Failure to Withdraw – A Game Design Puzzle

Since Summer of 2021, my club Bucks has been using an in-house ruleset that I call “Weighted Right-of-Way” (WRoW), in which all clean hits are worth 2, and a double hit with priority is worth 1. Before I get into it, what is “right-of-way” anyway? Basically it’s a method of resolving double hits (both fencers hit each other at close to the same time) in which the outcome is decided by who was more tactically “correct” based on the ruleset. In general, in our ruleset and almost every other right-of-way ruleset, the attack has priority over the counterattack – what that means is if fencer A initiates an attack and fencer B chooses to attack into it instead of defending and the exchange ends in a double hit, fencer A will get it, because fencer B should have defended. Fencer B of course will still score if they hit without getting hit, RoW only matters if the exchange was a double.

The theoretical idea behind the weighted ruleset is to resolve doubles via priority by rewarding the fencer who was less at fault in the action, but also encourage fencers to attempt to land clean hits by rewarding them with a 2 point hit. At the time of this writing, we have used the ruleset in 7 small local tournaments, and there have of course been tweaks and fixes over time as new situations and edge cases arise. One in particular I found to be very interesting, and to solve it I had to dig deep into the purpose of this ruleset and what it is really trying to accomplish. It goes like this:

Situation: A lands a quality attack on B, B lands a no-quality counterattack at the same time

Question: In the weighted right-of-way ruleset, should A be awarded 2 points or 1?

Though it may seem like an insignificant choice, the more I thought about this problem the more I noticed it really gets to the heart of the idea of what the ruleset is trying to accomplish. The question for me boils down to this: when we reduce the score in a double, are we punishing the attacker for not hitting cleanly, or are we rewarding the defender for being able to land a hit? 

This reminded me of the wording of different types of “afterblow” (hit after being hit) rulesets. Some use the language “hit with afterblow,” which implies that the ruleset rewards the defender for getting a hit in, and some call it “failure to withdraw,” which in my opinion implies that the ruleset punishes the attacker for not hitting cleanly. In this article I am not talking about afterblows but doubles with priority, but it is the same idea.

Case 1: Punishing the attacker

In this case, we assume that the default score for a hit is 2, and if the person with priority is hit in the process, we punish them by reducing their score by one. There are two ways that I have thought of to look at this: from the internal perspective of the attacker, or from an external perspective of the exchange. From the fencer’s perspective, they earn their full score of 2 points as long as they hit without getting hit, therefore if they did receive a hit, their score should be reduced. Quality of the hit should not matter, because it isn’t the fencer’s fault that the defender hit flat or didn’t have a good arc, they got hit, and getting hit can’t earn you two points, end of story.

From a universal perspective, we reward 2 points for a clean exchange. Similarly, if both fencers got hit, even if one hit was not quality, the exchange was not clean. One fencer was certainly in a better spot than another, but ultimately the exchange could have been cleaner, therefore full points can’t be awarded.

Neither of these explanations are satisfying to me, because I believe they assign too much responsibility to the attacker on making sure the exchange as a whole is clean. While the reduction in points does make the situation non-ideal for the attacker, the purpose of the ruleset is to encourage the attack, and adding some extra possible reluctance does not help advance that goal. This also adds an extra level of subjectivity, especially in terms of the universal perspective, because it defines a clean exchange as an instance where one person does not get touched at all by the sword, which is outside the purview of the rules and their quality requirements. There is already enough subjectivity baked into the priority system in the first place, we don’t need to add any more.

Case 2: Rewarding the defender

Another way we can look at the reduction in points is as a reward for the defending fencer; they were less correct by the standards of the priority rules, but at least they still hit something. By landing a hit, they are rewarded with -1 points for the opponent, therefore modifying their gain from 2 to 1. If this is the case, then quality must be taken into account. In order for the fencer to earn the reward of subtracting a point for their opponent, their strike should meet the standards of any strike. 

My issue with this is that the intention of the rules is not to encourage fencers to hit without priority, it’s the opposite. If a fencer counterattacks to the leg instead of parrying, their opponent will continue to score, and that’s because that is the fencer we are trying to encourage. 

Case 3: Going back to the rules

It turns out that the solution to this problem already existed in the rules as written. A valid hit earns the fencer 2 points. When two valid hits occur, the priority rules are used to determine who earns one point in the exchange. The key is right there, a valid hit; a hit with no quality is not a valid hit. Therefore, if the attack is quality and the counterattack is not quality, only one valid hit occurred, and the attacker earns 2 points. The same is true if the attack is not quality and the counterattack is quality, even though the attack has priority, the only valid hit was the counterattack, and thus it earns 2 points.

By doing this, I get to sidestep any issues of who we are punishing or who we are rewarding in each scenario. What counts as a valid attack has already been established, and we are simply following the existing procedure. If I were to come up with a different outcome for this exchange (call it 1 point instead of 2), I would either have to change the definition of a valid hit, or modify the point criteria in some way to indicate that sometimes non-valid hits can score. This would in essence create a new class of recognized hits (no hit, hit that is non-valid but can subtract a score, valid hit). It would also muddy the waters in a situation where a non-valid hit has priority in a double against a valid hit – who would score in this situation? 

I made a similar choice regarding tip cuts in the quality criteria. I wanted to discount cuts that just skim, but the language for that already existed in the ruleset. A cut is defined as a strike with the edge of the sword, therefore cuts with the tip (and with the flat) do not apply. 

Mindset for dropping the score from 2 to 1How that relates to scoringScore for AWhy I like or dislike it
A is punished for not hitting cleanly (“failure to withdraw”)A would earn 2 points for the act of making a hit without getting hit, so because they got hit, regardless of quality, the score is reduced.1Dislike – Assigns undue responsibility on attacker for the quality of the exchange as a whole
B is rewarded for landing a strike (“hit with afterblow”)B’s counterattack is a reward for B, earning -1 points for the other person, therefore it must be a quality hit in order to be awarded.2Dislike – Implies incentive of counterattacker to double or afterblow
We acknowledge that A was more correct, but the exchange could have been cleanerBecause both fencers got hit in some way during the exchange, it could have been cleaner, therefore should not earn full points.1Dislike – It makes a subjective judgment on what a clean exchange is supposed to look like
When both fencers score a hit, we use priority to decide who was less wrongBoth fencers did not score a valid hit, therefore right-of-way is not in play.2Like – It is more objective, we have objective quality standards, and does not rely on our opinion of what a clean exchange is supposed to look like

I don’t have any specific takeaway or conclusion for this one. I just wanted to share my thought process when trying to come up with a solution for an edge case situation in a ruleset. I always like to think about how the possible outcomes for each case relate to the intent of the rules, and what else might have to change in order to accommodate a theoretical edge case resolution.