Famously, Wittgenstein used the word "game", and how hard it is to come up with a concrete, precise definition of its meaning, as an example of his theory of meaning-through negotiation. More prosaically, the word "football" describes a huge range of sports, some of which don't even involve using a foot on a ball most of the time (yet, outside of only semi-serious argument, we generally accept that all of those games are a "kind of football", even if they aren't "our kind of football").
That very tendency of communities to fixate on one example of a range of possibilities as "the prototypical" or "paradigmatic" example of the entire range is something which all endeavours focus on. Whilst it's useful as a shorthand for thinking, getting too focused on there being a "one true version" can also blind us to the range of options we have to change, adapt or otherwise improve.
As a mathematician and physicist by training, I tend to think naturally of rulesets as occupying a "point" in some abstract space. Similar rulesets are close together, more distinct rulesets far apart; but we can imagine a range of "directions" corresponding to different ways in which we can change the rules (or equivalently, different aspects of the rules being considered).
We can imagine, in this model, "Roller Derby" as a region in this space of all possible sporting rules. We have, as with Wittgenstein and "games", a bunch of examples of concrete rules which are definitely (or at least, strongly debatably) within this region, and I'd like to use comparison of their various differences to discuss what we can say about the range of possible rulesets which could exist.
The rulesets I'm mostly going to be comparing to are: all WFTDA rulesets (from v1 through to 2019), the current JRDA rulesets, the USARS/World Skate rulesets, the RDCL rulesets, the Japanese Roller Game, Roller Derby Sevens, Sur5al, PivotStar rules, Short Track Roller Derby, and the German "Street Roller Derby" ruleset. I'll also sometimes point at variant rules which have only really been used as one-offs in experiments or fun bouts, mostly to establish limits to various changes. For the WFTDA rulesets I'll use numbers and symbols in parentheses after WFTDA to specify if I mean a particular subset - (YEAR) is that year's edition only, (YEAR+) is that year and later rulesets, (YEAR-) is that year and previous rulesets.
All Roller Derby rulesets have a basic "skeleton" of assumed things in common. We might establish other limits on how different a ruleset can be, and still be "Roller Derby", but we're sticking to as small a set of axioms as possible here.
Our axioms are:
- Roller Derby "Games" are subdivided into 1 or more "Period(s)", each of which is further subdivided into 1 or more "Jam(s)".
- Roller Derby is played on a Playfield, which is an approximately Oval annulus: two Semicircular annuli connected by approximately Straight lanes (which may taper along their length). - That Playfield may be on a flat surface, or banked. (The majority of rulesets are flat track, and we'll focus more on these as banked tracks are inherently self-limiting in popularity due to the infrastructure requirements.)
- Games are contests between 2 or more Teams, each consisting of a Roster of Skaters (who will mostly use Quad Skates).
- Each Jam, at least 2 Teams field Skaters into the Playfield.
- The Skaters are divided into two classes: "Blockers" and "Jammers". - Jammers score points by lapping the opposition. - Blockers from both teams must stay within a certain distance of each other, to form "The Pack". - There may be a class of Blocker called the "Pivot", who can in some circumstances, replace the Jammer as point scorer.
- During play, there may be only 1 Jammer per Team on track.
- During play, there will be at least 1 Blocker per Team on track.
- Skaters may engage in [physical contact of some kind] in order to impede opposition skaters, especially the opposition Jammer.
- Skaters must stay within the Playfield ("Track") bounds.
- Penalties of some kind will be issued for violation of the rules.
Everything which isn't explicitly stated here is something which varies between the various rulesets which collectively represent "Roller Derby"
- Roller Derby "Games" are subdivided into 1 or more "Period(s)", each of which is further subdivided into 1 or more "Jam(s)".
We can look at division and provision of playtime from two sides: how much time is allocated to each, and how many subdivisions we make of the various units. Many rulesets mix these two ways of talking about playtime and hierarchy.
