Game Design Theory: A New Benchmark For Modern Skirmish Games


Way of the GunThanks for all the film suggestions folks.

As you might expect, I’ve already got a number of them, though that doesn’t mean I’ve watched them recently. Even those I’ve seen are well worth re-watching.

One comment that I thought was worth discussing more was about Collateral. This Michael Mann directed Tom Cruise/Jamie Foxx movie has a number of scenes containing firefights and is well worth watching for action/thriller movie fans. Tom Cruise with very blond hair always looks a bit odd, but that’s possibly just me.

Warning: the rest of this post contains spoilers. 

Perhaps the most interesting scene, from the viewpoint of modern skirmish games, is the shootout in the nightclub.

To recap on some basics: there’s a busy nightclub with hundreds of people in, dancing away. At the back is a Bad Guy, surrounded by his many bodyguards. So far, so busy. It’s loud, crowded, and full of armed guys intermixed with largely unaware civilians.

Into this comes Tom Cruise, Bad Guy number 2, with his hostage Jamie Foxx. Tom is set on killing Bad Guy number 1.

Pursuing Tom are the LAPD and FBI. More guys with guns. They also think that Jamie is Tom. At least, some of them do. It’s complicated.

So, you’ve got multiple factions, all with different agendas and mostly with lots of weapons in a very crowded and confusing space. How would your favourite skirmish game cope? Find some popcorn, watch the scene, and have a ponder.

There are several major challenges in modelling this fight. One of these is the noise and lighting that makes target identification very difficult. Here you have some factions that will happily shoot anyone in their way, and others that will try to avoid causing injuries to civilians.

Your rules have to cover differing rules of engagement, and accurate target identification is a part of making that work. Don’t want to shoot the wrong people. And, even if you don’t care who you kill, you really want to take down the guys that are trying to kill you before they return the favour. Got to pick the real threats out of the crowd, and do it quickly.

Non-combatants are a huge factor in this scene that most games don’t do well. Here we have way more panicking and unpredictable civilians than we do combatants (though we still have plenty of them). Although it doesn’t happen in this example, we frequently read about off-duty cops, servicemen on leave, and all manner of brave individuals who’ll have a go at the shooter when things like this happen in real life. Can our combatants trust the frightened civilians to conveniently run away? Not entirely.

Even when all the crowd wants to do is flee, they are dodging in different directions, running from multiple shooters in several directions; escaping through exits that cannot cope with the movement of that many people that quickly. There are bottlenecks and panic. people fall over. Are they shot or did they merely stumble? Does it matter? Should you be helping them? Back to objectives.

Do we need a referee to control these crowds, or just some AI system?

In the middle of this mess the factions are all trying to do their own thing. Need to have robust mission and objective rules to cope with that. SOP is important as well as individual mission specifics.

Factoring confusion into games is always difficult as we naturally start with perfect information. This is the opposite of reality where information is only dragged reluctantly out of a situation. On the tabletop we can typically see exactly where every participant stands, and know exactly who everyone is. We know the odds of each shot before we take it, and can foresee every modifier. None of that is available to the real individual on the ground in a real firefight.

The more I think about it, the more I think that recreating this fight should be a textbook test case for every skirmish game that purports to be able to model modern firefights well. Admittedly it is somewhat far-fetched in its complexity, but if you read AARs of real actions it’s not as far off as you might think. The confusion is certainly well done.

I understand that some players won’t want to include non-combatants in their games, and that’s fine. Much simpler that way (both mechanically and morally). But I like to be able to model as much of “anything” as I can, and this would seem like a good way to open up all sorts of interesting games I couldn’t otherwise do. Reality is messy, and combat doubly so. Games have to do a certain amount of putting things in neat categories whereas reality is always fuzzy at the edges, and it’s squaring this circle that makes game design such an interesting challenge.

For me, this nightclub shootout is a new benchmark against which I will be testing Eternal Battle, to see if I can replicate it well on the tabletop. Very challenging to include all these factors and still be slick and fun to play, but that’s been EB all over 🙂

Can you think of a system that could replicate this scene already?


This entry was posted in Eternal Battle, Game Design Theory and tagged , . Bookmark the permalink.

32 Responses to Game Design Theory: A New Benchmark For Modern Skirmish Games

  1. James K says:

    That is a really interesting scenario.

