|
[The latest installment of Ernest Adams' Designer's Notebook series once again takes a close look at major and minor design flaws made by developers -- when it comes to interface, A.I. behavior, or pure game design -- and highlights them so you know what to avoid doing.]
Welcome to the 12th installment of Bad Game Designer, No Twinkie! I think we're having a positive impact, folks -- many of the games mentioned in this edition are fairly old, so it's just possible that newer games are starting to avoid the worst design mistakes. I'd like to hope that this gripe-fest is helping. However, old or new, we can always learn from the errors of the past, so I'm going to go right on documenting them. As always, thanks to my many contributors, and you're welcome to send more complaints to notwinkie@designersnotebook.com.
As usual, I've tried to mix and match between small things that are incredibly annoying (but easy to fix) and larger, harder problems. This year we have eight.
Uninterruptible Text
I read fast, and I don't like to re-read something that I've just read five minutes ago. Waiting for text to... scroll... slowly... by... drives me nuts, especially if it's text I've already seen.
Joshua Gault wrote, "This is most annoying when you save, then there is a long conversation between you and some guy, and he turns out to be the boss.
"This is particularly bad at the end of all of the Mega Man Battle Network games. I don't like breaking my A button from anger because I must retry the boss, which requires me to go through four pages of text."
It's very simple: non-interactive text should be interruptible, just as movies should be.
Oh, and don't put save points before long non-interactive sections, either -- text, cinematics, or empty regions the player has to walk through. But you knew that one, right? I mentioned it last year in Bad Game Designer, No Twinkie! XI.
Rapid Non-Stop Text
The flip side of the foregoing is text that goes by too fast. Shairi Turner wrote, "I have a problem with dialogue moving too quickly. We don't all read at the same speed. While I may have found pressing X or clicking to be tedious in the past, I miss it when it's gone." This is a basic accessibility failure. (Most games have terrible accessibility.)
You need two buttons: Advance to Next Page (which should happen instantly, not in a slow scroll or worse yet, a letter-by-letter display -- TeleTypes were old news by 1985, okay?) and Jump to End, which should take the player to the next point at which she has to take action or make a decision.
False or Pointless Alignment Systems
I've always thought that explicit alignment systems, a la Dungeons & Dragons, were kind of pointless anyway; they constrain role-playing and discourage enacting characters with flaws or complex personalities. Is Hamlet good or evil? Well, he killed Polonius, so I guess that makes him evil. Whew. I'm glad we've got that sorted out.
It's even more of a problem in computer RPGs that keep track of what you do, but only in a simplistic way. Luke Bainton wrote,
Generally, you have to decide from the outset whether you're going to be the white knight or the black knave; you will only get benefits from relentlessly pursuing one or the other. On top of this, the contrast between being "good" and "bad" is usually so far divided it's impossible to relate. In BioShock your choices are between being a loving caretaker for the rest of your life or turning into a comic-book supervillain with destroy-the-world intentions. I have no desire to do either!
The concept of neutrality is usually poorly implemented as well. In some games it means you need to balance good acts with evil ones, which gets hard to swallow. Roughing up one pedestrian for a few coins and then helping another across the street doesn't make you neutral, it makes you schizophrenic.
If you want to track the player's behavior and generate consequences for it, by all means do, but the consequences have to be proportionate to the activity. And if you're interested in rewarding moral, or immoral, behavior, it's better to do it via some in-world system than an arbitrary alignment. For example, if players want to be evil, let them join the Crime Guild and work their way up, gaining benefits as they go. They shouldn't be thrown out for the occasional act of virtue, nor should they be thrown out of the Heroes Guild for a little burglary in a good cause.
While we're on the subject of alignments...
Forcing the Player to Violate His or Her Alignment Unnecessarily
The example for this goes back a way, but it's a good one. Benoit Girard chose to play a good Jedi in Jedi Knight II: Jedi Outcast, but...
You could get a powerful Jedi mind control trick allowing you to change an enemy into a neutral NPC permanently. You could use it to neutralize your opponents before killing them (not very Jedi-like), or just neutralize them and continue exploring the current level.
