Player democracy, the FemShep dilemma, and why do we have unpopular classes/ races anyway?

So Bioware recently decided to put a female version of their lead character on the box art for Mass Effect 3, and held a poll on Facebook to let people vote on how she should look. The long-haired blonde won by a large margin (I think my fave was number 3), but that isn’t the point. The point is that they decided to do it in the first place. Not so much the voting – brands have been doing that on Facebook since (it feels like) forever to drum up customer interest. But that they decided to switch the advertising to show off the female version of the character.

I guess it can’t hurt in a world where sex sells product to have a pretty blonde woman with boob-enhancing body armour on the front of a box either.

82% players can’t be wrong, can they?

Last month, Bioware discussed this decision and revealed two interesting statistics. Firstly 18% of Mass Effect players use a female model for their character. And secondly, 13% of players go with the default (male) option and don’t tweak it at all.

So the question is, if they know that only 18% of their players pick a female version of Shepherd, does that really justify that cost and time involved in putting in all the extra female animation and voice work? Wouldn’t it be more cost effective to just not bother? This is a question that I predict we will see more often with respect to F2P games. If it’s part of the ethos of F2P designers that they provide more of whatever customers are willing to pay for, then surely the majority choice must get more attention unless the minority is willing to pay more for their wishes to be taken into account. After all, time is money.

This will be even more marked in SWTOR, because there is that much more voice work. And how many female bounty hunters and sith warriors do people really think there will be? There must be a level on which Bioware decided that offering diversity was worth it, which is one of the reasons that I love them.

Another point is that yes, lots of people are not interested in spending time customising their character, even just to pick one of the several basic settings. This implies to me a UI issue because even people who don’t care how their character looks could manage to select something in a Facebook poll. And I do think it is of value to get people to show a preference or two during character creation, it can certainly make the character feel more personal. The Facebook poll type of character customisation is much closer to how WoW presents character generation. You have to pick one option for face, hair, etc but can’t customise it beyond that. There’s a ‘select random’ if you really don’t care, and still you’ll find that some faces in particular don’t show up much among the player base (which implies that most people weren’t just hitting the random button.)

Race and Gender popularity in WoW

mmo-champion published some data about relative popularity of race/ gender options in WoW recently. There’s nothing new here to surprise anyone, the least popular options are Orc female and Dwarf female and female characters are less popular across the board than male ones with the exception of Draenei (I hope some lore writer is going to portray them as a matriarchal race to reflect this ;) ). Blood Elves are probably the race with the least dimorphism (ie. the females and the males don’t look so different) and their ratio of male: female is roughly equal.

But other than the Draenei and Blood Elves you might be excused for wondering why Blizzard do spend all this time on female models (including new dances etc) when a new race is brought in. Remember all the fuss about worgen females? Only 20% of the worgen players picked them. But sometimes, percentages don’t tell the whole story. And if those female models weren’t there, maybe those players would simply not have picked the game up at all …

So if I am not overly enthused about player democracy in games, it’s because I know that I am in a minority and I don’t trust the majority to vote with my interests in mind. It makes me glad that there are devs who won’t give them the option.

Improving Roleplaying: Measuring and Rewarding Roleplay

This is the sixth (and last) in a series about improving roleplaying in MMOs.

But in this last installment, I’m going to cut to the quick and ask how we can legitimise RP as a playing style. In order to do that, we need to find ways to measure and reward it in games.

Previous posts were:

People have been roleplaying for fun or for ritual since forever. Depending on how you define the term, roleplay can include anything from putting on funny voices to tell a childrens’ story, improvisational theatre, playing out sex fantasies, or acting out elaborate ancient rituals in full costume (some people would put modern religious rites into this category too).

But as an actual gaming system, roleplaying is bounded by rules and rewards. The standard tabletop session ends with the GM assigning points to people based on various criteria including their roleplaying and how well they completed their character goals for the session. (In D&D this was traditionally done by giving xp for monsters killed and treasure looted, since that was pretty much always the characters’ goal.) The players could then spend their points on improving their character.

MUDs and MMOs simply mechanised the process. You still get xp for killing mobs and completing quests. But somehow, the optional extra xp for good roleplaying disappeared.

This isn’t because good RP is subjective any more than good writing or good acting is. Sure, there’s a subjective side to personal preferences but you could imagine some rough rules. Give people a few ways to define their character. See how well the role was played, given the circumstances of the scene. Judge the use of language. How well did the person play with others and accept their RP. There are ways in which you could make a start on the problem.

But computers are not (currently) good at evaluating style. G Christopher Williams writes in Pop Matters about why you can’t have a Project Runway type of computer game. He’s also musing about whether style has its own internal rules, which could be coded into a game.

And the biggest problem of all – before we measure anything, we need to define what it is.

Automatic Measures – Activity, Language

In the text-based days, we knew what we meant by roleplaying. It was all to do with playing out scenes with other people by typing in sentences of text. You wrote your character’s action, then waited for other people to respond. The action might include something they said, or something they did, but you’d input one pose at a time.

We experimented with automated measures. Activity was the most obvious one – measure how many hours a week a character has spent in a scene with other people where all of them were active. Of course it could be subverted by macros, so you weight that metric down accordingly. You could also add a factor if they have interacted IC with a lot of different people, as opposed to always the same one or two. Giving a bonus for attending large scenes adds an incentive to do it.

