Episode 178: Unit Customization and Game Design

[Troy Goodfellow]

Sixth Beatle Soren Johnson returns along with new guest, Cliff Harris from Positech - designer of Gratuitous Space Battles and the new Gratuitous Tank Battles. This week, how does the idea of unit customization fit with general design principles? At what point is this mechanic an intrusion into good 4x game design? Where do designer and player expectations collide?

Listen

Rob's column on this issue at Gamespy

[King]

As the guy who did design Heart of Iron 3 I suppose I should get in first here. First things first, I should hold my hand up and say we seriously overreached with the game deisgn, we simply tried to do too much. The game was designed with the notion that you would focus in one part of the conflict and you would leave the AI to run the secondary theatres. The reason this failed was in information, the AI couldn't give you enough useful information to really make use of it properly. If I had unlimited time and budget I would love to redo the whole information interface to give the player better information to make give the player better information and why, what do I want and why do I want it. Then the player could better tailor his/her strategy better. Well that's my defence.

[RoutineMachine]

I think the interesting decision with Galciv II's unit customization was balancing the desire to specialize against your nearest neighbor, with being well-rounded enough to be prepared for the unknown. I liked that aspect of the game, as it was a balance that had to be adjusted throughout the game depending on how threatened you were on any particular turn.

The busy-work of updating my new ships with incremental improvements in technology wasn't so much fun.

[SamS]

This was a Good episode. Great work by Troy coming in with an early statement of definitions to stop the show falling down the semantics wormhole and then a late game home run with the observation on lack of starship engineers in popular culture.

One game though that I wish you guys would have played and been able to talk about is Aurora. It's a real outlier that would make a good reference point for discussion. It doesn't quite model the plumbing as Cliff says, but it does go into insane levels of minutiae regarding customization, yet still holds together.

[spelk]

Enjoyed the episode, good to see Cliffski get some exposure.

I am a gamer who struggles with high levels of customisation in a game. Ultimately, I want to play the game and be gently led into the mechanics, I want to explore the design first, until I totally understand it, and am totally familiar with it, I never feel able to begin the customisation of the units. If the customisation is limited at first, and you're taken through how things make differences, then I can handle that. Having a wealth of options just thrown at you, go play and make it work, I crumble.

Both of Cliff's recent titles overwhelm me with options right at the beginning. GTB was a little more approachable than GSB. Chris Park's AI War has evolved into so much with almost infinite variety that I start it up and am lost again. I think it could just be the way I enjoy games, I enjoy the journey, the progression, the dawning realisation. I don't enjoy the thrown in the deep end, work it out yourself approach. For me to play the game, it has to have a draw of understanding, playability and payoff. Thrown in a complex sandpit with endless options, there's just no payoff, no plans to follow, no clue on how to start. Ultimately, you're beaten down more often than not, for not totally understanding the mechanics at play, and that's when I walk away.

I think customisation in moderation to begin with, is brilliant. As long as you are shown the ropes. Once the mechanics have been imparted, and you're hooked, then open the customisation up to the "Creative" level.

Progression first, lavished with feedback and Understanding, then Customisation to reward those who really do want to challenge themselves within the myriad of options.

[robc]

Disclaimer - I heaven't listened to the postcast yet, but I will. I feel similar to spelk about unit customization. It should be slowly introduced. The benefits for customization should be clear - no customization just for the sake of customization. I rarely find the act of customization enjoyable and it ends up being tedious. Reasons why your designs succeed or fail shouldn't be a secret. For all the stats Gratuitous Tank Battles gave I didn't feel the feedback was clear.

One thing I think games should incorporate to unit customization is a design cost. The player shouldn't be able to spam unlimited designs without incurring some type of cost. Design work takes time and resources, games should reflect that. Designs shouldn't be instantaneous to produce. Ok, I see my enemy has a unit I want to counter. I shouldn't be able to instantly design said unit and put it in production in the blink of an eye. Maybe more designs should increase production time to reflect factories needing to retool themselves between production runs ( a la Hearts of Iron?). In any case there should be costs associated with designs.

[hexgrid]

I'm inclined to think customization works fine in moderation. The problem comes when the game throws you a heap of nearly indistinguishable components and tells you to construct a fighting force, and then forces you to revisit it periodically as new research comes in. See the old DOS game "Stellar Crusade" for an example of how to do it right.

