Saturday 20 January 2018

Game Design #76: Uncertainty in Activation

From the start the initiative and activation mechanics have fascinated me. Even as a teenager, I always wondered why games like Warhammer had simpler/less interactive activation than games like checkers.  Most rules were clear on the how - how to shoot, move, melee etc - but tended to cursorily skip over the when.  And the when is just as important. When someone shoots matters a lot to the target; if his head in below the trench at the time, or above it....

This topic is in my head at the moment as I experiment with aerial/space dogfight games.  More precisely, the certainty in activation. How predictable is your turn? How far can you (and should you) plan ahead?

Let's skip through a few common activation types, to orientate ourselves. Blog regulars will probably skip this next bit and go to the next main heading.

Common Activation Systems
This has a high level of certainty. You get to move, shoot etc - resolve all actions with all your units - without the opponent being able to intervene. Your opponents stand around like dummies as you can act with everyone, without any interference.  The upside: it is simple to track whose turn it is. And you can go off for a toilet break in your opponent's (usually long) turn.  Popularized by 40K, this was once the default activation, but is less common now.

Alternate Move (aka similar to Chess-style move)
This has less certainty then IGO-UGO.  Each player takes turn moving a single unit each, until each side has activated all its units. This allows more interference and organic "reaction" to your opponent; he moves and makes all the actions with a single piece, then you respond. It's still pretty predictable as you know who is going next, and you get to choose which (of the so far un-activated) units go next. Probably the most popular nowadays. (It's not exactly like chess, as chess allows you to move any single piece as long as it's your turn; alt-move requires you to act with every piece in your army before you can act with a unit - aka piece - a second time),

Group Move 
This is a mix of the above two styles; players take turns moving clusters of units (i.e. 3-5 bases in a platoon, or squad, or even members of a fire team).  One player moves a group of 3-4 models, then another player activates with a group. Usually each group makes all its actions, then the next player acts with his cluster or group of units. Unsurprisingly, this method is more predictable than alt-move, but less so than IGO-UGO.

Reaction Moves
Some of the above methods have some sort of baked-in method of reacting to opponents;  often forfeiting their turn to fire later (overwatch). However some games have strong reaction mechanics where units can react quite often (sometimes an unlimited amount of times) to enemy actions, giving a huge scope for interrupting enemy turns  and making them less predictable.

Momentum (chaining activations)
I've explored this topic in more depth, but basically it is alt-move, but offers the player to "chain" or "follow on" by acting with a second unit without allowing the opponent his action.  A bit like player A moves a unit, but he does not want Player B to have a turn with one of his, so he (in some way, maybe by passing some sort of leadership roll) "follows on" and Player A then acts with a second unit in a row, without letting Player B have his turn. A bit like in, say, chess, if the white player moved a piece, then move another piece without allowing black to move his - basically, messing with the "normal" move sequence.

Okay, now we know some common systems. Let's talk about sequencing.

Predictability in Sequencing
What I mean by sequencing, is the order units activate. Is the order units activate dictated by the player, or by chance?  How far can you plan ahead?

Games like chess have a predictable sequence - you freely choose which unit to activate. A game like Savage Worlds where each unit is assigned to a card and acts when its card is drawn is random. You have to act when your card comes up; you don't get to plan ahead for the best time to activate the unit, you have to do the best you can if and when your card turns up. Whereas most games, you can choose which unit acts next. Some game devs like Too Fat Lardies allow semi-random sequences; they roll a dice and you can choose between certain (but not all) units which meet the dice roll criteria.

This topic has been one I've considered a bit lately, due to many aerial games making initiative/move sequence in order of best pilot->worst pilot or vice versa.  This is a quick, easy way to emphasize good pilots, but makes dogfights quite predictable - you know the move order ahead of time, allowing Chess-like "thinking ahead" and squadron-wide teamwork in a way that is nothing like the chaos of a real dogfight, facilitating cheesey tactics like co-ordinating and focussing down rookies first, so you can outnumber the enemy elites.

