One question I’ve been pondering as I go about designing a strategy game is that of clarity in game mechanics, and in combat systems in particular. Committed players of strategy games (and many other genres, for that matter) have long taken joy in pulling apart the math behind the combat resolution systems that drive the games they play, in part to seek an advantage or upper hand and in part out of simple fascination. This description of the mass queen experiment from the StarCraft2 Hacks blog is a great example. It runs through a bunch of math, based on things like minerals, gas, queens, hatcheries, food, game time, energy, “transfuses,” injections and more, and comes to the colorful but perhaps surprising conclusion that “Queens are better cost effective healers than Medivacs.” Food for thought.
What’s interesting to me is the question of where a combat system might lie on the spectrum from transparency to obfuscation (through complexity or by other means), how that stacks up against the combat results you as a designer want to produce, and whether a game is any the better or worse for your equations and calculations being deeply buried, or riding on the surface of play.
The Battle for Wesnoth, for instance, which is a really impressive (and free!) turn-based strategy game, is exceedingly clear about the math behind its combat mechanic. At any given moment, you know exactly what your chances are of winning any given encounter between two units, and you can discern, with hardly any poking around, why your chances are what they are (because Undead are highly resistant (60%) to most Physical attacks, for instance).
Others, like the Total War combat system analyzed in this 2006 forum post, take pages to explicate and are difficult to understand even with extensive documentation provided by the developers.
The system I’ve been working on lies somewhere between the two, though certainly closer to the transparent/Wesnoth end of the spectrum. If it survives in anything like its current form, it will be relatively easy to understand with perhaps a bit of calculation on the player’s part, though in many cases not so simple that you can do the math in your head. Note that the choices I’m making definitely do not grow out of a desire to make the system more or less transparent; they are driven by the kind of gameplay I want to produce, and the immediate and tangible questions it raises. E.g., how do you resolve combat between two stacks of units with various advantages and vulnerabilities, without requiring the player to make any additional decisions after sending his or her units into the fray? There are more answers to questions like this than at first meet the eye.
The questions I have about questions like those are these: What does it mean for a game to have a transparent mechanic, versus one that’s more obscure? Do players (those who are paying attention, at any rate) feel better knowing how they’ll fare before they go into battle? Does having a mechanic that’s too obscure to figure out (if there is such a thing) mean players will be less engaged? Does it take away some of the challenge if you know beforehand whether you’ll win or lose? Do transparent mechanics provide clearer paths through the content? Or do they lead players to reduce the number of alternatives they explore, since much of the exploration has already been done for them?
Or… do these questions apply to such a small portion of the player population that they’re not even worth considering? I’d argue that even if it is only a few who are so engaged as to do the math, this is still worth thinking about, for a host of reasons I won’t get into here. In any case, it’s interesting stuff to ponder. What do you think of it all?
I think clarity is a good thing up until the point at which you know the outcome of a fight before it happens – and it stops there. That could be cool for a game that’s pointedly about something else – I’ll check out Wesnoth – but if it’s about fighting, I want the judgement call about whether I can win this to be my own.
One of the biggest pleasures in strategy games for me is building a new unit, or perhaps just a rare one, and pitting it against stuff to see how easily it crushes it. However awesome I’m expecting it to be, I’m always delighted with the results. That might be lost if it had a little “Chance of success: 100%” thing every time I right clicked.
Tangentially, I love strategy games with a huge, clear delineation of power between different units. I love SupCom because when I build a Tech 2 unit, I can *see* it’s going to be more powerful than a whole bunch of Tech 1s, and when it fights them it’s gloriously apparent exactly how superior it is.
One of the reasons I don’t get that excited about playing StarCraft 2 is that there are very few big, clear, simple power differences. Except for one or two top and bottom-end units on each side, everything is roughly the same size and has roughly the same value. There are strong counters, but they’re not clear or intuitive, they have to be learned. Are Marauders good against Roaches? There’s probably a definite answer, but it’s not clear to me even after seeing both in action many, many times.