In a Bespin level, you ended up in an arena-like place where dark Jedis would attack you, and from which you could escape only after defeating them. The problem is that the appearance of the dark Jedis was triggered by the slaughter of all the other opponents in the level... something I figured out after more than half an hour trying to find the secret exit I might have missed somewhere in the level, while a bunch of formerly-hostile NPCs were walking around randomly.
 So Benoit, being a good Jedi, courteously neutralized all his opponents without killing them, then couldn't find the exit because it didn't exist yet. The only way to get out was to massacre all the harmless neutrals in cold blood, thus triggering the appearance of the dark Jedi, and then kill all of them before the door would open.
And we wonder why some people are concerned about violence in video games. Note that this is what you're supposed to do when you're the good guy!
I already covered Illogical Victory Checks back in Bad Game Designer, No Twinkie! VIII, and this is certainly an example. (Normally, you don't have to kill people to unlock a door.) But apart from that, it compels the player to violate his alignment, requiring him to do something that the game has told him not to do.
Lying to the player about how he's supposed to play the game is almost never a good idea. (Yes, I know about Shadow of the Colossus, and I'm not convinced.) Worse yet, it fails to recognize lateral thinking. Benoit neutralized his opponents without shedding a drop of blood. That should be rewarded, not ignored.
It's one thing to put the player in a moral dilemma for dramatic effect. But this was no dilemma, it was just bad level design... and a Twinkie Denial Condition.
|
Printer-Friendly Version
As it is units can be kited out away from defensive structures if left and the player who is controlling them is given the expectation that they should be keeping an eye on them. If the player is given the (false) expectation that the units will be smart, they'll learn to get slack and not have any concern for what their troops are doing.
So unit AI is kept deliberately simplistic to allow the player to control them with basic instructions and allowing emergence to sort out the details.
Not sure I agree with that - the AI behaviour Starcraft exhibits is equally open to exploitation, is highly unrealistic and makes the game much more difficult/frustrating for those who aren't adept at micro-management.
@David and Steven:
There's a very, very good reason that Starcraft didn't have better AI: it was released in 1998 and targetted a 90-mhz machine with 16mb of ram. Each individual unit (up to 200 per player) could therefore only have a small amount of CPU time and memory allocated to it.
(arguably, Blizzard *could* have implemented better AI - game reviews at the time noted it was fairly limited - e.g. http://uk.gamespot.com/starcraft/reviews/starcraft-review-2533189 - but that would probably have meant sacrifices elsewhere!)
Starcraft 2 could have implemented better AI - after all, Moore's law has significantly marched on the last decade. However, with such a dedicated "competitive" fanbase having grown up around Starcraft, making such a change would most likely have been commercial suicide; whether it was good or not, many fans would probably view it as "dumbing down" and reacted negatively...
Most of the optimization mentioned could exist, but as pointed out, would seriously harm the competitive players who need fine grained control and predictability. A great example is how the musketeer concept (one for all) can backfire in one of many ways. For example, how often/how does the one for all check work? For example, Unit A and B are muskateers together/ Unit A moves to fire on something attacking Unit B which exposes him to a new enemy so unit B moves to attack what's firing on Unit A which exposes HIM to a new unit and so on. Before you know it your units set to Deny Passage or Defend Position are 1/2 way across the map or pulled completely out of position.
The reason why there's no 'hold fire' in Starcraft is it's unnecessary. Cloaked combat units aren't good scouts and even if they were a move command prevents attacking. In all other situations attacking immediately is a better option to holding fire or accounted for in the balance of units in the case of splash damage units.
In Starcraft's scheme it has the minimum necessary for all possible actions (assuming human intervention) so that's what it went with. There's nothing extra that could have been removed, and nothing that could be added that a player can't already do.
Basically in short Better AI is NOT the same as better unit control.
"I think that's a really dangerous attitude to take Jamie. You shouldn't be so willing to dismiss a team's design decisions as pandering to fans when there are plenty of legitimate reasons that people can bring up in support of the decision. By dismissing it as fan service you run the risk of not learning from someone else's work."
