Perhaps even more than the particulars of attack and defense, there is a game design decision that influences the entire feel of the game, especially in combat. That decision is how much time passes between actions.

For some games, that decision is precise and, in a very real sense, defining. For others, it borrows a now-marketable line from Dr. Who:

Each method has its advantages and disadvantages, and each can both promote and inhibit drama and willing suspension of disbelief. This decision also dictates tactics.

But first, let’s take a look at our Fab Five. This is going to go a lot faster than usual.

Ah, turn length. Turn length in D&D has moved around quite a bit, and Random Wizard has created a handy compilation of how this has changed over time. The current version adopts the same six-second convention used from D&D3 onwards.

Overall, turn length has varied from (more or less) one effective action per minute to as many as ten, thus the convenient (for easy division) into the otherwise perhaps-odd choice of six seconds per round.

Each player and each side gets to go in initiative order within that turn. The variety of methods to determine who goes first are many and varied. Fifth Edition has each player roll a die individually, and that sets the turn order. Some groups reroll initiative each round, while some will do it once per fight.

There are held actions that allow one to take their action after someone who would usually act afterwards, which enables team tactics, or simply delaying your moment in the sun until the tactical situation has resolved itself more.

Likewise to D&D, the turn length used in Savage Worlds is mentioned roughly once, stealthily inserted into the text at the beginning of the combat chapter on p. 64 of Savage Worlds Deluxe: “When a fight breaks out, game time breaks down into rounds of six seconds each. Ten rounds, then, is one minute.”

The key here, again, seems to be that there are a nice ten rounds per minute.

Who goes first is done by drawing from a deck of playing cards, and resolving actions in descending order, from Ace to Deuce. It is possible to delay an action to subsequent turns, by going “On Hold,” which allows you to attempt (there’s a die roll involved) to interrupt the action of any other character at any time. You may hold an action as long as you like . . . but you’re not doing anything else but holding your action until then. The die roll means that if you lose, your foe still gets their action in before you do. You’ll be “dealt back in” for initiative order next time.

If GURPS isn’t the finest resolution game ever, it is certainly the finest resolution in turn length presented here. Each turn for a character in GURPS is one second long. The trick here is that there is no universal turn In GURPS – each character goes in order, and each character acts once per second, but the turns of all characters are interleaved, and if there are four characters in a fight, that is not saying that character B goes 0.25 seconds after character A.

Characters act in the order of descending Basic Speed (a figured quantity), with ties broken by who has the highest Dexterity, then skill level, and then whatever the GM and players decide on to break ties.

Delaying who goes when is done by a particular choice – a “maneuver” called (unsurprisingly) Wait. A Wait allows action if a very specific trigger, player definable, occurs. “I’ll hit my foe after Bob goes” is definite and legal, as is “I wait for my foe to peek around the corner, then I shoot him in the head.” But “I wait until something interesting happens, then I’ll do something based on that” is not a well-defined trigger, nor does it invoke a specific action afterwards. The conditions may be broad but not vague, and are basically IF . . . THEN statements.

Normally one does not simply walk into Mordor and lump games together. And yet Fate and NBA both make the same basic choice with respect to the length of a turn, which is that a turn takes however long it needs to, no more, no less. Quite simply, both games embrace the shrug-driven philosophy of J. Michael Straczynski when asked precisely how fast certain classes of ships (my brain tells me it was the White Stars) was capable of moving through space. His reply?

“They move at the speed of plot. What of it?”

So it is with these games. A turn might be a split second, or an hour of weaving through traffic in rush hour. It might be a full round in a boxing or MMA match, or it might be a single blow.

There’s more or less no point to holding an action in Night’s Black Agents. By and large, attacks and other aggressive behavior is against a fixed target number, so going second is no help. Even if you do narratively go “second,” your action might be “I wait for the minion to make a mistake, then dart in and try and hamstring him with my razor-sharp vampiriblade!”

If the player has the initiative, she can describe the action in any suitable fashion.

Fate takes a similar abstract tack on things. Movement is separated into arbitrary zones that take an action or so to cross, assuming you can Overcome whatever obstacles (people, things, situations . . .) are in your way. Including other characters trying to do you in.

How does timing drive the feel of a game? It drives choices and complexity. It speaks to tactics – how much can a character do; how much can an adversary do before that character can react to it? It also has a surprising effect on group dynamics.

Starting with the finest resolution game, GURPS, is instructive because it takes very small snapshots of what happens in a fight, encouraging/mandating very discrete actions. While there is some small amount of abstraction in a GURPS fight, mostly each move and countermove is declared – and with optional rules turned on, the specificity of the described action can be declared is quite high.

