GDC 2015 The Delayed Impressions of a Games User Researcher

This year was the first time I have had the privilege to attend the Game Developers Conference in San Francisco, and also importantly as a Games User Researcher the associated Games User Research Summit. I had been lucky enough to be selected to talk and somewhat naively I had vague ideas of writing up my impressions of attending the conference on a day to day basis.

Welp, thanks to the wonderful people I met there keeping me busy into the early hours every single day that (happily) never happened. But before these memories fade I wanted to get some of my impressions down. These impressions are based on my notes and the notoriously bad recording device of human memory. So apologies in advance if when describing a talk I get something wrong. Please feel free to correct me. The content below was first posted on my blog on a day by day basis, but has been combined for this post. Where possible the titles below are linked to the talk on the GDC Vault. 

Day Zero – Sunday – Hello San Francisco

This is not directly conference related, but the first day there (Sunday) was mainly spent exploring the city of San Francisco. Highlights were enjoying a long walk out to Castro Street via as many comic and RPG gaming stores as I could find. Then taking it easy by heading back through town to the waterfront, until finally registering at GDC where I had found that I had made it big with my name spelled out in cardboard for all to see at the speaker registration.

A-L, Lewis-Evans – Z. My name up in cardboard, I had arrived (and excellent quality photo I know, my camera is broken).

Day One – Monday – A Day of Stories

After a lovely breakfast at The Grove, it was time for talks!:

First up was Computers Are Terrible Storytellers – Let’s Give the Humans a Shot by Stephen Hood from Storium.

As part of the narrative track Stephen laid out his “narrative wormhole” of computer based storytelling where the computer can only accept the limited input that it has been designed to receive. As an alternative Stephen promoted allowing players to improvise and make their own stories. To do so he laid out six techniques to help players ; 1. Sharing, let players tell stories together, 2. Leave whitespace, a gap for players to fill, 3. Creative constraints, give players a framework and structure to work from, 4. Use dramatic economies, stories told with limitations (such as point totals that are spent) on players so they have to make interesting choices, 5. Provide setup, give players writing prompts and they can build worlds, and 6. Provide pacing, set an end point for players to work towards.

This was followed by Narrative Review Process – Getting Useful Feedback on Your Story by Jonathan Dankoff from Ubisoft Montreal.

This talk has already been covered well by Polygon but in essence Jonathan, a fellow Games User Researcher, laid out a method for working to improve narrative in games and put forward an argument as to why this was important (i.e. current feedback on narrative often too late in development, leading to a lack of iteration, the feedback that is received on story is often informal, and to gain stronger buy in to narrative across the development team). The method involved forming teams of 6-10 people, half of which were writers and half of which were other developers on the project (e.g. designers, leads, not writers). This team then had time to give notes on a very early treatment of the narrative beat, which was then collated and analysed, put into a report, presented to the writer, discussed one on one between the researcher and writer, and then discussed in a group with a focus on finding solutions and understanding comments. The success of this method is yet unproven as the games it has been used on are not yet out, but Jonathan was able to say it had helped writers find issues they didn't know were problems, and to work through issues they knew were problems but couldn't solve in isolation. As well as increasing buy in across the team and allowing for more early iteration to occur. Sounds like a win to me.

After lunch it was time for Strategic Community Communication – Managing Your Channels Without Running Aground by Maurice Tan from Deep Silver.

In this talk Maurice covered the difficult process of choosing which social media channels to spread your precious time and resources into. With the assistance of plenty of humour Maurice laid out the importance of focusing on the utility a channel will give you, the goals you have for the channel, the role the channel will play, and understanding your own resource limitations. He also stressed the importance of monitoring success, but not letting metrics rule your life. His final message was that it was important to be mindful, to not spread yourself too thin. Not just for yourself or for your fans, but also for your organisation. It is important that organisations buy into and understand what a community manager can do for them, and without this support communication is much less likely to be successful no matter the channel.

Next was Meaningful Play: Monetising F2P Without “The Pinch” by Scott Rigby from Immersyve.