Enemies (from the non-active player) can also interrupt the sequence with reactions - in many cases reactions are limited (i.e. you trade entire move for the ability to fire in enemy turn aka "overwatch") but sometimes they are very strong - in Infinity reactions are unlimited; each time a unit activates in line of sight, every opponent can fire at him. They can do this every time, every opponent acts (unless prevented by being suppressed etc). 

Sometimes units can seize the momentum by becoming the active player; perhaps if an active player is hit, the other side becomes the active side. This really messes with the "expected" sequence - we've already discussed this under "chaining" and "following on."  Allowing the status of the active player to switch at least semi-unpredictably between players within a turn removes chess-like planning ahead and forces players to "live in the moment."

Predictability in Actions per Turn
I've explored this before; it's how many actions a unit can make when it is their turn. Most units, in most games, can move and shoot; that is two actions per turn, or 2APT as I will abbreviate it. If a unit is guaranteed to be able to move and shoot each turn, it is very predictable. But in some games, actions per turn vary quite a bit.

In the post linked in the previous paragraph, I posited that too many actions-per-turn is bad; in making the active player too strong, and allowing them to do too much in their turn. It "freezes" everyone else too long.  It's like watching the Flash; he zips around and does all this stuff while they are frozen. But in a wargame, we want to reduce the "Flash" effect and make it seem more fluid. In contrast, a model that can take one action per turn (move or shoot) is more limited. We've broken the turn up into smaller chunks. 

However, sometimes we can allow a few actions per turn, but make them unpredictable.  Song of Blades and Heroes does this by allowing players to roll between 1 and 3 dice. Each dice that beats a set score (typically 3+ for elites, 4+ for average, and 5+ for rookies) allows an action.   There's a twist as well; two failures means the turn ends, and the other player player gains the momentum, becoming the active player until he too fails in his turn. That said, actions per turn are semi-predictable, but not completely so. A rookie is unlikely to roll successfully and get all 3 actions, and is likely to get only one. A hero might average 2 or 3.

Besides my dogfight homebrew rules, another reason for this topic is Inquisitor. I found an secondhand rules set from 2001.  It's a RPG/narrative wargame, with d100s and dashes of Necromunda. But what caught my attention was the activation system.  Players activate in order of Speed (this is quickness of movement and thought); i.e. Speed 4 would move before Speed 3 or 2. So quite predictable. However, each Speed allows you to roll a d6. (I.e. a Speed 4 character rolls 4 d6). If you get a 4+ you get to perform an action. I found it interesting in that actions had to be declared in advance (I've done this in my aeronef rules) and if you run out of actions... just don't get to do all you said you would.  It makes the amount of actions unpredictable. I.e. you might say "my gunner will move to cover here, crouch down, then shoot."  The Speed 4 gunner only rolls, 3, 6, 5, 1 - so he only moves, and crouches - but does not have a third action to fire.    There is a predictable move sequence - i.e. fastest Speed to slowest - but an unpredictable amount of actions per turn.   I think these rules neatly illustrate the difference between the two.

So - we have uncertainty in activation, uncertainty in actions per turn, and the ability for opponents to interrupt.  All these add unpredictability.  So which is best? More predictabilty, or less... 

Is war predictable? Plans always survive contact with the enemy, right? The ability to react to the unpredictable and impose order/execute plans within the chaos of war is a hallmark of good commanding.  For example, allowing perfect co ordination in the swirl of a WW1 biplane dogfight seems silly. A more random method such as a random card draw seems indicated.  With radios in modern aircraft, co-ordination with wingmen seems likely. Perhaps the ability to "follow on" to a wingman or make a group move would work better here.

However, some games need predictability.  Take Warmachine. It's a game focussed on super-OP-combos which often need to be "chained." It's more CCG than wargame. Using IGOUGO makes perfect sense - or you'd find it impossible to execute any combos or sequence attacks. 

In summary, it's not as simple as unpredictability good, predictability bad - the amount of predictability can (and should) vary depending on the genre and era. Even how this unpredictability is introduced - be it from unpredictable sequences or amount of actions, or reaction mechanics. 

I think the key takeaway is determining where your game fits on the spectrum and picking the right activation to suit.