    • Quirkworthy says:

      Certainly is. And you can play about with the number of players too. How many of the (many) factions are player controlled and how many referee or AI? That would make things even more varied.

      Plus, if you did a solo version you could play as any of the factions, that would be pretty cool. I’d expect it to be rather different for each side.

  2. Chris says:

    I think only RPGs routinely have a crack at such a scene and then by cheating (the GM gets to handle it without a system). Hell it is hard enough to get civilian models. I am always on the lookout for some because I have never liked fighting over deserted towns and cities!

    • Quirkworthy says:

      After I posted this I started thinking some more about civilian models. Not only are they rare in the first place, but because individuals dress differently for different environments, you’d need huge numbers of them to accurately reflect the scene. For example, the crowd in the nightclub would mostly dress differently for work. So if your shootout was in a mall, then you’d need different civilian models, even if they represented the same people. At least militaries wear uniforms (and zombies have an excuse) 🙂

      And you’re right about RPGs. I’ve done similar things to this in several RPGs over the years. IN fact, the barroom brawl was a staple for many years, both in fantasy and SF settings. The technical challenge is getting it to work without a GM on hand to paper over all the cracks.

  3. Arthur Monteath-Carr says:

    Well, for starters, I’d probably run this scene using action-movie RPG Feng Shui, but that’s probably not the answer you’re looking for. 😀

    In a wargame, I think I’d want to abstract the crowd behaviour into an environmental hazard. Models have a reduced movement as they push through the crowds; certain actions will increase the ‘panic’ level by a greater or lesser amount (Tom’s HTH take-downs wouldn’t increase the Panic level, a single shot would cause it to go up a certain amount, and a full-auto burst would max it out straight away). Panic goes up, but also ebbs a bit at the end of the round.

    When you are taking a model’s turn, you determine line of sight, then refer to the panic level to see what the odds are of hitting your target, or hitting a bystander. Any model in any faction can choose to take the shot regardless of what the odds are.

    The effect of Panic is twofold. At a low level, it gives you a straight -1 to ranged attack rolls, and a low chance of hitting a bystander. You can also attempt to hide in the crowd. At medium levels, you suffer a -2 to hit, and a higher chance of hitting a bystander; you stand out more in the crowd though, so you can’t blend in the same. At maximum panic the bystanders are actively attempting to flee the scene. For the next 2 turns everyone is at -3 to hit and a missed shot *will* kill a bystander. However, once that is over, the area is clear and play can continue as normal.

    Depending on which faction you’re playing, the panic levels and dead bystanders mean different things to you.

    If you’re the assassin, you want to keep the panic low for as long as possible. If you make your ‘blend in’ roll, then anyone in the LEO faction won’t be able to target you at all, and anyone in the Gangster faction will have a very high chance of killing a bystander if they elect to try and shoot you.

    If you’re the cops, you want to minimise civilian casualties – dead civvies earn you negative VP, and if an LEO kills a civvie, then that’s double negative VP. They want to either close to the Assassin to take him down in HTH, or, if they manage to raise the panic level to get the club to vacate (maybe by a special action (waving a badge, firing into the air, pulling a fire alarm) then after the chaos the battlefield is clear.

    If you’re the Gangsters, though, you can also try to ‘blend in’, but you don’t earn negative VP for dead civilians, so you don’t care about them *as* much as the cops, but a low panic level has it’s advantages to. It’s a choice, really, to try and hit the “panic” level at the opportune time.

    I guess the alternative to abstracting it would be to have counters or models representing the civilians, but the prospect of rolling d3″ scatter dice for 20+ counters doesn’t really appeal to me.

    • Quirkworthy says:

      Thanks Arthur. An excellent starting point. I agree that treating the crowd as an abstract environmental hazard is a good way to go. Not sure that’s the whole story, but it’s certainly a heck of a lot easier than tracking each of them individually 🙂

  4. Ian says:

    This scene isn’t far removed from a pulp scenario I ran a few years back where 2 opposing groups were hunting for a missing professor in a crowded London train station.

    This was achieved at the time using a set of rules called .45 adventure, these work best as a semi-narrative game with a GM and a bit of upfront work. I suspect you could achieve the same kind of scenario with my current favourite pulp rules which are Pulp Alley which require essentially no GMing and just a little scenario building to cover some of the random effects.