To me this talk felt a bit like a sales pitch, and I don’t really find self-determination theory particularly compelling. To me, self-determination theory is a nice idea that has its merits, but its proponents tend to see the world as very black and white (or as very intrinsic and extrinsic perhaps?). HOWEVER, this aside, Scott argued that some F2P models can make players feel shamed and duped in the name of short term profit, which I can agree is unlikely to be very productive in the long term. Even if I am less certain that it is because specific “needs” are being violated. That said, autonomy, relatedness, and competence, which Scott promotes as basic psychological needs, are motivating for sure and there is certainly more room for games in the F2P space to grow to help meet (for example, by providing opportunities for players to be altruistic, cooperate, and for others to benefit when others spend in a F2P game), rather than suppress (for example by taking advantage of spenders by increasing prices after several purchases) these desires. He also discussed the fact that metrics can tell you what is happening, but not necessary why things are happening (e.g. there may be some time lag between motivation being lost and drop offs occurring in metrics). Essentially Scott’s message came down to the fact that players do not like feeling like they are out of control, that they are being controlled, or generally being screwed with. As Scott says, players are not dumb, and in the long-term treating players poorly will hurt F2P as a model.

I then had to meet some people, so I missed the last talks of the day and instead checked out some of the games in the West Hall. The train jam games in particular were interesting to check out and see exactly what could be done in 56 hours on a train. Please do check them out on the website, they are good fun!

Day Two – Tuesday – Games User Research everywhere!

On day two it was time to get up early and get a bus out to the Games User Research Summit held in Sony’s impressive offices in San Mateo. The Games User Research Summit is an annual opportunity for those interested in Games User Research to gather, network, and share their experiences (assuming they can make it to San Francisco that is). The summit was well attended and organised and included two tracks. As such, I wasn't able to see everything, but I did manage to catch a bunch of interesting talks. I should also say that you can see the official photos on Flickr, which were taken by Tiffany Young:

First up was a welcoming speech from the appropriately purple clothed Jordan Lynn (Games User Researcher at Volition and an organiser of the summit), which was to the point and well put. Then it was on to Rage Quitting: Predictors and Long Term Impacts by Elizabeth Schmidlin from Sony.

Elizabeth shared the results of a large scale survey she had carried out among undergrads at a Texas University looking at reasons behind rage quitting. Some of the headline findings were that men were more likely, even when controlling for exposure, to report raging in FPS games, whereas women were more likely (again when controlling for exposure) to report raging in racing games. The biggest determination of rage quitting due to in-game features was reported as players feeling like they were not getting enough out of game compared to the effect they were putting into it. Followed up by getting stuck on puzzles. Whereas, in terms of more technical features a game feeling like it wasn’t responding as expected was the other most commonly cited reason for raging. Finally, social features could also lead to raging with the three most important features mentioned (tied in terms of importance) were another player being overly aggressive, another player cheating, or another player being perceived as just too good.

UX CYA FTW – Triangulation Flexibility & Integration: Making Bioshock Series Awesome by Bill Gardner from Xperienced Points.

In this talk Bill covered his time at Irrational Games and how Irrational handled user experience (UX). Going from System Shock (before he joined) all the way to Bioshock Infinite. Interesting titbits were that 8 months out from shipping the game an external playtest was run, only to find that players were scoring the game three out of ten and reporting that they had no idea who they were, who the robots were (they had no good understanding of the Big Daddies), and why they were in the ocean. Players, apparently even referred to the game as a cheap Half-Life 2 rip-off. Based on this feedback the plane crash was added and the introduction to the Big Daddies that is in the game now was also added.

Similarly, while there was more focus on testing earlier in Bioshock Infinite the introduction was initially not hitting with the players who found there was no sense of urgency. Based on this the note and corpse were added to the Lighthouse, therefore increasing the feeling of tension at the start of the game. Interesting stuff. He also discussed internal arguments over keeping the controls from Bioshock or moving to an Aim Down Sights model. These control methods were apparently playtested and it was found that while the aim down sights model didn’t impact on vigour use (a predicted outcome) it did lead to people standing still and aiming more often, an outcome the developers were not happy with. Finally, Bill discussed the Skylines in Bioshock Infinite as being an exciting feature, but one that players struggled with due to the high cognitive demands of maintaining control, shooting, and navigating the skylines at the same time. As such, a Skyline Playground zone was built to test with and allow fine tuning of the mechanic. In the end he wrapped up by stressing the importance of user research at developers, recommending starting with expert reviews and internal playtests, and then moving to external playtests (but not waiting too long for this).

Recruiting Tactics for a Modern Audience and Lab Space Design Panels.

These panel sessions were a joy for a methodology nerd such as myself. In them the different recruiting methods for getting participants for user research were discussed and the hassles and victories that can come with setting state of the art user research facilities were examined. Hint, phone screen players to make sure they have the knowledge and experience that you need, and consider the noise that can come over cable trunking. Oh and OBS was rightfully unanimously praised as being awesome.

More Brains: GUR in PvZ Garden Warfare by Veronica Zammitto from EA.