  1. Advanced song of blades and heroes now had a reaction system if you roll 1 failure when rolling for actions your opponent gets a chance to roll 1 reaction

  2. "...the when is just as important."

    - Agreed. I see this as a factor in not only combat/firing, but also variable movement systems. Most such argue command control, troop variability, and (mostly) terrain in explaining the results of, say, rolling a die for movement.

    To me, the biggest factor is timing, ie: How far you move before your turn ends indicates how far you got before someone else (either friend or foe) gets a chance to react to your action.

    1. I considered touching on variable movement systems, but decided it's probably it's own topic.

      While I understand WHY people do variable movement, my background as a PE teacher screams against it, so I am unnaturally biased.

      I KNOW people have vastly different speeds and endurances (hence why the --everyone moves 6"-- has never sat well with me; likewise real life speeds tend to be consistent - a 1d6 move makes me grit my teeth as it is soooo variable; it FEELS wrong - like you're saying someone who covers 60m in 10 seconds on day, will only cover 10m in 10sec the next.

      I KNOW that's not what it's doing - variable moves are more about flexible time/reactions as you say - but it just FEELS and LOOKS wrong to me on the tabletop. And I doubt I'll ever get over it, even though I see the benefits for it in terms of game design.... :-)

  3. You keep mentioning "chess like" thinking not matching real dogfights. I disagree. When you read the accounts of WWI and WWII aces, they did have a chess like thinking to the situation. They would think if I do X the most likely response will be Y, and if they do Y i will position myself at W to shoot at them again. However, the aces could do this very quickly and almost intuitively. Most of us gamers can not.

    Sorry for typos as I am posting on a mobile.

    1. When I see the term 'chess like' applied, I think:
      Aircraft are not chess pieces with sure, predictable moves, nor are pilots capable of predictable performance in executing whatever plan they may have.

      The trouble I have with plotting/planning games is that it is, as pointed out, a very poor representation of the reality being modelled - which is why I dump the whole thing altogether.

    2. I have a feeling we are all coming at this topic all from different angles. Kinda talking about different things.

      Eric - I think you are referring "chesslike" as to the pilot thinking ahead. Fine if the player ONLY is the pilot of a single model. I agree if this is the case - aces CAN plan ahead.

      I am using "chess like" in a different way. I am referring to the wargamer who controls perhaps 4-6 aircraft, being able to co ordinate their actions.
      The "wargamer-as-flight-controller" is what I am referring to - not "wargamer-as-single-pilot" - which I think is how you mean it.

      Once the dogfight begins, a WW1 dogfight would have limited co ordination and strategizing between pilots. That's what I am referring to - the sequence of activation from one pilot to another, and how predictable it is.

      For example, if rookies go first and ace pilots always move last; the wargamer-as-flight-commander can predict the EXACT sequence of moves for ALL planes involved. Which I think is unrealistic and allows you to "game the system."

    3. Yes, I see your point. I probably am having a hard time with this as I enjoy many aircraft games and am therefore wedded to some of their conceits. :)

    4. If you've progressed from 1970s rule sets, then "modern" air games probably seem great in comparison. If you compare them to the innovation seen in, say, skirmish rules - they don't measure up. I haven't "grown up" with the genre, so I tend to recoil at what perhaps is taken for granted.

      A bit like age of sail - I find the vast majority of these rules hideous and clunky, but probably most existing AoS players are perfectly content with 30 minute turns and tracking 100s of hitboxes, crew points, etc.

  4. MacavityandMycroft23 January 2018 at 17:52

    Okay.... but that Inquistor speed thing. If you divorced the quickness of movement and the quickness of thought, does that not sound a lot like machine vs pilot?

    One stat for how much the machine (plant, tank, Mech) CAN do, and then vary the success role on how competent the pilot is (or give better pilots a re-roll per turn or something)....

    I like it!