The simplest possible ruleset in terms of subdivision is already represented in Sur5al. A Sur5al "Game" consists of a single "Period" which is also a single "Jam". Sur5al's duration is set by the duration of a Jam, which it takes from WFTDA Derby: 120 seconds. In all rulesets, the "Jam" is the quantum - the smallest subdivision - of playtime, so it's useful to think about how its duration affects play at this point.
We could imagine longer or shorter "Sur5al variants" based on other Jam durations: Jam durations of everything from 60 seconds through to 120 seconds are in use in various active rulesets. We've seen "fun" games play with Jam durations of up to 600 seconds or more, but not in anything resembling standard play.
|Ruleset||Jam duration (seconds)||Early calloff?|
|Short Track Roller Derby||60||No|
What effect does changing Jam duration have?
Changing Jam duration has some obvious, and some not quite as obvious effects. Most directly, it determines the length of time that skaters (and skating officials!) must be physically exerting themselves, and mentally focused. Shorter Jams allow for more intense bursts of activity and concentration; but can also make it harder for skaters to flip in-and-out of "focus mode". Conversely, lengthening the Jam duration makes competition more of an endurance sport, but is simultaneously more tiring on referees, who must also focus for longer periods without a break. Slightly less obviously, longer Jams (especially without call-offs, see below), reduce the amount of "luck" in the game (other than penalties) - if we consider a Jam as a series of measurements of how effectively the two Jammers can get through the pack [one measurement per pass], then having longer Jams gives more measurements, and therefore a better set of statistics (variance of the sample mean). Longer Jams without call-offs also allow very mismatched teams to avoid scoreless blowouts for the lower-rated team, by giving their Jammers more time to eventually break through the pack at all.
Most active rulesets allow a special skater the right to stop a Jam early, before it reaches its default duration. In almost all cases, this skater is one of the two Jammers on track, and they are designated "Lead Jammer".
WFTDA-family rulesets differ from the MADE and USARS rulesets as to how "Lead Jammer" is defined and retained. In WFTDA-family rules, the Lead Jammer is the first Jammer to complete their initial pass legally and within bounds - the status can be lost by receiving a penalty subsequently during the jam, and cannot be transferred with the star to the Pivot, but is otherwise a permanent property of that Jammer during the current jam. In MADE/USARS family rulesets, the Lead Jammer is the Jammer currently in the lead at any point in time - if one Jammer passes the other, they will also gain Lead Jammer status from them at that point. (This is the same in RDCL, which also permits Lead Jammer status to be held by a Pivot who has successfully taken a star pass.)
The two families of Ruleset also differ as to when a Lead Jammer may use their ability to "call off" a Jam: WFTDA-family rules allow this at any point, even if the skater is out of bounds or out of play; MADE/USARS family rules have additional requirements (usually that they be in play).
Some rulesets, most notably Short Track Roller Derby out of the active rulesets, do not use Lead Jammer status: all Jams run exactly to their specified duration (unless injuries or other emergencies require a stop of play).
What difference does allowing call-offs have on the dynamics of the game?
Call-offs, especially with WFTDA-family rules, have a tendency to amplify the advantage of the stronger team. This is trivially the case, in that a team which consistently gets its jammer out first can always guarantee that the opposing team never has a chance to score - the jammer just needs to call the jam as soon as the opposing jammer completes their initial pass (barring penalties). Without call-offs, the weaker team will usually have an opportunity to score a few points each jam; it's unusual for two teams to be so unevenly matched that a jammer is held for the full duration of a jam.
Simulation of jammer versus jammer statistics, allowing and not allowing for call-offs, suggests that this effect is particularly strong for "moderate" differences in team strength, resulting in effects on the order of...
More obviously, call-offs mean that jams are of uneven length. For rulesets where Period duration is measured in total time, not number of jams, this means that the fraction of Period time consisting of actual "on playfield competition" can vary dramatically, depending on the closeness of the teams.