The abstraction is usually in terms of making sense of a series of die rolls from turn to turn. As an example, a fighter might be attacked by four thugs, and he might parry a blow from one, but then dodge two others, retreating against one of the attacks. The fourth missed.

This is resolved with three die rolls, but would be understood as mostly simultaneous movement. Still, this starts as many arguments as it explains at times.

The blow by blow, second by second declaration means that there’s a conflict of expectations for players used to more exposition of actions. Drawing a weapon might be a full-second action (a Ready maneuver), moving into combat range might be a few seconds more, and then blows might be exchanged one or two at a time until a victor emerges. A bout of modern longsword fighting that has a primary exchange lasting perhaps two seconds is resolvable down to nearly the millisecond.

That’s a lot of agency to hand players and GMs, and for those that like resolving fights where very little falls to the realm of abstraction, this is a great thing.

However it has its issues. I’m always shocked at how much people assume they can actually accomplish in a single second, and how long some events take. It is easy to lose track of how long (on a one-second scale) actions can take. Stooping to retrieve car keys is probably resolved with a single one-second Ready maneuver, but in reality, it might be several seconds to pick them up, depending on how much casual fumbling one does.

Consider this Pankration demonstration (at least I think it’s a demonstration – there’s not really enough resistance by the defender). I copied the video, and used an editor to add a quarter-second time clock to the demonstration, but one thing that stands out is that the actual throw – the time the recipient spends airborne – takes more than one second from initiation to completion. For a game that is resolved second by second, taking a turn to pause while the foe (and the thrower) sail through the air is highly unusual, and realistic or not, would represent a perceived loss of agency. Players will decide to forgo otherwise-excellent tactical choices because of a gap in their ability to choose a viable action each turn.

On the other end of the scale, games like Fate and Night’s Black Agents attach no particular amount of time to any action or round. A declared action could as easily be “I drive across the city” or “I run five miles” as it might be “I shoot the vampire in the head.” The sub-second resolution required for fine resolution of dramatic moments in “bullet time” is as accessible as hour-long actions involving two agents shadowing a vampiric minion.

While this works well when every player wishes to take actions of equal length, chronological dissonance can result if one character is taking actions that can fit well within (or drastically exceed) the span of time of another’s.

While split-second increments can lead to frustration if each turn is not filled with action, vague time increments can also lead to confusion or situations that result in a breakage of suspension of disbelief, as one action might be “I shoot the bad guy” and the other might be “I reprogram the computer on the super-dreadnaught to process food instead of supralight navigation.” That can lead to the same feeling of agency loss, as those not taking long actions may be given many choices in a scene, while the person reprogramming the computer is basically off camera the entire time. That’s not unrealistic, nor does it break immersion or suspension of disbelief. It is, however, boring for the player taking the long action.

Still, there’s a lot to recommend to abstract action timing. A flexible GM can adjust timing and action order on the fly in order to both preserve immersion as well as ensuring that the passage of time is as elastic in the game as it can sometimes seem in real life.

A boxing match might be fifteen three-minute rounds. Each round is thus 180 turns in GURPS. In contrast, one might treat a boxing round as a few exchanges in Fate, representing a lot of punches and footwork, with an opportunity to inflict stress or consequences only a few times per round. We’ll return to the boxing match in a moment – it’s a very convenient example for what happens in a fight (though a circumscribed one – no takedowns, elbows, or other varieties of violence – but the rules can be acknowledged without making the example irrelevant).

Intermediate between GURPS and the pair of NBA and Fate are games that seem to have settled on a six second time scale.

The choice of six seconds was probably a matter of convenience, originally. A nice, even 10 turns per minute, and you can see this division in the progression of turn length through the various editions of D&D.

While the six-second time span may have been a matter of convenience, it happens to represent (in my personal experience) a near-ideal division in terms of granularity. It also makes some amount of sense when benchmarked against real-world sporting events, such as boxing. CompuBox is a website that quantifies such things, and this data talks mostly about records – but it also compares to averages. The junior middleweight average “power punches” per round (180 seconds) is 12 landed punches. That’s a solid hit every 20% of turns. In D&D terms, it’d be about 1d20+5 (say, a low-level Fighter with +2 proficiency bonus and +3 for STR 16 or STR 17) against AC 22, assuming that a “power punch” equates to a hit good enough to roll for HP. If we look at total punches landed, it’s closer to 20, which is a 33% hit rate – a saner low-level fighter striking AC 19. Still, you can see that fighters, even low-level ones, are a cut above average.