The Stellar Crusade model is as follows:

- a ship is just its components

- with a couple of exceptions, components are unit size; a ship can be from 1 to 15 units in size

- each component has 1 hit point, ship health is the sum of those hit points

- damaged components don't function until they get their hit point back (ie: repair)

- there are four types of weapon

-- long range normal - one shot, lots of damage, a missile, essentially, becomes "damaged" when fired

-- long range i-space - as per long-range normal but hits ships in i-space

-- short range normal - a gun turret

-- short range i-space - a gun turret that hits ships in i-space

- two defenses

-- point defense intercepts attacks against this ship

-- area defense intercepts attacks against the whole fleet

- two cargo types

-- regular, can carry almost everything, but can't be used in assault drops (except with special forces)

-- assault, can be used as dropships

- special modules

-- command module - has sensors, allows a commander to take charge of the fleet

-- sensors

-- i-space drive; this beast takes 13 units of space, but lets the ship enter i-space

-- maybe something else? I can't recall

You didn't need to specify the structure or layout of the ship. If you added a component, the ship had that component. No new components were ever added to the component list; research in the game boosted a multiplier called "efficiency", and when a ship is designed it's baked with the efficiency that the player had when the design was finished, so updating designs is trivial and for the most part unnecessary.

The simplicity also means it's pretty easy to see what needs to be done. Want to take a colony to a new world? Well, colonists are a 15 unit cargo, and you can't split cargo between ships, so your colony ship is 15 units of cargo space. You can't put more than 15 units on a ship, so the design is done. Want a space submarine for commerce raiding? The i-space drive is 13 units, so you've got 2 units to spend on weaponry; that could be a missile (making it a single-torpedo submarine, since long range weapons take two units of space) or two guns. Want a cruiser? Well, now we're into some choices about weapon and sensor mixes, but it's still pretty simple to suss out.

Find a DOS emulator and go play Stellar Crusade. I think it's up on Underdogs. Ignore the EGA graphics, put up with the learning curve, and there's an excellent game hiding in there.

There's no reason that model can't be abstracted to cover other situations. The core of the idea is to give you simple rules with enough capability to construct units to suit tactical needs (trading off speed, attack, defense and range along with terrain considerations) without bogging you down in the details or forcing you to tweak your designs constantly. That way, if the player decides they really need some sort of cavalry analog (fast, hits hard, then gets out of the way for the heavies), it's easy for them to whip that up. Going to higher levels of abstraction is the key.

I've got a design on the back burner for a game that uses a similar ship construction and research mechanism, though it has some twists that I think will add to the appeal. I may revisit it after Rob's comment about German vs. British fleet design; I think there actually is room there for a game where you play the research department of a government at war.

[ShadowTiger]

Customization is very important to me. I can have fun playing games with little or none of it such as Age of Wonders, Master of Magic, Heroes of Might and Magic... but the more the better.

I think that Master of Orion 2's system was really fun to play around with. Fallen Enchantress's system is pretty bland in comparison.

The key is perks. This is a new overarching philosophy of mine, but good strategic gameplay in any game, of any genre, needs well designed perks. +5% damage vs +5% armor is a stupid decision to offer a human. You offer a human exploding fireballs or bouncing lightning. Make there be only a few perks so its easy to balance, maybe 10 or 20, let them choose 3 at a time, but make every combination have its own unique personality that is memorable and changes the experience.

In the end, its probably easier to design static units than let players build dynamic ones. There is no lore justification for customizable units, historically it takes a long time to develop new techniques and equipment, usually the war will be over before significant changes can be made. Maybe there can be an exception for a mechwarrior type universe.

In my spare time I am designing a fantasy 4x game and I originally wanted to have customizable units. I then realized that it opens up a huge can of worms for balance, in addition to watering down your worldbuilding and storytelling. What does it matter if you pick an egyptian inspired race if you are using guerrilla warefare and blitzkrieg tactics. If you want to make the player feel the inspiration in your design, you have to restrict them to play a certain way. You can still have a detailed unit customization system, but you have to limit their tools for creation so that they can't build the same thing every time. You have 3 areas to differentiate: Faction choice, Technology Tree, and Exploiting resources on the map.

So if you combine that advice, I think you can make something really good, but to play it safe, I will be sticking to pre-made units for my first 4x game.

P.S. I tried out GSB, and while it is interesting I prefer to have direct control of my units. Star Wars: Empire at War is more my style.

[Gormongous]

Star Wars: Empire at War is more my style.

That game and its expansion got rightfully slagged for a lot of reasons, but I've never encountered a cleaner and more attractive space combat model, even if it was only two-point-five dimensions.

[Codicier]

Did anyone here ever play Warzone 2100? Although it got quite a bit of play in my local gaming group I never fell in love with the game myself, partly because it often fell pray to just the sort of "MUST OPTIMISE AT ALL COSTS" mentality that got mentioned in the podcast.

Overall my personal attitude is that I like customisation best when it gives scope for personal play styles to emerge, not when it just allows the creation of one super unit that you then pump out 10,000 copies of.

At the start of the podcast Rob talks about unit customisation being a Toolbox from which you can select the right tool for the job. The problem is if you have a sonic screwdriver your not going to pick any of the other tools. I think any game which has units which are just plain 'better' without any appropriate cash or opportunity costs basically risks becomes boring very fast indeed.

Cliffski mentions the Football Manager series, which is interesting because while the Gratuitous series focuses on customising units and FM focuses almost entirely on strategy.