The message from Games User Researchers is, rightly, test early and test often. Do it, all the time. Playtest. Please. As soon as possible. Like when you have paper and/or digital prototypes. But that doesn’t mean that there is no value in testing later in production. In this talk Veronica focused on the user research carried out in the last 6 months before the launch of PvZ (with playtests being run as late as two weeks before launch). This testing was mainly focused on map and class balance and Veronica stressed the importance of using large sample sizes, with longer play sessions and a mixture of repeated quantitative and qualitative measures across a long play session. Asking players as a day of play went on how motivated they were to continue playing, how engaging the game was, and how much enjoyment they were getting as well as looking at their behaviour. She also stressed that user research at this time should be about fine turning and needs to be focused to address issues that can be fixed relatively quickly.

Know Your Enemy - Combantants in Destiny by Jennifer Ash from Bungie. 

In this talk Jennifer discussed the work she had carried out on Destiny focused on the combatants in the game. She outlined the early work where there was a focus on players evaluating concept art and looking at images that were presented to them rapidly (for about 5 seconds) to see if enemies could be spotted amongst various backgrounds and if their design gave appropriate indications of their threat and faction/race affiliations. There was also a focus on how enemy weapons looked, the threat they communicated, and on what abilities and weakness players expected different enemy types to have. Jennifer then discussed some of the subjective measures used to gain deeper understanding of player perceptions. Many of which she said were tracked over year long periods, allowing comparisons to be made across tests as the game developed. These including general questions about how combatants behaved and interacted with the player, indications of how dangerous enemies were, and how players approached bosses and how satisfactory they were to fight.

She also discussed the objective data that was collected, such as event based data on when players were killed, how, by what, etc. One interesting case study from this was the “Shank Flank” where it was found that the low powered Shank enemies in the game were getting too many kills. This was, they found, because Shanks would spawn behind the player and stay there. Then due to their low damage players weren’t noticing and Shanks were finishing the players off when they were engaged with more powerful enemies that they could see. This was fixed by changing the Shank spawn points (World Design) and changing their behaviour to attack from the front (AI).

Destiny User Research Q&A featuring Jennifer Ash, John Hopson, and Nick Hillyer from Bungie. 

In this panel three user researchers from Bungie faced a barrage of questions (I may have snuck in two…) about their processes and various factors in Destiny including, but not limited to, the Cryptarch, loot drop balancing, class balancing, colour blindless, simulation sickness, Phogoth’s tummy, the move to aim down sights, the surprise of how the community reacted to the Iron Banner, and how they work in terms of carrying out user research within Bungie.

User Profiling by Anders Drachen from Aalborg University.

Anders gave a comprehensive presentation on how to start to deal with the masses of player data that is often being collected. Specifically, he outlined four steps: 1. Knowledge discovery, 2. Selection, 3. Interpretation, and 4. Application. He stressed that you must go through these steps and that it was cyclic process as player behaviour changes over time, as do games, which echoes into the data. He then discussed player profiling, mentioning bottom up profiling (or “explorative profiling”) where you work from data coming in, which can lead you open to biases. The alternative is top down profiling, where you go out with specific hypothesised profiles and see if they work with the data. Anders advocated for the idea of building “protean” profiles from day 1 (e.g. persona like profiles) and then following players through the whole life cycle of play (from start, until they abandon the game).

Anders also made sure to point out that data profiling is not an objective process. This is because as a profiler you have so many different choices to make about processing data and what you chose will change the outcome. He also stressed that you can get pretty far even with just descriptive statistics and segmentation of player groups. Anders then went on to cover some other methods such as clustering. All and all his talk was jam packed with details.

At this point, my iPad, on which I was taking all my notes crashed. As such, from this point on (and for the rest of GDC) I am pretty much relying on my memory. So, if there are less details it is not that the talk wasn’t interesting, it is just that memory is not as good as note taking (not even close).

Dissecting the Dragon: UXR for Dragon Age: Inquisition by James Berg from EA.

For this talk James took us through some of the challenges and situations that came up when carrying out user experience research on Dragon Age: Inquisition. These included working on a good plan to make sure a diverse and LGBT recruit would be carried out, and then finding that due to Bioware’s relatively inclusive history in this area that a normal recruit could just be run and would result in a diverse crowd coming in to test the game anyway. James also stressed the difficulty of carrying out user research in such a big, open game – which was also still under development so constantly changing. He cited one specific example where a player spent three hours just wandering around in featureless, content-less mountains and generally having a great time. Great for that type of player, but not so useful if the playtest session was to test out class balance or some other mechanical aspect of the game. Much like Bungie with Destiny, concept art was also tested in Dragon Age. In this case they looked at how recongisable the characters where, romantic interest, enemy threat, and so on.