For example, in WFTDA-family rules, the average duration of a jam (for games where records exist) is closer to 60 seconds than the full 120 seconds of a natural jam - but the length of a given jam in a game can vary between 30 and 120 seconds.
Periods & Games
Periods consist of one or more Jams (as above). Other than Sur5al, most Rulesets have more than 1 Jam per Period, but the manner of subdivision differs.
Most rulesets prefer to specify the total duration of a Period in Time, and then essentially have Jams until we run out of Period time. (Depending on the ruleset, the Period may end strictly when its clock expires, or wait for the current Jam to end first; early WFTDA rules did the former, but changed quickly to the latter over time - tightly scheduled games may revert to the former policy as it makes duration more predictable).
Whilst many rulesets have either 1 or 2 Periods per Game, there are a few which have options for more than 2. In these cases, it's almost always the case that the total Game duration is conserved, and Periods get evenly distributed fractions of this. (So, for the case of MADE, for example, a Game consists of 60 minutes of playtime - but you may play 2 Periods of 30 minutes, 3 of 20 minutes or 4 of 15 minutes.) RDCL is a bit unusual in that it provides a 2 Period version which simply halves the length of the result game (we've tabulated this as RDCL(Short) below).
|Ruleset||N Periods per Game||Period Duration||Game Duration (track time)|
|Roller Derby Sevens||1||21min||21min|
|Short Track||2||10 jams||20 jams|
Rosters and Lineups
The number of skaters in a Roster varies by less than an order of magnitude across the rulesets we're considering. By definition, the smallest size a Roster can be is determined by the number of skaters expected to be on track each Jam: Sur5al then once again provides our minimum value here, as with a single Jam per Game, all Rostered skaters play.
We can express other Rostering limits for rulesets in terms of the multiple of "Skaters on Track per Jam" ("lineup") they represent.
|Ruleset||Roster size||Multiple of Lineup size|
|WFTDA (<2018)||14||2 4/5|
|Roller Derby Sevens||7||1 2/5|
|Short Track||7||2 1/3|
Other than Sur5al, it seems that most Rosters try to be around 2 to 3 times the size of the "active lineup", presumably as this provides a nice strategic sweet spot for bench coaching decisions. Roller Derby Sevens is unusual in being a multi-jam game with a very short Roster for its Lineup size: the reduced duration of the period compared to most versions makes this less of an endurance challenge than it might otherwise be, but it's still a significant tradeoff. (It has been argued that this reduced roster size contributes to the "tactical challenge" of Sevens play, as you have a much tighter set of limits to work with.)
The total lineup size varies less dramatically: almost every ruleset sticks to the model of 4 Blockers and 1 Jammer on track per team. The one significant variant, Short Track Roller Derby, has 2 Blockers and 1 Jammer; but also shrinks the Playfield size significantly (in fact, to <50% of the total area of a WFTDA or USARS track).
This reveals a natural and obvious fact: the lineup size is inherently tightly coupled to the track size, as this determines the difficulty of scoring. For rulesets which penalise skaters by sending them off track during play, this is obvious by observation - losing a blocker from the track significantly reduces the effectiveness of that team defensively, even if the opposing team also has a reduced number of blockers on track. Conversely, adding more blockers to both sides makes it much harder for Jammers, simply due to the reduced free space on track, before any other effects add to this.
It's worth noting that exactly one example of a track significantly smaller than the Short Track exists, and that experiments with "Micro Derby" suggest that just a single blocker is sufficient at this scale. This is a natural limit at the small end of lineup and track sizes.
At the larger end, Helsinki Coast Quads possess the only "Big Track" - a Roller Derby track erroneously laid in one of their venues to a larger scale than the standard WFTDA, World Skate sizes. They've, mostly for fun, explored various alternative rules for playing on such a large playfield,
There are also combinatorial effects on the number of blockers on track during a jam: more blockers allows more possible combinations and coordinations between them. The currently popular braced-wall needs 3 blockers to do - the equivalent as a dyad is less effective due to the lack of "width" provided by a single player, as much as any other factor.