> “I love SupCom because when I build a Tech 2 unit, I can *see* it’s going to be more powerful than a whole bunch of Tech 1s.”
This is an excellent point. I had in mind only the numbers, but of course there are loads of other ways that a design conveys the information that X > Y. WoW gear as well is a great example.
Pingback: The Sunday Papers | Rock, Paper, Shotgun
“I think clarity is a good thing up until the point at which you know the outcome of a fight before it happens – and it stops there.”
I dont see a relevance here. In the strategy genre, the outcome of a war is rarely defined by a single battle. Strategy games are pretty much resource management. You can estimate the outcome of every single battle almost perfectly, but that doesnt help you win on its own in a well-designed strategy game, the game is not about that. For example, not wasting the powerful unit with cleaning up the small fry, when he can be used somewhere else to its full potential. Setting up attack queues so that the enemy will not have time to recover. Knowing when to go full out and when to hold back. I dont see anything that would detract from a game experience just because you know the outcome of every single battle.
Transparency in game mechanics — especially in RPGs, where it’s fast waning — is very important to me. But I’d say the main reason is one that you didn’t mention: it’s about the choices I make *before* the fight. I don’t so much care whether I know exactly what my odds of success in any given action are, but if the game is asking me to make a choice between two skills I could learn, or two stats I could upgrade, or two weapons I could buy, I really really want to know what difference they’ll actually make. If it’s just a matter of producing unit X or Y, you can try each and eventually at least get a good sense of the difference. But in a long RPG or hybrid-RTS campaign, an early skill choice can lock you in to an option the value of which you may not truly understand until hours later. Having to make such decisions on the basis of insufficient information is agonizing. Perhaps more importantly, it replaces the joy of direct mental engagement with the game with the grinding necessity of looking things up in forums or FAQs, which inevitably spoil far more than the bit you were looking for. And that last point holds for a lot of “pure” RTS and TBS games as well. The less transparent the system is, the more time the player has to spend trying to second-guess the designer’s intuitions, and that’s (usually) a distraction from the real fun of the game.
Coming in from the opposite angle, Dawn of War 2 is a game that’s often disliked by Starcraft fans (Personally I’m a big fan of it, but I won’t bore everyone with the reasons why here), and one of the key reasons is that it doesn’t natively spell out things like DPS and armour types.
At the same time though this is mitigated a lot by the nature of the gameplay itself, because in most situations it’s less about the actual DPS and more to do with the actual rules by which combat resolves itself, and the interplay between them. Which is probably why it’s more important that individual abilities are explained in terms of their functionality (a holofield cloaks a specific area for a time, explosive shot from a shotgun causes knockback etc.)
In DoW2 there’s very hard delineations between things like ranged and melee combat, and normal and vehicle armour, and that probably shapes most of it. Units that are good in ranged combat won’t be good in melee, and vice versa.
Combat is the majority of the time decided less by actual damage numbers, more by the situations that these numbers are being applied in. Melee units take damage coming into combat, but once in combat they tend to dominate ranged units because they can’t respond. If a suppression team keeps the from coming into range then that changes the situation completely, and so on.
Even there though, there’s a lot of obfuscation and a fair few edge cases, which as much of a fan of the game as I am, can be really problematic for new players to get to grips with (Sentinels for example are the earliest vehicle anyone can get in the game, but they’re classed as a form of “soft” vehicle armour that can be damaged by standard infantry weapons and concentrated fire). This results in the devs having to try and convey a lot of this non-obvious information via tooltips on the loading screens. Hardly ideal, I’m guessing the game could have used a bit more of a training mode for multiplayer.
Basically I think a fair amount of obfuscation of the raw stats can be totally fine as long as you’re working by obvious and spelled out mechanics, and clearly illustrate in other ways in which one unit may be superior to another (like bigger unit sizes, which IIRC there’s a post somewhere on the internet about the reasons that Blizzard didn’t go with super sized units in Starcraft 2). A game like DoW 2 for example, could use clearer visual markers for what marks “heavy” infantry armour compared to “normal”, because that can have a big impact in ranged standoffs.