You can also reward people for writing backgrounds, writing chronicles of their play, or some other form of non-direct RP.

Measuring language is a more difficult prospect. This would involve weighting scores towards longer and more involved sentences. So the people who wrote more would basically get some recognition. There’s no guarantee that this meant better roleplaying, but the person who never posed more than “X nods and smiles,” would be encouraged to write a little more.

You could imagine a smart parser that was able to do a more thorough review. But it’s a lot of work and a lot of computing power, and even if it was possible, you end up having to argue with a player as to whether or not their character would really have said or done something or not. The point at which you say, “Sorry, we don’t think your character would have done that,” is the point at which you’ve failed, even if you were right. Because it has to be their decision on how to play their role.

MMOs add more possibilities for people to act out roles, and they won’t be purely text based. People use emotes, whispers, and all sorts of props and external details to get into role and interact with other players.

But if you had, for example, some roleplaying instances, you could measure activity within them. Even if it was just time logged in. What might a RP instance involve? It might be no more than an environment which could be manhandled by players. Imagine a town hall where you could rearrange the chairs, a theatre with optional backdrops and sets, anything that is designed purely for roleplaying purposes.

And even that would still exclude RP that happens in other places. The act of introducing measurements already limits how people will choose to play.

Human Measures – Voting, Parsing the Logs

The other way to measure any social gaming metric is to use social measures. Let people vote on each other’s RP. Give everyone a handful of votes which they can cast every week. Make it easy for them to nominate anyone with whom they have interacted. Then every week, tot up the scores and assign some kind of weighted average points. Maybe even reward people for using their votes to vote for someone they hadn’t nominated in the last month to encourage them to mix.

This is subject to all the usual voting issues. People can form into cliques. People can organise tit-for-tat voting. But ultimately, as long as they are voting for people with whom they enjoyed RP then the system has some value.

The other big issue with this type of system is that it rewards certain types of character more than others. The socialite who flits from bar to bar has a great excuse to get into lots of scenes. The reclusive, hunchbacked, xenophobic wizard (whilst probably being a really cool character) probably doesn’t.

Another method is for expert judges to scour the logs and watch the scenes and give out bonuses when they observe good RP. Again in text games, we occasionally used to do this. It was considered rude for staff to sit invisibly in a scene without telling anyone, but we encouraged players to send us the logs of any scenes they’d really enjoyed. Partly so that we could put them up on the website to encourage new players and demonstrate how the game could work, but also so that we could reward people.

However, what is practical in a small private game isn’t always smart in a massive MMO. It’s a huge amount of time and effort to judge scenes. Even beyond that, judges can be biased.

So using voting is viable, if you limit the effect of cliques and cartels. Using activity monitors is viable, as long as you are careful not to reward time spent in game too massively. Using GMs to judge selected scenes is viable, if you can find a way to be fair to all the playerbase while doing it.

What sorts of rewards do roleplayers want?

Although xp was the traditional award for roleplaying, it was never the only reward given out in tabletop games. GMs used player RP as a stepping stone to introducing new storylines, new NPCs, and maybe even giving out loot.

Roleplaying in games became synonymous with planning your character’s next moves. I always felt that this was a trap. The strategy of ‘what is our optimal best move to do next’ wasn’t always the same as ‘what would my character do next’. So we have to be careful of rewarding strategy when we actually wanted to reward RP.

But opening up new storylines for the character is one of the ways I used to try to reward good roleplaying. If someone wrote me a great background about their long lost friend who had become a vampire hunter and I was impressed with their RP, there was a good chance that the long lost friend would show up in the game.

Really, the best rewards for RP are props, plot hooks, and new storylines that open up avenues for further roleplaying and empower the player to go ahead and drive it. If you let a character find an ancient book which is being hunted by powerful secret societies, she has a strong hook by which to involve other characters in her story.

Proactive RPers will happily suggest plot hooks which involve their character playing a key or starring role in some storyline. If they are also good RPers and good social players, then one option is to let them have their plot hooks. They can take the responsibility to go run with the plot.

RP as a pure style is probably best removed from xp altogether. Because it was never about making your character more powerful by sitting in bars and talking about the weather. Or rather, those things may make your character more influential, which is a form of social power rather than a shiny new skill or weapon. But players love xp and use it as the only form of measurable advancement in game. So it is a dilemma.

If you give out the rewards that they want and will use, then those rewards won’t be valued as highly as the unimmersive xp point. I suspect that until we have good social systems that measure influence and gossip as well as how many monsters you have killed, we’ll never get to the bottom of this problem. It is an issue that remains to be solved, not only for RP, but for any social activity in game.

Perhaps having a good social network has to be its own reward. There certainly are side-benefits for whatever type of game you are playing.

Should we even try to measure and reward some styles of play?

This is the big question. As soon as you introduce game metrics, people will jump in and figure out how to game the system.

But it would be cool to find ways to reward people for interacting in character, for taking decisions because it’s what their character would have done and not just because it leads to the optimum quest reward, for writing clever backgrounds, for involving others in their RP, for entertaining each other and telling smart stories. Because as long as we don’t reward it, it can never be a mainstream playing style.