I'm sure there's other reasons, but fundamentally, when you have such a heavily entrenched fanbase, you're not going to change the formula; people tend to react negatively to change, even if it's for the better in the long run. "Long term" systems such as Facebook, Hotmail and the Xbox 360's dashboard can ride out the storm; "short term" systems such as video games (which rely heavily on the initial sales surge) cannot. Look at how fan-reactions affected Dragon Age 2 (i.e. rating-bombing on sites such as Amazon and metacritic); even something as relatively simple as the announcement of Diablo 3's "always online" requirement triggered a huge backlash (http://multiplayerblog.mtv.com/2011/08/04/blizzard-vp-surprised-over-fan -reactio
n-to-diablo-3-online-requirements/)
"Most of the optimization mentioned could exist, but as pointed out, would seriously harm the competitive players who need fine grained control and predictability. "
How would it harm them? They can still manually monitor and order each of their 200 units as needed, and unit behaviour would still be deterministic/predictable. Then too, it should be fairly trivial to build in an option to disable specific "AI-assist" options for competitive-multiplayer.
"In Starcraft's scheme it has the minimum necessary for all possible actions (assuming human intervention) so that's what it went with. There's nothing extra that could have been removed, and nothing that could be added that a player can't already do."
Hmm. Theoretically, everyone can use a manual gearbox and there are many good reasons to do so (e.g. better control, better fuel economy). But the vast majority of cars sold in the USA have an automatic gearbox. Similarly, people can cook their own food (and it's generally healthier/cheaper), but yet: fast food shops are popular.
There's a lot of people who enjoy micro-management games, but I'd argue that there's a hell of a lot more who don't. Personally, I take the view that automating manual, repetitive processes is what computers are designed to do - I don't want the computer to fight the war for me, but I'm quite happy for it to handle some of the grunt work. And there have been game series which have improved/streamlined unit management, such as the Oddworld series.
I'd also point out that in real life, generals don't micro-manage individual soldiers: they give orders to groups and then expect the soldiers within those groups to carry out those orders in the most appropriate way. And that's what I want: I want a unit-control/AI system that I can give orders to - for instance, instructing troops to ignore enemy fire, retreat when fired on, or only target specific enemy types. Better yet, I'd like to be able to give groups timed orders, so I could set up a multi-point assault where everyone moves into action at the same time. This would then mean I could fight in a more strategic way, rather than having to deal with things on a short-term, tactical, micro-management way.
Now, you could describe this sort of "order queuing/AI assist" dumbing down - personally, I think it's more like the difference between American football (strategic play) and rugby (tactic play). And I suspect it'd actually lead to more interesting competitive gaming...
Well, Dark Reign, released a year earlier, has much, much more detailed unit controls, including some of the features Ernest discusses. Computer specs were not the limiting factor in Starcraft, though perhaps they're simply inherited from the earlier Warcraft games (where specs may have been important).
As you said though, it could well be that Starcraft was limited by the fact that it was (presumably) built on the Warcraft II engine, which was originally designed for a 33mhz 486 with 8mb of ram. And to be fair, it was also designed to scale up to 8-player battles (i.e. up to 7 AI players and up to 1600 units), so there was definitely an argument for keeping per-unit ram/CPU usage to a minimum...
Well really if that's true then it can't be a bad design decision to make that sort of game can it?
I don't at all think that a Starcraft style game is the only way to do a "control an army" game and games with broader strokes and less precise, but more intricate, unit control could very well work well as part of a strategic game. Even on top of other RTS games it might work better, a game like Company of Heroes is a much better environment for something like timed order controls because of the cover and suppression mechanics. In SC2 the commands are often just extra baggage because executing a flank is often as simple as positioning units properly and then hitting 1a2a (or sometimes just 1a). It's just that overall it's two sides of a coin, much like Call of Duty and Quake are both the same genre, but completely different. What could very well work in the genre just doesn't fit in the space that Starcraft has carved out.
The reason I think it's dangerous to ignore what SC has done as fan service, is you need to understand WHY it's like it is so you can do something different. If you simply take Starcraft and slap on top of it all the unit control features you want, the game is going to be TERRIBLE. Instead you want to understand WHY Starcraft unit control is the way it is, and how that influences future design decisions. Then understand how what you want is different, and use all that knowledge to build up from your unit control style into a game that fits it well and works great.