James also covered how DA:I originally had a Skyrim like compass navigation system, but users just couldn’t use it, leading to the (much, much better in my opinion) system that is in the game at the moment with the bottom mounted radar. Other interesting nuggets of information were that about 25% of players use tactical mode all the time, with most others only using it where necessary, that bears were the bane (and ultimate opponent) of many playtesters, and for a long time Solas had text to speech for his lines, which some players thought suited him.

After a little bit of housekeeping from David Tisserand (current head of the IGDA Games User Research SIG and researcher at Ubisoft), it was time for the keynote UX Invaders: We come in peace! How we can collaborate with developers to frame “fun” by Celia Hodent from Epic Games.

She had the best t-shirt of the day and discussed different ways to approach user research and getting developers on-side. This started off by addressing common misconceptions, like user research is just “common sense” (no such thing, this is hindsight bias!) and that the claim that Dark Souls shows that user research isn’t needed misunderstands the process. User research isn’t about making games easy, it is about making the best games possible for a certain target audience. As such carrying out user research on Dark Souls would be entirely possible, you just have to consider the audience, and in this case make sure the game was punishing enough for them. Ultimately, Celia was stressing that user experience is not just the concern of user researchers and user interface designers, but that it is about the whole studio achieving their goal. A sentiment that you just can’t disagree with.

She then went into her usability plus game flow = good user experience model. To do so she went through the heuristics she uses for usability (Signs & Feedback, Clarity, Form Follows Function, Consistency, Minimum Workload, Error Prevention & Recovery, and Flexibility) and for game flow (Pacing, Motivation, and Emotion). Celia provided clear examples for all of these, calling mostly on the work that she has been carrying out at Epic on Fortnite. The talk in general reminded me of listening to a lecture in Human Factors class, and I mean that in a good way (I know some people had bad experiences at University, but I loved it and had great lecturers). Celia finished up her talk by outlining four steps for collaborating with developers in user research: 1. Debunk UX misconceptions, 2. Insist on testing early, even with paper prototypes to demonstrate early, easy, wins, 3. Come up with a “model” (her use of quotes, not mine) that can be communicated and understood by developers in order to get buy in, and 4. Put user experience at the centre of the studio, as mentioned earlier it is up to everyone, not just the user research team.

After that it was time to get back on the buses, head back to SF, and to party until late with the wonderful user research community. Good times were had.

Day Three – Wednesday – Get to talks 30 minutes early or you miss out

I really wish my iPad had been playing nice on this day, because on day three I made it into several extremely packed talks (I turned out 30 minutes early, but many others were turned away). These talks were packed because they were interesting, and thus it sucks I don’t have any notes. Anyway, again, these descriptions and impressions will again be mostly based on my memory.

Hearthstone: How to Create an Immersive Use Interface by Derek Sakamoto from Blizzard Entertainment.

Derek delivered a well-paced, laid back, interesting, and humorous (some really great groan worthy puns were made) talk about how the user interface for Hearthstone was created (an on-going process that was carried out over the six years of the games development). This can apparently already be watched online as part of Gamespot's coverage. He shared some really early (and very different looking) screens of the early user interface for the game. However, the main thrust of the talk was about finding the “seed” for the interface. The idea that everything flowed from. In the case ofHearthstone this seed was the idea that the game was played and held inside the “box”. This seed took about a year to come up with and many other decisions flowed on from this. For example, the question “well, where is the box then?” lead to the idea of the game being played in an Inn, in turn leading to the introduction UI for the game and other associated “Inn like” touches.

Furthermore, it lead to a commitment to a “physical” design for the game and the UI. Interestingly, this was not just aesthetic, it also impacted on the mechanics of the game. For instance, the fact that there is a played card limit in the game arose from not wanting to shrink down the cards to make room for more – yet that limit also forces players to consider their moves and restricts strategy in interesting ways. Similarly, the deck size used to be larger (60), but that wasn't very workable in the UI. So it was reduced.

That is not to say that aesthetics aren't also important for the experience. In fact, Derek shared that his child’s idea of “playing” Hearthstone is to get him to load up the Gnome Workshop level and launch the interactive rocket – a “physical” extra element added to the game to serve this UI principle, but that has no mechanic impact on the game itself. It is there simply to serve the player experience and the “seed”. All of this said, Derek did admit at times their desire to stick to a physical design had caused some usability issues with multiple steps being needed to perform some common actions that could be done faster. However, he also stated that this would be addressed while staying true to the “seed”.