All of the existing active rulesets consist of contests between exactly 2 Teams. This does not mean that it is impossible to conceive of versions of Roller Derby with more than this: we're aware of "fun" games being played with 3 or 4 active teams at once.
As we discussed in the section on Packs, the Pack size is strongly entangled with the Playfield design, in order to control the "difficulty" of the game. A narrow playfield can be effectively defended by very few skaters; whilst a wide Playfield would provide no challenge for a Jammer without enough skaters to effectively interdict across its width. If we increase the number of Teams in a Game by allowing all of them to have skaters on track simultaneously, then we also need to adjust the number of skaters per team per pack, or adjust the side of the Playfield.
It's hard to model precisely how much harder it will be to pass a pack with additional skaters provided by more than 2 teams, but if we consider the worst case of linear scaling (that is, doubling the number of skaters doubles the width we need), then we can make a chart of how pack sizes (per team) would scale using a standard WFTDA-width Playfield.
|N Teams||N Blockers||Total Skaters on Track (inc Jammers)|
Simultaneous 3- and 4- competitor games are not uncommon, outside of track sports - many card games have competitive forms with this many players, as does (for example) Mahjong; there are also multi-competitor versions of games like squash and bowls. In general, adding additional competitors introduces significant new tactical and strategic space into the game, as you now have more choices to make concerning prioritisation. It would be interesting to see more experimentation with this in the space of Roller Derby, as the tactical space is already very rich due to the simultaneous offence and defence involved.
Making this change would also result in decisions being needed in the realm of Scoring, as presumably each team's Jammer would be eligible to score points for every member of each opposing team, not just a single opposition.
We imagine that these variants would be somewhat harder to officiate, both for on-skates officials (due to the addition of more team colours to differentiate), and for non-skating officials (who would have more teams to track in every role); the reduced number of on-track skaters for each team will probably not compensate for this fully.
The other possibility is to allow for Games with >2 teams by assigning different pairs of them to the track for different jams. (This is not quite the same as the multiple games of "2 team/ binary" roller derby you could imagine being interleaved here, as your tactical choices for roster selection will be closely entangled with the ordering of your opponents.)
In this sense, the only existing version which is similar is the Japanese Roller Game, which has only 1 Jammer per jam [the teams alternate, jam by jam in playing "offence" with a Jammer, or "defence" without one]. You can imagine this as a sort of "1 and a half" team limit of as few teams as possible, only possible via alternation.
Including the Roller Game alternation mechanism, then, there are two obvious ways of doing alternation here:
- complete round-robin matching of all teams ([AvB,AvC,CvB] repeated motif) across the jams.
- differentiating "active" and "passive" teams each jam, with only active teams fielding a Jammer.
Both approaches need consideration of fairness in the number of jams provided: ending a Period without completing a complete set of "matches" would be unfair to at least some of the competitors.
In every active version of Roller Derby, and most historical versions, scoring is via the team's Jammer (or Active Point scorer in versions with Pivots) passing members of the opposing team in laps.
There are two main subdivisions on this, which themselves have some internal variation.
WFTDA, USARS and the majority of rulesets derived from them award 1 point per skater. Passing all the members of a 4 skater lineup as a Jammer would give you 4 points; "incomplete" passes are usually awarded a number of points equal to the number of skaters who were passed.
PivotStar and Short Track Roller Derby (and some others) award 1 point per pass. A "complete pass" awards 1 point in total; incomplete passes usually award nothing.