I just know that over the ages many games have given extra unit commands that just don't really work all that well and end up being vestigial or not part of the core game play (which is bad design I think you'll agree). Things like odd formations that are never used and are tactically unsound or weird hold fire commands that are always a bad decision. Meanwhile there are some gems worth learning from too so it's nice too look at other RTS games. A good example is Rise of Legends which featured infantry squads that had individual soldiers that could be killed, and importantly killed by splash damage or being run over. So in their system it made a ton of sense to include a 'scatter' command that felt great and worked well to keep units alive, though since it sent units in random directions it also had a drawback in how it felt to use it. That's exactly the sort of detail I think it's important to notice, understanding why 'scatter' in a game like Rise of Legends felt good, but in a game like Starcraft it'd feel bad, is key to understanding how to build your own quality unit control system.
Also keep in mind with something like Starcraft there's the secondary issue of good spectating which has it's own concerns and motives that don't necessarily line up with other aspects of good design. With less unit commands it's easier for a spectator to understand what's going on and how a player is accomplishing a task. A spectator doesn't need knowledge of 5 different types of unit behaviors to make a prediction on what's about to happen all they need to know is that without player intervention the units will either return fire while perusing an enemy, stand still and return fire, stop moving to attack, or move without attacking. For a spectator's benefit it helps when the player's involvement in the game is brought more to the forefront over automated tasks or harder to understand strategical decisions.
>> There's a lot of people who enjoy micro-management games,
>Well really if that's true then it can't be a bad design decision to make that sort of game can it?
I never said it was bad, I'm arguing that it limits the audience - to fully quote myself from above, what I said was "There's a lot of people who enjoy micro-management games, but I'd argue that there's a hell of a lot more who don't".
These days, I think the audience for that sort of "hardcore" gaming experience is far smaller - while there's a lot of other factors which could be involved (not least of which is the fact that SC has been available for over a decade), it's worth noting that Starcraft II has sold approx. 5.5 million copies, which is significantly below the 11+ million sales of the original. High precision, micro-management gaming was perhaps more accepted a decade ago; as Nintendo showed with the Wii (and as per the general shift to console gaming), people are perhaps not looking for that sort of challenge these days.
> The reason I think it's dangerous to ignore what SC has done as fan service, is you need to understand WHY it's like it is so you can do something different. If you simply take Starcraft and slap on top of it all the unit control features you want, the game is going to be TERRIBLE. Instead you want to understand WHY Starcraft unit control is the way it is, and how that influences future design decisions. Then understand how what you want is different, and use all that knowledge to build up from your unit control style into a game that fits it well and works great.
Agreed: the history of gaming is littered with examples of games which failed because the developers attempted to shoehorn in badly thought-out features. And there's definitely wisdom in the "if it ain't broke, don't fix it" approach to design. But I still think that SC2's conservative design was at least partly due to changes being viewed as commercially too high-risk - as per above, even if they were for the better, a significant minority would view them negatively and actively campaign against them.
But we'll probably never know for sure, so we may have to agree to disagree ;)
> I just know that over the ages many games have given extra unit commands that just don't really work all that well and end up being vestigial or not part of the core game play (which is bad design I think you'll agree)
Again, agreed - but as per Chris's comments above, games such as Dark Reign and Total Annihilation showed that it was possible to build better unit-management tools a decade ago - and TA in particular is still relatively popular for a decade-old game. Are we really saying that after a decade of game-design evolution (and the march of Moore's law, which has increased processing power anywhere up to a hundredfold since 1998), the best we can come up with is more of the same?
> Also keep in mind with something like Starcraft there's the secondary issue of good spectating which has it's own concerns and motives that don't necessarily line up with other aspects of good design.
Um. Is Blizzard really that bothered about designing the game for spectators? With the exception of Korea, spectating video games has never really caught on, and while the support for SC in Korea is near-fanatical, it's not a big enough market to justify implementing a "spectating" system which could impact the game's playability and damage it's reception in other markets.
