By that I mean, no matter what the relative skill levels of both participants are, they can both participate in this drill and try as hard as they can. This is possible because it gets easier as you get closer and more difficult as you get far away. There will always be a distance that is so close that the attacker can’t help but hit, and a distance that is so far that the attacker has no chance of hitting. If both players have a very high disparity in skill, the game will proceed the same as usual, but the lower skill player will end up attacking at much closer distances, and the higher skilled player will end up attacking at further distances. 

The self-scaling aspect of this drill is extremely useful. Usually in a competitive drill, if there is a large skill disparity, either the more skilled side will end up dominating and have a lot more successful reps than the less skilled side, or the more skilled side will lower their level to meet the less skilled side. The first method is okay for the more skilled side but less useful for the less skill side, and the second is the reverse. If all drills were self-scaling, this would not be a problem, because both sides would be able to try as hard as they can and still get something out of it. 

Okay, so self-scaling drills are good, so let’s make all drills self-scaling, right? Unfortunately that’s easier said than done. The mechanic that makes the direct attack drill self-scaling is starting from a different distance every rep. The problem with this is that it can’t be used when constantly adjusting distance is an aspect of the drill, which it almost always is. Nearly every skill, tactic, technique, and strategy in fencing involves manipulating distance through footwork, and if free footwork is allowed, then you can’t manually mandate a specific distance in the drill. 

It is possible that there are other ways to make self-scaling drills, such as giving a side a certain number of strikes that can increase or decrease, or changing the value of a successful rep throughout the match, but so far I have found none that are as elegant as the direct attack drill. Others have come up with some ideas for this, which you can find if you search the “self_scaling” tag on the Game Archive. Finding a drill that is both as self-scaling and tactically/technically applicable as the direct attack drill would be very valuable. Finding a formula for creating self-scaling drills would be a holy grail for game based teaching.