You can also see, that at least in this sporting competition, that most of the time in a three-minute round is not spent throwing blows. It’s movement, feints, footwork, and recovering breath. The frantic pace of GURPS combat (and the unlikelihood of actually gaming out a single boxing round, much less ten to fifteen of them) either begs for a lot of chances to spend time “doing nothing,” or should encourage other things to do with that time. The typical boxing match seems to average a thrown punch every three seconds (about 50-60) with only a third actually landing (in the neighborhood of 20). So in a six-second turn, fighters will be moving, evaluating, gauging distance, and throwing roughly two attacks. Given three such turns, one might expect two hits. At least for boxing.

On the flip side, most people will think of a die roll as throwing a single punch, or shooting an arrow, or firing a gun. Given that many games ask players to keep track of individual bullets or arrows (Fate and Night’s Black Agents do not do this, which shows excellent consistency in game design philosophy), there’s only so much abstraction that can be subsumed.

The six-second turn allows time to flow by at a reasonable level, while still assuming that a few attacks per round are possible. It’s been shown by trick-shooters that firing an arrow every second is possible with a low-weight bow, and for the cinematic world of RPGs, assuming it’s doable with a 100-200-lb. draw warbow isn’t going to be the most unrealistic thing going on at any moment on a battlemap. The expected rate of fire for an English Longbowman seems to have been one arrow every six to ten seconds – when bending a bow that’s 150-200 pounds draw, it pays to take time to do it so as not to exhaust the shooter. That number varies, though, depending on the source. So with trick shooting, six attacks per turn might be conceivable, but sustained fire with a high-power warbow would realistically take place at between one and two turns per attack. And crossbows? Fuggeddaboudit.

One does have to be a bit careful: in a modern game going full-auto on an M16 (with a cyclic rate of 75-90 shots fired per 6-second round) will empty a 30-round magazine in less than half of the usual time for a turn. This is offset by the fact that it’s possible to swap a magazine in a scant few seconds (though it will often take longer). This means an expert can blaze away (though, of course, an expert will not blaze away) for a half-turn, do a speed-change of a magazine, and repeat until part of the rifle fails from the heat.

My own experience GMing GURPS showed that the five-second turn was ideal for group movement. It allowed a cautious walking advance (1-2 yards per second) for a few seconds, interspersed with Perception rolls (Evaluate or Concentrate maneuvers in GURPS terms) and ending with a Wait, so that if a threat appeared you were ready. In practice, this meant a full move (which is usually 4-5 yards per second, or 8-10mph) spread over three seconds, one of Ready, Aim, or Evaluate, ending with a Wait. It made for plausible rates of advance, and breaking the pattern to run across terrain was scary, since it denied the Evaluate and Wait parts of the turn, it left one open to attack. That meant when people would give up their paced movement for all-out sprinting, they’d move from cover to cover. Hey, real-world tactics as emergent behavior? Bring it on. I would, however, break back down into single-second time frame when the bullets started flying. This might or might not be realistic, but it was necessary.

Any time scale can work for a game. It depends on what is happening, and how the GM wants to describe the action, and what the players are content with. The important thing is that all parties understand what can happen between (as Lorien might have said) “tick and tock.” I very much remember a D&D 2nd Edition game – my first and only such – where my archer got caught between tick and tock, with a ready bow, and a thug popped onto the battlemat, moved his allowance, and hit and killed my character with him just standing there. In retrospect, this probably means I managed to not see a guy until he was within one move. That underscores the point, however. An archer can be set upon and killed (potentially) so long as his foes are inside of 10 yards. Want to ensure you can’t get hit? Stay farther away, because you will not get a chance to react

That same situation in GURPS? Assuming that our mercenary started 10 yards away, he’s moving with a typical speed of about Move 3 or Move 4. So the archer might see him on turn 1, Aim on turn 2, and shoot once right as the fighter gets within hitting distance. Perhaps this would be an Attack of Opportunity in D&D, upon closing to within 30′ of an archer with a ready bow.

The most convenient time increment for combat seems to be something like 3-6 seconds, with 5 or 6 being a very good compromise. That it happens to make ten turns to the round?

Not a bad side effect.

Still, one of the best options would seem to have the ability to telescope time as required. Fate and Night’s Black Agents can do this, under GM control. Other games might have to look at choosing discrete scales – a geometric progression that looked at 3 seconds, 10 seconds, 30 seconds, and roughly two minutes with smooth rules for moving between scales would probably be fun to play. It would provide a good “pick from a list” set of options for each time scale, and players and GMs alike could set expectations accordingly.

Because time after time, expectations mismatch is a great way to kill a game, never mind killing characters.