Clash of Clans: Designing Games That People Will Play For Years by Jonas Collaros from Supercell.

Another packed out session had Jonas, a programmer on Clash of Clans laying out the reason for their success. To start out Jonas clearly stated that they had been lucky. He then moved on to discuss the other factor that Supercell believe is important, their commitment to focusing on the player experience and designing for long-term play. To do so he outlined the history of Clash of Clans development, from the 5 person initial team to the 13 (yes, 13!) people that currently maintain the game. Over, and over, Jonas clearly stressed that the team working on Clash of Clans is always looking into the long term and not at short term gain and expansion. As part of this he, right near the end of the talk after not mentioning it otherwise, stressed that monetisation should always come last and player enjoyment first. To back up his claim that Clash of Clans focuses on long term engagement, he shared that 1 in 10 people that would log in to Clash of Clans that very day would have been playing for two years.

Essentially, his message was focus on making a really good core experience for your players, make sure it works, and only then add in things like replays, clan wars, heroes, and so on. This is a sensible message, however, I can’t quite help to think that because of the success of Clash of Clans that doing this has become very hard for anyone looking to follow Supercell into the Base Builder genre. Certain features, which Clash of Clans added later, are now expected as standard – making it much harder for developers to start focused and then grow. But, perhaps, this is also part of what has made Clash of Clans so dominant. Clash set the bar, and did so slowly and carefully, but now expectations are raised.

User Research On Destiny by John Hopson from Bungie.

Having already heard about user research on Destiny at the Games User Research summit, you would think this held no appeal. But not so! It was again a “get there 30 minutes before affair” and John shared many interesting bits of information about the user research effort on Destiny. My iPad came back to life for some of this session, and then crashed again, but I did manage to make some notes. To start with John talked through the history of user research on Destiny, a history that began four years before launch, a much longer period of user research than had been carried out on Halo titles in the past. As part of testing so early the user research team had to convince the rest of Bungie that nothing would leak and build trust within the studio. As such, all players had to go through metal detectors (to detect recording devices and mobile phones) and essentially players were constantly monitored and they were never, ever, left alone with the game. Destiny was also a new challenge for Bungie, having primarily worked on relatively linear experiences with clearly separated multiplayer and single player components in the past. Since Bungie wanted to bring along the fans of Halo, but also appeal to fans of other genres (e.g. RPG’s), this presented a challenge for the user research team to recruit and understand their players.

This lead John to explain Bungie’s recruiting process, which, I have to be honest, I am jealous of. Basically, since they sign people up via Bungie.net they have access to those players Halo gameplay data. This means they can, without having to only rely on subjective report (which is still used, via the signup survey, etc), see exactly what type of Halo player each person that signs up is. Meaning that if they wanted to recruit only individuals who never touched multiplayer and only played campaign over and over, they could. This, of course, raises the usual issue of the danger of only testing with fans (i.e. fans are biased and may not reflect the general buying population). However, John stated that they had done comparisons between their bungie.net recruited players and players recruited via outside agencies and found no significant difference in terms of the results during testing. They also used weighting based on the general population to further control for any biases. All and all this recruiting process lead to them being able to recruit very quickly and with a good certainty that they were getting the types of players they needed.

John then went on to talk about the clustering methods they used to identify “types” of players based on Halo: Reach data. Saying that the most interesting thing about this wasn't the clusters that they settled on, but rather how tweaking categorisation variables would alter the clusters (again, the methodology nerd in me has to agree). However, the three (or five) categories they ended up with were Campaigners who tended to just play campaign, Omnivores who did either a little or a lot of everything, and specialists who fell into many subtypes. It was decided that Destiny would be designed to encourage Omnivore-like behaviour, basically the idea was to provide a variety of options for players and encourage them to try their hand at everything (thus the bounties, etc in the game).

Among the other things that John covered were the fact it was not expected that people would mainline the story missions the way they did during testing, leading to the need to rebalance the game. Furthermore, John stated that in general the fans were about 50% faster in completing the content at launch that Bungie had estimated and that one are the Games User Research effort had missed focusing on the transition at level 20 from xp leveling to gear leveling – stating that around 4% of players in the launch game appeared to get stuck here (a small but significant number). John also said that Bungie was surprised by how people focused more on random drops than the loot from marks and vendors. Given John’s most famous article on Gamasutra (the article that in fact alerted me to the fact that Games User Research exists) is on Behavioural Game Design, I am personally a little surprised that they were surprised by the appeal and power of the variable ratio schedules associated with random drops.