(and it's also worth noting that - possibly partly due to the legal spat between Blizzard and KeSPA - SC2 has not proven to be that popular in Korea, anyway.)
> With less unit commands it's easier for a spectator to understand what's going on and how a player is accomplishing a task. A spectator doesn't need knowledge of 5 different types of unit behaviors to make a prediction on what's about to happen all they need to know is that without player intervention the units will either return fire while perusing an enemy, stand still and return fire, stop moving to attack, or move without attacking. For a spectator's benefit it helps when the player's involvement in the game is brought more to the forefront over automated tasks or harder to understand strategical decisions.
Um. Pretty much by definition, a spectator is a fan of the game and will therefore have significant knowledge of the game, in much the same way as soccer fans understand the rules for free kicks, throw-ins, penalties and free kicks.
Also, the sort of things I suggested (e.g. queued/stacked orders) is something that's easy to replay and commentate over (e.g. "Player 1 has set up a camoflaged sniper on the far side of the enemy base, with orders to target the commanding officer when the first wave of the assault hits, as well as two separate groups who will carry out feints before the attack starts") - and it's also far more open to analysis than the "300 actions per minute" of a top-level SC player. But as I stated above, I don't think innovation was ever a part of the plan for SC2s design, as it would have hugely increased the risk of commercial failure.
Just two points though.
1. On sales numbers... It is 10 years for Brood War vs 1 year for SC2 so that's going to be a factor in total sales (keep in mind pretty much up until SC2's release BW was available in video game stores). Also you now have the added factor of digital sales which don't always get reported the same way and sometimes get missed or are kept secret.
2. The two paragraphs that start with Um. are out of touch with the way SC2 has been received. Spectating SC2 in US/EU has blown up to be a pretty major thing. Within the year SC2 has taken over MLG from other games like Halo and tons of new leagues and major prizes have popped up. Popular SC2 streamers can pull in 10-15k concurrent viewers and major tournaments can pull in 90k concurrent viewers. Dreamhack 2011 had 1.7 million unique viewers. On top of that part of that crowd rarely or never plays Starcraft 2; it's not just all die hard Starcraft fans. While any long term watcher will pick up the rules, the idea is that someone should be able to be hooked just by watching a single game with no prior knowledge (much like Soccer, Football, etc.).
There is some reception issues in Korea for SC2, but it's still doing fairly well. The GSL is a major tournament with pretty serious prizes (50-100k per tournament). Not everyone has switched, and many Koreans still prefer BW, but there's still a sizable Korean SC2 scene.
Meanwhile every major tournament has to strike a licensing deal with Blizzard and you can bet that's part of their plan for their long term monetization. It's almost definitely a factor in why the game is online only. So I think it'd be a little short sighted to claim they didn't factor in the spectating.
Sure we can go back and forth on how unit commands would affect spectating quality and that's fine and something I don't care to argue at the moment, and you certainly may be right, but I just wanted to draw attention to what the spectating scene is like and how big it has grown.
It's been a good debate :)
> 1. On sales numbers... It is 10 years for Brood War vs 1 year for SC2 so that's going to be a factor in total sales (keep in mind pretty much up until SC2's release BW was available in video game stores). Also you now have the added factor of digital sales which don't always get reported the same way and sometimes get missed or are kept secret.
True - as I noted above, SC has had a decade of sales - and I know several people who have rebought it several times as discs have gotten scratched/lost. But I still think the market for "hardcore" titles is smaller now than it used to be, and I think this is at least in part because the entry barriers are getting higher for newcomers - having watched non-gamer friends and family on games like GTA3, Dragon Age and Zelda:TP (which, ironically, is a Nintendo franchise which is meant to have low entry barriers), it's scary how much I take for granted or implicitly understand, thanks to several decades of gaming.
But that's something for another time and article ;)
> 2. The two paragraphs that start with Um. are out of touch with the way SC2 has been received. Spectating SC2 in US/EU has blown up to be a pretty major thing. Within the year SC2 has taken over MLG from other games like Halo and tons of new leagues and major prizes have popped up. Popular SC2 streamers can pull in 10-15k concurrent viewers and major tournaments can pull in 90k concurrent viewers. Dreamhack 2011 had 1.7 million unique viewers. On top of that part of that crowd rarely or never plays Starcraft 2; it's not just all die hard Starcraft fans. While any long term watcher will pick up the rules, the idea is that someone should be able to be hooked just by watching a single game with no prior knowledge (much like Soccer, Football, etc.).