    Both these rules are very open and suit pulp gaming very well where you want to be able to adapt the rules and scenarios to basically create film scenes on the tabletop, we’ve done everthing from Indiana Jones temple runs, scenes from the Mummy, Car Chases, 1930s film noir detective games using these rulesets.

    Quite a few of the games are on my blog:
    and on my brothers:

  5. edenchanges says:

    Delighted that my suggestion has hit the target! (No pun intended – okay maybe a little). Your thoughts about modelling it are very interesting. I will have a good think about this when not at work! Cheers, Stephen

  6. WhelpSlayer says:

    There’s a very good vs board game called Kings and Assassins where one player commands a group of knights guarding said king and the other 3 assassins. What makes the game stand out is that this takes place in a crowded market place and the assassin is hidden as one of the townsfolk. During their turn both the king player and the assassin player move villagers around and the idea for the assassin is to kill off the kings guard, kill the king, or just delay him long enough.

    A similar principal could be used here. Your combattants are hidden by standard crowd markers and can use their counters to check others. This doesn’t give away your position entirely but your opponent will have an idea of where one of your guys is and where they may be. Of course, once the fight breaks out the tokens will start making for the exit with maybe some “off duty cops” or “undercover agents” could make their appearance.

    It’s an interesting concept and I like the principal of theorycrafting some ideas for it!

  7. In game design terms, I don’t think this is necessarily a good test, because you’d be wanting to build in all sorts of mechanics about things like random movement or an NPC AI. Very few people will want to play such a game regularly, or have a miniatures collection or terrain set that would support it.

    Having said that, there are a few skirmish games that could be modified to some extent to do this. Infinity already has rules for camo markers, which are very like unidentified potential hostiles. You have to Discover them before you can engage them. If everyone in the scenario began with the Limited Camouflage skill and each player had a number of civilian counters within their control, you could play some great head games. Infinity also has a range of different visors that make it easier to discover concealed troopers, something like wearing NVGs in a dark nightclub…

    • Quirkworthy says:

      I think it depends on what you want your game to do. If you expect it to ever need this sort of capability (and with EB, I do) then it makes sense to build it in at ground level.

  8. Guillaume Gentile says:

    Do you think the crowd can be a terrain?
    We always say that the position of the minis is an approximation of their true positions. Can it be the case here? do we really need each member of the crowd to be represented by a mini?
    In fact when the crowd moves, it’s like the sea on a beach.
    In that case, the crowd can move in response to actions made by the factions, closing and opening like the sea…

    • Quirkworthy says:

      That’s similar to what Arthur suggested, above. I think you need to be abstract in some way when you’re dealing with crowds rather than just one or two bystanders.

  9. Renzo says:

    I think maybe you could try to simplify it using the crowd as area terrain and to shoot through it you do it the standard way for your game, but add a penalty dice. This dice can be a negative bonus to your shoot or can define a threshold from which you kill a civilian instead of the target. The panic of the crowd would modify the number of penalty dice.
    That way, you can guess your chances to hit your target, but can’t be sure, because of the chaos of the place.
    Just my two cents.

  10. Smakkit says:

    I have always wanted to use a crowd as “forest” terrain. Cant shoot through it and difficult to shoot into.

  11. Jimmy To says:

    for something like 40k killteam it could be something like slow and purposeful for going through the crowd when tom was using the guy as a shield maybe that could count as a feel no pain save

  12. Pingback: Game Design Theory: Benchmarks for Other Skirmish Games? |

  13. vaultage says:

    spectre and black ops rules have something with civilians rules as well as hidden opponents models among the civilians.
    tomorrows wars has also interesting asymetric rules that could be transposed to this.
    but you d need to look deeper in paper rpgs to get something that approach gun scenes including lots of civilians.

    • Quirkworthy says:

      As you say, RPGs can do this sort of thing well. One of the things I’m trying to do with Eternal Battle is to bring some of that aspect to skirmish games. Which is largely how I got to this point 🙂

  14. Dan says:

    My initial thought is that the Chain Reaction series from Two Hour Wargames might be able to handle this. Chain Reaction itself covers modern combat, but they also have rules for other settings. I’ve only played their games a half-dozen times and it was many years ago, so I don’t have full confidence in the suggestion. Many AARs on their website do have civilian NPCs in them, but I’m not sure how active a roll they played in the scenario. Chain Reaction leans heavily on AI-driven gameplay. Most of their games can be played solo and will force reaction behavior on models when they are attacked, regardless of who is controlling the model. Diving for cover, shooting back, dropping to the ground are common reactions.