This is getting long (it was an information packed talk), so I will just mention one last thing that John showed, this was a way for players to provide emotional feedback to the team without having to pause the game. The players would hold the left hand side of the touchpad on a PS4 controller and then the face buttons could be used to say that a moment or situation was confusing, awesome, etc. This was then overlaid over a map of a level and designers could go in, click on an “awesome” (or whatever) and watch the moment from the video in the playtest in which corresponded to the rating. This was good for getting buy in from the team, but also as John says we must remember that while Games User Research often seems to focus on the failings of a game, it should not be forgotten that it is also about the AWESOME.

The next talk was another packed session, which was good to see, but also somewhat sad as the topic was Game Developer Harassment: How to Get Through a panel featuring Nehia Nair (Storm8), Donna Prior (Green Ronin Publishing/Salish Events), Zoe Quinn (The Quinnspiracy), and Elizabeth Sampat (Optical Binary).

I say sad simply because harassment shouldn't be problem, but it is, so it is fantastic that this panel was on the schedule and that so many turned out to see it. My iPad once again stopped working, which is a real shame as this panel deserves to be summarised way better than my memory allows – so please head here, here, and here to read some of the more detailed summaries of this event. Seriously, please, go read, and anything else you can find on this talk. Go watch it on the GDC Vault, it is free. Even if you don’t come back and read anything else I write. Go. Now.

Still here? Ok, what I can say is that each member of this panel got up and spoke with passion, emotion, and eloquence about their own extensive experiences with harassment and then gave guidance on how to deal with it; from the personal of stressing how important it is to have support networks that respect you around you (meaning that they know when to help and know when to just be there if you want help) to the more technical in terms of outlining the steps (e.g. two step verification on everything, etc) you can take on the internet to minimise the impact of harassment (or do damage control if it has happened already). There were also important lessons for the industry brought up, with Donna Prior stressing the need for game companies to have policies in place and to support their staff if harassment occurs. Again, please, go read those other write ups if you haven’t already. Game developers should not have to deal with this crap and as an industry (and as fans) we need to do something about it. It is not appropriate to shrug and say “that is the internet/that is games”.

After that it was time for me to head off back to my Hotel. I was giving my talk the next day, so I need to be ready, plus I had three parties to go to. What follows was a businessy time at the UK Game Industry Drinks, a funky time at the Wild Rumpus, and an unreal time at the Epic party. Followed by, very responsibly, getting back to my hotel so I could sleep and be ready to present on day four!

Day Four – Thursday – The day I talked

I didn't quite make it to as many talks on day four, because I gave a talk of my own and suck around for some post talk discussions. But here is what I did see. My iPad was still broken, so these impressions are still mainly based off my memory.

I didn't see anything in the first session in the morning. Which was a pity as there are definitely sessions there that I will try to check out on the GDC vault. Rather I got up relatively early, had a relaxing breakfast, checked out the indie games in the West Hall, went to the speaker prep room to check my presentation one last time (a nice productive space with bagels, cream cheese, and cool, cool, refreshing water), and then got to my talk the requested 30 minutes before it was time to go. My talk was on Anti-Social Behaviour in Games: How Can Game Design Help?.

I’m not going to talk about what I said there that was interesting, please ask someone who went as their impressions are more valuable than mine, or go watch it yourself on the GDC Vault (no membership needed). But I do want to talk about the experience of being a GDC speaker. First of all it places me in an even more privileged position that I normally enjoy (pretty is pretty privileged anyway), got me free access to GDC, free lunch every day, access to the speaker prep room, and also the ability to wear a badge that said “speaker” on the bottom, which gets you occasional interested squinting from random attendees. Secondly, it was all very smooth. I had all the information I needed and when I did get to my session I was well briefed by the technicians and the volunteers. Very nice. As for giving the talk? Well that was exciting and great fun. I used to be a University lecturer and giving talks and teaching people, is something I miss from that life. So to get up and talk about a topic I think is interesting and important is always a wonderful and fun feeling, and opportunity. The final thing I want to say is that the debrief room idea is a clever one that more conferences should copy. It was good to be able to say “I will be in the debrief room” and then have people come and ask questions or catch up after a talk. A really nice opportunity to grab speakers and learn more.

For the next presentation I again got there 30 minutes early and it was another packed house for More Science Behind Shaping Player Behavior in Online Games by Jeffrey Lin from Riot Games.

This is the second talk that Jeffrey has given on this topic at GDC, the first being The Science Behind Shaping Behavior in Online Games, which can be (and should be) watched for free on the GDC vault. In this talk Jeffrey focused very clearly on the importance of the (lack of) feedback mechanisms in solving the issue of toxic/anti-social behaviour in games. Feedback was something I had stressed as an important part of the picture in my own talk and I fully agree with Jeffery when he says the problem with the internet, and with games, is that there is basically no feedback loops to let people know that toxic/anti-social behaviour is not ok. This essentially makes it the norm, or even worse, makes it the encouraged and rewarded way to communicate.