In the "1 point per skater" family, there are differences as to how "successfully passing" an opposing skater is determined. WFTDA(2019), for example, considers it impossible to "pass" the opposing Jammer - there are no "Jammer Lap points", whilst WFTDA(2018-) all consider this possible. USARS and WFTDA differ on how
Penalties & Contact
Most Roller Derby rulesets have broadly similar concepts of what "permissible contact" consists of.
Low-Contact & No-Contact
The JRDA rulesets provide alternative definitions of "permissible contact" for Level 1 and Level 2 play.
This same approach has been copied over in some regions as a modification to WFTDA rules play. Especially in Australia, you see "LOCO" bouts advertised, which essentially use the same modifications as Level 2 JRDA but playing otherwise WFTDA regulation games. These are especially popular for introductory or rookie games or scrimmages.
In a similar way, the Short Track Roller Derby ruleset has "No-Contact" and "Low-Contact" variants, again using the same changes as JRDA uses for Level 1 and Level 2 contact.
Where do we talk about track cut relativity?
Most modern rulesets follow the WFTDA & World Skate models of penalties being issued and served during play. Penalised skaters leave the track and head immediately to a penalty area, where they sit for a period of time (60 seconds in WFTDA(2014-), 30 seconds in almost all rulesets now).
RDCL does not require penalties to be served during play [except when it would be dangerous not to remove them]. Instead, skaters serve a penalty by sitting out for the subsequent jam [or later jams if the penalty box is full for their team already], requiring their team to skate with 1 fewer skater in the lineup.
Short Track Roller Derby is unusual in that penalties are served by subtraction of 2 points from their team's score. (As Short Track is a 1 point per pass game, this is a significant effect on their team's total points.) Skaters do not leave the track except in cases of expulsion.
German Street Derby
An unusual special rule, which we've seen used in a small number of "fun games", and is also a component of the recently publicised "Roller Derby X", is exchange. Rather than having skaters assigned to the track for an entire jam, exhange allows them to "tap out", swapping with another skater on their roster, whilst the jam continues.
We'd expect rulesets implementing exchange to require more Officials than rulesets without, as the exchange process will require its own monitoring.
Exchange also requires changes to
Almost all revival rulesets on flat tracks inherit some version of the "tapered" track which appears in the earliest formalised versions of the rule. This tapered track implies a particular rotational direction - the tapers make it easier to corner when moving "derby direction", which is usually counter-clockwise.
There is, of course, no reason why a particular ruleset needs to use "derby direction" as counter-clockwise - you could just mirror the track tapers and skate clockwise instead. Mirroring has no effect on the rules themselves, although skaters who have trained on an oppositely-handed track for years will themselves have an unevenness in development.
If, instead, the track is made perfectly symmetric - removing the tapers on the straightaways so that they're just straight corridors - there is no longer an inherent bias towards a particular direction. Rulesets with a non-tapered track can therefore alternate direction of play between units of play.
The only extant ruleset which does this is Short Track Roller Derby; the first period is played clockwise, and the second counter-clockwise. (The Short Track is also physically smaller than the WFTDA and other standard tracks, as well as being symmetric.)
You can easily imagine rulesets which alternate direction every jam, although this would involve more potential error in tracking the orientation of play between jams. Alternating direction between periods, as in Short Track, adds almost no officiating overhead.
The effect of alternating direction within a game is limited with respect to the gameplay itself. (Reality itself appears to have no bias between clockwise and anticlockwise motion, physically.) However, on a training level, alternating directions "evens out" the exercise bias which players are exposed to - leaving them more symmetrically developed, and potentially reducing long-term issues in development.
Most revival rulesets include the third player type - the Pivot - who act as a Blocker usually, but have some Condition which allows them to replace the Jammer as the Active Scorer. (In all rulesets with this rule, the Jammer loses the ability to score once the Pivot has activated via this Condition.)
In WFTDA-family rulesets, Pivots "activate" by physically receiving the Jammer's helmet cover (with a star on it). In USARS-family ruleset, Pivots "activate"
There is no particular reason why a ruleset needs to allow Pivots, of course. Short Track Roller Derby does not have a Pivot position - skaters are either Blockers or Jammers.