Actually, one final thought, you’re talking about all these mechanics but I’m presuming this is all discussion for multiplayer. The rules can change a fair amount depending on the goals in a singleplayer game.
To follow up a little on what subedii said, I think there’s a difference between games which have simple mechanics but complex outcomes that are hard to predict, and games with complex mechanics that allow intuitively predictable outcomes. Starcraft fits into the former, in that the mechanics are fairly simple (apply damage to armour type until unit explodes, for the most part), but the outcomes can only really be learned from experience, or reading guides.
The Total War games and, say, Company of Heroes are more like the latter. The mechanics are mostly too complex to be worth the trouble of calculating, but the outcomes can be predicted in advance, for the most part. So, pikes beat cavalry, deployed MGs beat infantry squads but not vehicles, and flanking always gets better results.
Dawn of War 2 is an interesting one, in that it uses very similar mechanics to CoH, but in a fantastical setting, where the fighting abilities of the units are far less intuitively clear (what are Warp Spiders good against?). This one is most likely an example of complex mechanics, complex outcomes. Wesnoth, on the other hand, is a decent example of simple mechanics, simple outcomes.
> “an early skill choice can lock you in to an option the value of which you may not truly understand until hours later. Having to make such decisions on the basis of insufficient information is agonizing.”
This is a great point. Whatever the system and whatever the transparency, designers will generally need to convey enough information to allow players to make educated choices, especially where such choices limit your alternatives down the line. (I can imagine a game that didn’t convey that information, but it would play to unusual tastes, I’m sure.)
It’s interesting to consider just how educated players need to be, though. That’s one dial a designer has access to that can make for very different experiences from game to game, as described by the commenters here. Where the sweet spot is is probably more a matter of taste and flavor than anything else (provided you stay within certain bounds).
That’s a fairly relativist answer (to my own question), but maybe the more interesting point is *how* the mechanics and information are conveyed, as the last two commenters describe. It may be the numbers that are informing your strategic and tactical decisions, it may be the unit types, it may be the visuals — preferably all three (and more) in combination. But some games lean on one at the expense of others, and to good effect. The StarCraft – Total War – CoH – DoW – Wesnoth continuum is interesting to contemplate in that regard.
I don’t think there’s any shame in having complex mechanics as long as those mechanics work well. It’s just that you’d need to be firm on the fact that you’re not gunning for any sort of mass market appeal there, but a more hardcore base.
There’s always this stigma attached to the ideas of “manuals”, “training” and so forth in games, there’s often a commonly accepted wisdom that if the player can’t grasp the mechanics automatically then you’ve failed. But some of the most fun games I’ve played have required investment to learn the rules. Heck, the base rules for chess aren’t immediately intuitable just by looking at the board, but without its mechanics it would just be checkers.
The problem is that more complex and involved mechanics can make a game more fun and give it more depth when implemented properly, and that’s really important for RTS’s when you’re talking competitive play.
There was a good podcast on Three Moves Ahead recently where they were discussing Men of War, and they kind of touched on this subject. Because Men of War goes well beyond the mechanics of something like CoH, and tries to run itself as, in some ways, a battlefield simulation, where nothing is certain, and there are vast arrays of possibilities in how things play out because of that simulation aspect. It leads to a game that’s pretty hard to grasp (barring the occasional grognard, and a lot of that is admittedly due to the interface), but that’s intrinsically rewarding because it allows much greater scope for heroic events to occur that nobody was really predicting. And more importantly, even the more mundane sounding things can seem heroic because you were the one that pulled them off, on the fly, and against the odds.
There’s a lot that can be done in things like visual design (especially with sci-fi / fantasy games) and interface design to make even complex mechanics seem easily readable and understandable. But like you said there’s still a sweet spot there depending on who your audience is.
I think it’s fair to say that the fun in strategy games is ultimately not from learning the rules but from mastering them. The clearer the rules, the quicker players can get to grips and the bigger the focus on decision-making rather than information-recall.