As part of this talk Jeffrey shared the results of the development of a machine learning algorithm that Riot has developed based off player data. This algorithm is now able to automatically recognise toxic behaviour (as defined by learning from the reports and communication of players) and enforce appropriately without having to wait for the relatively slow process of player reporting. This helps met the need of feedback to be swift and certain, greatly increasing its potential to work to change behaviour. As part of this Jeffrey shared some amusing anecdotes about a player flaming himself for poor play in chat, resulting in an automatic ban, and that while “bronze scrub” (the lowest league in LoL) is detected as insult in most of the world that “silver scrub” is detected as an insult in South Korea.

Jeffrey also talked about the power of so called “Shadow Bans” where users on communities such as Reddit can continue to post, but their posts are not seen by anyone but themselves. This means that these shadow banned individuals may not realise they are banned and instead it may look to them like they are just being ignored, which denies them the positive feedback of a reaction. Jeffrey stated that toxic behaviour can be contagious (remember, just seeing it and seeing no negative consequence for it gives the feedback that it is ok and fine to communicate like this) and so by shadow banning just a few high profile consistently toxic accounts a big impact on the negative tone of the League of Legends subreddit was able to be made. As in his first talk, Jeffrey stressed the fact that the people at GDC, as the designers and influencers of games, have the power to do something about this lack of feedback and to reduce the amount of anti-social behaviour occurring in their games.

More Than This World: A Design Retrospective on SID MEIER’S CIVILIZATION: BEYOND EARTH by David McDonough and Will Miller from Firaxis Games.

Without notes I have to admit that this talk was a bit of a blur. Likely because the week was catching up with me and also because I personally find double act presentations difficult to maintain focus on. Still, what I do remember is that this talk was remarkably open and frank. Time and time again the developers openly admitted that they had been too conservative with their initial design decisions when working on Beyond Earth.

The developers stated that the initial thoughts were that they could take the tried, true, and successful formula of Civilisation V and follow it when appropriate when creating Beyond Earth. However, this unfortunately failed to account for the significant framework in terms of understanding and meaning for players that a grounding in history gives Civilisation. Without this, things like the faction leaders and the wonders in Beyond Earth just didn’t have the impact that was intended as they didn’t have the familiar historical framework to rest on (and the pre-packaged emotional reaction and meaning that comes along with that). Essentially, there was no Gandhi nuking you moments. The wonders system in particular was singled out as just not feeling very wondrous, something that Firaxis are working to address via updates to make each wonder feel unique and like it really impacts on the world.

The developers also laid out the challenges in adjusting something like a tech tree to future tech. This has the problem, which they did realise from Alpha Centauri days, of technologies being unfamiliar (e.g. everyone knows what “horse riding” is but “Xeno-domestication” may not be immediately clear). In order to get round this they created the technology web, which with its UI and groupings tries to use familiar language and clearly outline the benefits of each upgrade. Which, having played the game at lauch, I think they did well.

Efficiency for Game Designers: Lessons Learned Making Gunpoint Quickly Without Going Mad by Tom Francis from Tom Francis.

Tom’s talk was delivered in a very calm, relaxed, and laid back manner. He gently moved through topics like how to be your own boss (basically don’t be a dick, even to yourself), how to manage workload, and how to decide how much effort into putting into an idea when developing a game. This last part basically boiled down to asking if you think a feature would be good and how long you thought the feature would take to implement. If you think the feature would be good and wouldn’t take much time, then, of course do it. If you think it would be good, but not are not sure how much time it will take (or think it may take significant time) then give it a day, see how it goes then make a decision. If you think it will be quick, but not sure if it will be good, then just give it a go. Then finally, if you don’t think it will be good or quick, then forget about it.

Tom also spent some time talking about how to prioritise and get over being a perfectionist. His tip for this was to put everything in to a ranked order list and only work on the stuff at the top of the list. Be constantly asking yourself “am I doing the most important thing I could be? And if not, why not?” He said that by keeping a list like this you can keep on top of your inner perfectionist by saying “it is on the list, I will get to it” even if this turns out not to be true in the end. All of this advice was interspersed with examples from Gunpoint and of Tom’s upcoming game Heat Signature. It was an interesting talk, for which again I wish I had notes.

The Neuroscience of Game Audio by Zak Belica (Epic Games) and Seth Horowitz (NeuroPop).