    1. I use a range of versions of this in various rules; many have set amount of rolls, with the success die varying (i.e. 3-4 rolls, better pilots get action on 3+, average 4+, bad 5+ - basically SoBH style); the other, I have units able to chain reactions until they fail (again, better pilots have better success %, so they are more likely to chain a few actions before failing)

      What you are saying is basically Inquisitor, but varying the success % of the dice; i.e. Mech might be SPEED 5, but human pilot may roll an action on 3+, 4+ or 5+ depending on skill level?

      Hmmm I'll have to try it.

    2. MacavityandMycroft24 January 2018 at 07:39

      Precisely! I often am disappointed by the attempts to replicate experience or veteran savvy in games, this sounds like it could work.

  5. It's funny you should post about this, because I had been thinking about the exact same thing with my project Warstack.

    Long story short, I implemented a Magic-like "stack" for resolving Interrupts, with Interrupts triggered on intent to attack. Each unit could take a max of 2 actions/turn: one or both in an activation, or only 1 via interrupt.

    Rather than randomize activation order (be it like retaining Initiative in Epic, or die draw in Bolt Action), a player gets an amount of Strat Points per turn (ala focus) which are used for chain activations or interrupting interrupts (both with incremental costs) and a few other bits related to activation order. However, the amount of actions is certain, reactions do not require prep but are less economic, and the game feels less like "two players taking 20-minute dickpunches" so to speak.

    1. I'll have to steal this
      ......the game feels less like "two players taking 20-minute dickpunches" so to speak.....
      I'm trying to figure out a way to insert it into my next rules review.

      Your method sounds interesting and I get what you're trying to do; but I'm not sure I could replicate it to try it myself - would you mind posting it up in the google group? (right side of tab) or linking to your blog if you've got demo games?

  6. Since the google doc is still a rapidly changing wip, I think I may create a blog for this after all.

    The game is generally "one unit attacks one target", unless you clip other units with melee/aoe, ala 3rd-7th 40k. In exchange, you can place "target markers" next to individual units.

    So if Player A has infantry squad 1 attack Player B's infantry squad 2, A would place a 0 next to squad 2, a 1 next to squad 1.
    If player B wanted to preempt this, player B would place a 2 next to a unit. This doesn't necessarily have to be infantry squad 2.
    Should B wish to use that interrupt to attack squad 1, player A would spend a SP (due to already having a unit acting on the stack), and place a 3 down next to a unit (a unit can only act once on the stack). And so on so forth.

    Once a player declines to interrupt or performs a non-combat action, the interrupts resolve ...3 attacks 2, 2 attacks 1, 1 attacks zero. Assuming 2 and 1 survive of course.

    If after an Interrupt, a unit cannot attack its original target (ex: Player A has rifleteam 1 target player B's rifleteam 2, player B interrupts by driving a tank in the middle to block LOS), player A must redirect that attack against the tank. This doesn't alter who actually is on the stack.

    However, when its time to actually resolve, a player may spend a SP to "Fake Out" (have a unit not on the stack carry out the attack instead), or to "Stand Down" (not actually carry out an action).

    These eat into the 2-action limit as though they were activations. There is another SP cost to activate/interrupt with a unit that has already taken a single action.

  7. Fascinating discussion, and an excellent set of examples.
    In my opinion, interleaving and interrupt become more important as the speed of fighting units increase, or their numbers decrease.

    In other words:
    Fighters, spaceships, Mad Max cars and speedboats require some kind of mechanism to prevent "Drive by syndrome".
    Skirmishes need something because each figure might be an active personality.

    Contrast with an old-school mass combat wargame with a Macedonian pike-block lumbering toward a bunch of Persian foot.
    There's a lot more of "Just get on with your move so we can fight the melee".

    Of course the horse archer was the drive-by specialist of the ancient world.
    Old school gamers rarely had problems with them being abused; provided the designer selecetd suitable movement and shootiong distances.

  8. --"In my opinion, interleaving and interrupt become more important as the speed of fighting units increase, or their numbers decrease."

    Interesting you should point this out, I've personally been removing many reaction mechanics from my fantasy/medieval rules, for this very reason. As this genre is melee centric, reactions can promote shooting too much, and is unecesscary.

  9. I just re-read this post - so much great food for thought!