Your not customising to create the ultimate team (at least not initially, although it can get that way over time) your trying to figure out how to best the customise your strategy to make best use of the resources you have v's a constantly evolving set of opponents with different strengths an weakness.

I'd be interested to know if where anyone here who has ever competitively played a CCG's (like magic that gathering) or tabletop wargames fall on this topic, and how much of the fun of those game do they think comes from the customisation?

I've dabbled in both and its amazing how unbalanced the metagame has got at times. There have been times where one deck/list becomes near unbeatable, and I find in these cases the main limiting factor is the random card draws and dice rolls. To put it simply randomness helps give everyone a chance. However occasionally you get cards/units so strong, that they have had to be banned.

It's probably why the duel's of the planeswalker series has seriously scaled back customisation, no newcomer to a system is going to enjoy playing a game where you have no conceivable chance of winning.

[King]

When I think about customisation I have come to the conclusion that in Grand Strategy it simply doesn't work the way you want it. What you expect is to see people with various different types of the same unit all customised for different situations, what actually happens is people find the best unit and stick with that. When you combine customisation and technology what you are seeking to acheive is a sense of progression, but technology does that on it's own, thus customisation is not needed for the purpose. I am more of the opinion that a very limited choice of customisation that ensures that your units can fufill very different roles. When you give very limited but clear choices it is easier for the player to make them.

[juv3nal]

What you expect is to see people with various different types of the same unit all customised for different situations, what actually happens is people find the best unit and stick with that.

While I agree that my impulse would be to "find the best unit and stick with that," can't this be addressed by balancing? If the cost of generalist units is made prohibitive enough then people will look at the alternatives, no?

Reasons why your designs succeed or fail shouldn't be a secret. For all the stats Gratuitous Tank Battles gave I didn't feel the feedback was clear.

...

One thing I think games should incorporate to unit customization is a design cost. The player shouldn't be able to spam unlimited designs without incurring some type of cost.

I think it depends very much on the type of game. I see GTB as more of a puzzle game (in the same way that I think tower defense generally are puzzle games) rather than simulationist. In that context, the cost of trying out different possible solutions should IMO be minimal or it will only lead to frustration. i do agree about GTB failing to give comprehensible feedback despite its plethora of statistics though.

[King]

While I agree that my impulse would be to "find the best unit and stick with that," can't this be addressed by balancing? If the cost of generalist units is made prohibitive enough then people will look at the alternatives, no?

Yes and no, as anyone will tell you balancing is not that easy and the more customisation options you give the harder it becomes to balance. This also doesn't change the fact that some choices are basically crap. In HoI3 you can combine you super heavy King Tigers with Militia but you would feel just a bit silly doing it. So although balancing can make some of the choices more interesting, it doesn't get round the problem that some choices are simply bad and no matter how good your choices become there is always one choice that you will prefer when it is there.

[robc]

While I agree that my impulse would be to "find the best unit and stick with that," can't this be addressed by balancing? If the cost of generalist units is made prohibitive enough then people will look at the alternatives, no?

I think it depends very much on the type of game. I see GTB as more of a puzzle game (in the same way that I think tower defense generally are puzzle games) rather than simulationist. In that context, the cost of trying out different possible solutions should IMO be minimal or it will only lead to frustration. i do agree about GTB failing to give comprehensible feedback despite its plethora of statistics though.

Point taken on a game like GTB. Regarding having a cost associated to each design I'm thinking about more of a typical strategy game - Galactic Civilizations, Sins of a Solar Empire, etc...

[biz]

Customization can actually reduce the learning curve.

It's easier to learn about pizza ingredients and then order the pizza you want rather than memorize the 100 different names for all the different combinations.

If hyperthrusters enable +1 speed for everything, then you know the exact effects of researching hyperthrusters.

If they enable +1 speed for a subset of units, then you need to memorize.

If there are 30 technologies just like this, then you have to memorize a lot or spend a lot of time looking ahead in a tree.

[Procyon Lotor]

I wish there would have been more discussion about specific 4x games that have used customization, and what worked and what didn't. In my humble opinion, unit customization works great in MOO1 and MOO2, but not so much in GalCiv2 and Alpha Centauri, an opinion that I think is consistent with the general consensus (except, perhaps, rregarding GalCiv2). Instead, there was too much discussion about either liking customization completely (Cliff), hating it completely (Rob & Soren), or sometimes liking it and sometimes not (Troy). I think the more interesting discussion would be about why unit customization sometimes works and why it sometimes doesn't. With this panel, that would have required Troy to debate himself.

[Destrin]

I know it's off at a slight tangent but I think the best unit customisation I've ever had the joy to partake in is that used in the board game Eclipse which uses a very simple system to control the customisation. You don't have much in the way of options but you are limited in how far you can customise as it takes an action to play with the ship components. I've seen some very varied ship designs crop up in that game despite the simplicity of the implementation.