On spectating numbers: fair enough - it's not something I've been actively tracking, so wasn't aware of the numbers. In any case, a quick check threw up an interesting paper on SC spectating (http://jeffhuang.com/Final_StarcraftSpectator_CHI11.pdf), though it sadly doesn't try to provide any estimates on how the spectating demographics break down.
(for whatever it's worth: I went to Korea for a wedding a few years ago. It was quite surreal to turn the TV on in the hotel and watch a SC battle...)
On "rules": spectator sports generally have complex rules and terminology which are difficult for a newcomer to understand. I defy anyone to understand the rules of American Football/Cricket/Baseball/Snooker/Chess/poker on their first pass. American football is incredibly bizarre to a newcomer (I'm british, so fall into that category;) ) and cricket is chock full of wierd and wonderful terms such as googlies, overs, long leg, stumped, no-ball, lbw...
If it's entertaining (and this is driven at least in part by the "human" factor - there's a reason why most televised sports have a live audience *and* a commentator), people will watch and pick up bits as they go. After all, they're not actively playing, so they don't have to fully understand the rules and their implications: they just need to know enough to keep track of the gross flow of the game...
Ernest's "three musketeers" improvement of Starcraft's Hold Position command ignores the realities of the game's mechanics. David gives an example of a situation where holding position is essential; such situations repeat constantly throughout Starcraft. A hydralisk shot "for free" while its mates are holding position just outside firing range and the player doesn't bother to micro is a tolerable loss. Twenty "musketeer hydralisks" auto-advancing into siege tank range when their mate gets shot at is a total disaster. The Stop command provides roughly sufficient "musketeer" behavior when that is what the player desires. The "problem" of units pursuing enemies across the map is not one that actually emerges a lot while playing the game.
I think we'll have to agree to disagree. To use an sporting analogy: Starcraft is like soccer: a constant, free-flowing tactical battle. What I'd like to see is more akin to American Football, where you can pre-plan your tactics and then adjust them on the fly once a play starts.
Both are valid models. Both sports enjoy a significant fanbase. But the latter is arguably more of a strategic game *and* arguably involves the team's manager more (whose role has significant parallels with the player's role in SC), as he's better able to monitor and control the flow of the game. And it allows the player to focus more on high-priority actions, such as guiding an assault or defending an emplacement.
(Then too, arguably, the additional features wouldn't affect high-end multiplayer, as both players would have the same capabilities, negating their effectiveness. However, it would lower the entry barriers for new players and give less skilled players more of a chance)
But hey. If people like micro-management, that's fine. But it's not the only way to do things, and it's a bit of a shame that there's not been more investigation into throwing all that modern CPU power at something other than shinier particle effects...
I also like that Mirror's Edge rewards a player with some hefty gamerscore and achievements if you complete the levels/game without shooting anyone. I like to see encouraging creative - if not unorthodox - gameplay.
Showing players things which they can't get to is definitely a No Twinkie situation, unless it's made clear that there's a reason they can't get to it - GTA3 is a prime example of how to do this well, with it's closed bridge.
Unfortunately, this is something both Arkham Asylum and City fail at: as you play through the main campaign, you'll often see things that may not be reachable. This leads to several issues:
1) If it is truly unreachable, the player may waste significant amount of time trying to get to it
2) If it is reachable, the player may not realise this and move on
3) It forces the player to backtrack multiple times
All told, it's a surprisingly clumsy system in an otherwise excellent set of games...
Another annoying point is the fact that having finished the main campaign (admittedly without completing the catwoman levels - I have the DLC code, I just haven't typed it in), not only is the game only 57% complete, but I still don't have access to three of Batman's gadgets, as I haven't accrued enough XP. As far as I'm concerned, not having access to equipment *AFTER FINISHING THE GAME* is definitely a no-twinkie situation...