  15. Alex says:

    I’d say Infinity could quite easily handle this scenario.
    I’d effectively be designating the crowd as “terrain” and applying terrain rules for them. This abstracts them somewhat, but I feel it’s better than keeping track of 200+ individual NPCs.
    Start off by designating the crowded room as difficult terrain and a saturation zone. This slows movement and reduces the effectiveness of shooting.
    Second, designate the bar, DJ booth and a few other “well lit” areas as low visibility zones (-3 to BS).
    Third, designate all other areas as poor visibility zones (-6 to BS).
    Fourth select a number of ares as white noise zones (zero visibility to any model with a multispectral visor, no additional effect for anyone else) to represent flashing lights, lasers etc from the club ceiling.
    This sets up the “terrain”
    Play the game over three rounds:
    At the end of the first round, reduce the poor visibility zone to low vis and the low vis zone to normal.
    At the end of the second round, remove the saturation zone.
    You could run the police as NPCs or as a faction, if they’re NPCs I’d have a couple of officers turn up at each door at the end of the player turn where someone shoots for the first time, they’d take a reaction shot at anyone who activates in their LOF.
    Civilian casualties would be accrued through the failure category rule; any shooting attack that misses due to the poor/low vis zone’s modifier hits a civilian (eg, BS 12, -3 for low vis will hit your target on a 9 or lower, and will hit a civilian on a 10-12, missing everything on a 13+). Each faction would have a threshold for civilian casualties built into their win condition.
    Tom Cruise would be either limited camoflage (moves as a marker, can’t be shot until he is “discovered” or shoots) or holoprojector 2 (3 copies of him move, 2 are fakes, only 1 reaction is generated when all three move and your opponent guesses which to shoot at). He’d also have a sync’d civilian with him that he wants to keep alive till after he’s killed the target, then escape leaving the civilian behind, dead.
    Bad guys want to kill Tom Cruise and/or his sync’d civilian and escape without killing any FBI and with overall civilian casualties below (arbritary number here…) 20. 5 gang members, impetuous and irregular, 1 Gang boss, impetuous and regular, designate all fire doors as exit points.
    Tom Cruise wants to kill the gang boss, succeed in a WIP roll next to his corpse to “deposit” his sync’d civilian (fail and the civ runs off, dispersion roll, 4″ move, and needs to be re-sync’d, WIP roll, then brought back to the corpse and WIP’d to stay in place) and then exit through a door while civillian casualties stay below 10(?). He is regular with V:NWI or V:Dogged but recieves 5 orders per turn.
    FBI have 3 duo fireteams of two agents each (regular), some could have multispectral visors, they want to kill the gang members scoring a coup de grace on as many as possible (slapping the cuffs on him) as well as on Tom Cruise.
    Good scoring would be the issue, getting everyone’s win conditions to be fair.
    Gang: score 1 OP per member that escapes, 2 OP for killing Tom cruise, 1 OP if the civilian casualty level stays below 20 and 1 OP if no FBI were killed. (score out of 10)
    Tom Cruise: Score 2 OP if the gang boss is killed, 2 OP if no gang members escape, 2 OP if he successfully deposits his civilian, 2 OP if the civilian casulaties remain below 10 and 2 OP if he escapes. (score out of 10)
    FBI: 1 OP for each gang member “arrested”, 2 OP for Tom Cruise getting arrested, 1 OP if noone escapes, 1 OP if the civilian casulaties remain below 10. (scored out of 10)

    Tom Cruise would get the first player turn each round, with the Gang and FBI rolling off WIP at the start to see who gets initiative (goes after Tom Cruise, before the other faction each game round) or deployment (forces the other to deploy first).

  16. elgamers says:

    Just watched the first of the collated trailers for Blizzard’s Overwatch and thought of your post:

    … or should this be in the ‘what should I watch’ Comments? You decide. Either way I thought it was an excellent skirmish, albeit light on actual non-combatants. And a terrible Mary Poppins accent 😉

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s