This does not mean strategy games cannot be deep at the same time as being easy to learn, but I do think that most randomised or poorly explained elements tend to be to the detriment of the game as a whole, even if the end gameplay after learning these secret rules is incredibly rewarding. E.g. a ‘weapon type-vs-armour type’ matrix with too many combinations probably cannot be explained deftly in-game, whether through visual cues, UI unit tags or tooltips. Also (random) ‘special’ attacks during combat (and their chances of occurrence for different unit types) or other random elements like drops/experience points are bad because the player cannot see the rules they are constructed around. E.g. if a unit can sometimes do double damage, there is a value written into the game dictating the likelihood of this happening, which experienced players will know and newcomers will not.
The same applies to all repetitive-competitive games (repetitive-competitive being a game where you will typically play again and again in the same arena against an opponent, be they human or AI). E.g. in an FPS the maps need to be easy to read and learn because the quicker they can be learnt the sooner the decider of combat becomes down to the players’ decisions. No experienced player wants to be unchallenged and no new player wants to get creamed, so this rules benefits everyone.
I think the easiest way to test this theory to imagine the following scenario: if a player, after having lost an engagement/match, etc., were able to replay that exact same engagement and not do significantly better, there is a fault with the game. This is because there is some mechanic in the game that isn’t apparent from the enemy’s actions. I.e., the player should be able to learn from numerous encounters against an opponent and improve, as opposed to delving into the code or onto a Wiki to discover the hidden constituent of the equation for victory. If it is the player actions that decide the outcome, study of the player actions should be all that’s needed to improve.
If you are making a strategy game, ultimately I think your enemies are Obfuscation, Inconsistency, Haste and Imbalance. E.g.
Obfuscation
- I don’t know what counters this!
- Why is this doing less damage?
Inconsistency
- That shouldn’t happen!
- But they’re the same unit, how can one be stronger than the other?
Speed
- I’m not ready for this yet!
Imbalance
- Why would I not use that?
If you nail those, chances are you could have a great strategy game on your hands.
Forgot to mention one of the few ways ‘random’ can work: In Starcraft 2 the players start in randomly paired locations on the map. This is fine because the map itself remains unchanged, it is only the locations that were unpredictable, but are now known and need to be interpreted to inform a powerful strategy. It’s great how starting locations really close to eachother in SC2 tend to lead to rushes because of short travel times, and so when you discover this is the case you can either put your resources into a rush or counter-rush strategy.
Clarity is bad! Bad bad bad. At least in terms of suspension of disbelief.
I briefly enjoyed playing Europa Universalis 3 until I started reading the discussions on the optimal way to take advantage of how the other state’s diplomacy AI reacts to your state’s aggression. Basically the player’s nation has a reputation, which is a numerical value in the game engine somewhere, which is nicknamed ‘badboy’ on the forums. Do something naughty, like declare war on someone for no good reason and then assimilate their native terretory after winning the war, and your badboy value goes up. Badboy gets too high, and all your neighbours start doing uncooperative things like refusing whatever it is you might ask them for, or declaring war on you without themselves suffering reputation loss. And with time your badboy decreases.
But because the mechanic is fairly simple, it has been reverse-engineered and described in detail on the EU3 forums. The problem is I feel that diplomacy between nationstates should be a bit more complex than { if(badboy > 10) haveWar(); else havePeace(); } . Now, because I have learned of the details of this game mechanic, EU3 feels like a very elaborate graphical frontend on a spreadsheet full of numbers (one of which is my ‘badboy’) instead of making an alternative history where Sweden conquers France.
Good points, Stephen. The random map locations are a great example of one place random can work. Random loot drops are another. Slight random variances in combat outcomes are possible, I believe, though it’s best if the player knows roughly the range of possible results before engaging. E.g., I know one side may randomly get a 5 percent advantage, so I try to insure I’ve got a 5 percent margin on my opponent before engaging.
The quartet of pitfalls you point out is valuable as well, though I do think it’s acceptable to have some of those questions arise — as long as the player can answer them in short order as a result of more experience in play. I like the test you suggest: if the player can’t do better on successive times through the same scenario, something’s most likely wrong.