Rulesets with Pivots are harder to officiate; referees must be aware of the Active Scorer changing [and determine if the change itself was legal], and officials tracking lineup positions, and scoring data, must also be aware of these shifts.
Tactically, rulesets with Pivots make it easier to get your scorer through the pack. (They do this both by giving the Jammer the opportunity to pass over their role; but also by forcing defensive play to pay attention to both the Pivot and Jammer positions, if the Conditions for Pivots to become active look likely.) For rulesets which allow the Lead Jammer to end a Jam early, this should have the effect of causing Jams to run slightly shorter (on average) than without a Pivot - the second Active Scorer is less likely to be held longer after the Lead Jammer gets out; it may not actually help the lower rated team to score more effectively, due to the call-offs themselves. For rulesets without call-offs, we'd expect scores to be slightly closer; although the effect of a Pivot change is most significant for only 1 pass in the Jam.
Surfaces / Skates
The name "Roller Derby" implies some aspect of "Rolling"; but many leagues have also used variants of their favoured ruleset in which players wear no wheels on their feet. "Sock Derby" is the common name for this variant, which is usually used as a training tool to teach the tactics of "Roller Derby" to trainees whose ability to skate is still developing. However, there's no obvious reason why you couldn't play "Sock Derby" versions of any of the Rulesets in this article in a competitive mode.
Moving from "skating" to "running" as the mode of locomotion would affect many things about the dynamic of the game (velocity and acceleration) and make blocking easier (because feet and shoes have better sliding resistance than roller skates), but the game would remain recognisable as in the same family as "Roller Derby."
(There's a philosophical argument here as to if this is a sufficient change to stop the sport being a kind of "Roller Derby", but it is clearly very closely related.)
Similarly, by analogy with the existence of both Roller Hockey and Ice Hockey, one can imagine an "Ice Derby" differing principally by being played on an ice rink, by players on ice skates. (Exactly one test of something like this is publicly known: ). Given the sharpness of ice skates, and the risk of serious injuries, we would expect a putative Ice Derby to require additional protective gear - similar to the gear worn in Ice Hockey.
Despite the significant increase in injury risk, the other change to the dynamics of ice derby would be in blocking effectiveness. The sliding friction on ice for good ice skates is x compared to the rolling friction of quad skates y.
Finally, there are Wheel-chair versions of many sports, including full-contact sports like Rugby. As Roller Derby is already played with the assumption of Rolling motion, one can imagine a "Wheelchair Derby" ruleset family being possible, with players all on contact-sport-compatible wheelchairs. This change would almost certainly require changes to the track dimensions - wheelchairs are wider than most people on roller skates - and additional alterations to the contact rules, as contact zones clearly do not map to competitors in chairs.
In all of the above cases, we would expect that the officating load would remain almost the same as in the "Roller Derby" they are modified from.
In Medieval maps, the stereotype is that unexplored regions would be marked with impressive sounding warnings to fill space "Here be Dragons". This section tries to summarise the above discussion in terms of those "Dragons" for the space of possible Roller Derby rulesets - combinations which haven't been explored yet.
Some of these may prove to be completely uninteresting, definitely worse than any existing ruleset. But some may prove to be exciting games in themselves, just for want of exploration.
There are some interesting possibilities for rules which violate some of the axioms we started with for Roller Derby. These resulting games probably aren't kinds of Roller Derby anymore, but we'd be interested to see how well they worked.
Tagging for lapping
Roller Derby scores points by lapping opposing skaters. Without introducing a ball, the other common way to score points in contact sports is tagging - for example, in Kabaddi, and several attested games played in various Mesoamerican cultures.
A tagging-based sport on Roller Skates, with contact rules from Roller Derby, would probably need a different track (if we're not lapping, we don't need to be constrained into loops)