I have mixed feelings about this talk. First of all, it was another double act, which unless executed brilliantly really does break my focus and often weakens the message. Secondly, the content that was presented by Zak, who obviously great at game audio, was quite interesting. He showed how certain sounds could be layered together to change their quality and emotional impact (using a monster from Infinity Blade as an example) and how sounds known to cause discomfort (e.g. fingers on the black board) in some people (this sound has no impact on me at all. Nothing. But that is the thing with individual differences) can be used in monster/enemy noises in Fortnite. There was an example of sound overload situations from Gears of War, along with steps that could be taken as a sound designer to get around this. Also, he gave a nice demonstration of how perception is based on expectancy based on sounds by showing a bugged reload animation looking fine (to me at least) thanks to good clear reload sounds. He also suggested that the emotional impact of an emote (such as /dance) could be enhanced by adding audio of the person singing over top, which is true, sound is strong, although I would suggest that care should be taken to not over use it due to its strength. What might be a positive emotional response to singing along with an emote, may become negative if a player is exposed to it too often.

Ok, so that was pretty much all good. So why the mixed feelings? Well, the fact is that the other presenter, Seth, didn’t really hit home with me. This was not because he was a scientist and that he was “too academic” or anything like that. I am a scientist myself, with a PhD in Psychology, and I give scientific talks. Rather, it was because the parts he presented where often overstated, overgeneralised, and inappropriate/inaccurate in parts. Probably the easiest of these to call out in terms of inaccuracy is making a deal about there being 6 senses not 5. Seth did this even to the extent of dismissing a member of the audience who (rightfully) was pointing out that there are far more than 6 senses. He also, used a visual illusion on his presentation that (as he actually explained as he was showing it) relies on saccade motions of the eye in order to function. The thing is that this means the illusion has to be close enough that eye saccades are actually a factor, meaning that it does not work at all when up on to a presentation screen that is far away from the audience. Another example would be contrasting the cocktail party effect (hearing your own name amongst background noise and having it grab attention) vs visual search in Where’s Wally/Waldo (looking for a specific red/white target amongst distractors). This is an inappropriate comparison as one task is largely conscious search (top down attention) for a personally irrelevant stimulus (Wally/Waldo) versus pre-conscious bottom-up attentional processes targeting a personally relevant stimulus (your name). This stuff might seem small, but it was significant in terms of undermining his expert authority.

More importantly, his talk time and time again overstated and overgeneralised the importance of auditory feedback. He basically ignored the problem of individual differences, and ironically given that he mentioned it in terms of the quality of the audio setup in the presentation room, the problem that most of the more extreme effects he was describing rely on a very specific audio setup. This specificity and lack of generalisability of many impacts of audio is a big problem for using them to get certain effects in games. People who play games are diverse (both physiologically and in terms of the subjective affect they will associate to different sounds) and developers have basically no control over players sound setups and the environments that they are playing in (or even if they are playing with sound on at all!). This is a serious problem for trying to take advantage of the more niche effects that Seth was often broadly attributing to audio stimulus. This doesn’t mean you shouldn’t have sound or learn from the science around sound, but at the same time it doesn’t help to oversell and over generalise the impacts. Personally, I came away from the session with a feeling that if Zak had just presented himself that it would have been a much more satisfactory and informative session.

It was then time to head to the hotel, drop off my stuff, and make for the GDC Speakers Party. Thespeaker’s party was another privilege of presenting at GDC and was a wonderful night of chatting with interesting people from all over the industry. Then, it was on to the last day of the conference.

Day Five – Friday – An accessible end

On the last day of the conference I attended only a few talks. Instead took the time to hit the expo floor and visit the various stands around the conference halls. All and all it was a relatively laid back day, but also one that was bitter sweet. I wish I could have cloned myself to see more talks, both today and the rest of the week. GDC was an amazing experience and it was sad to know it was ending. I feel extremely lucky to have been able to attend and even more lucky to have been able to talk. I can only hope that I can come back again.

The expo floor was filled with server providers, publishers, and advocacy groups. There was a lot of VR, quite a lot of motion and face tracking, a decent number of haptic controllers, and generally speaking busy, but happy looking, people everywhere. I gave in and stood in a line for an hour to see the Smaug demo at the Epic booth. Which had an impressive sense of immersion and certainly looked pretty. I also made sure to get a photo of the booth of academy award winning SpeedTree (great name as we all know) and to try out a couple of the aforementioned haptic controllers (nice gimmick). I also wandered past the Valve booth looking like a hopeful lost puppy. But, no luck there other than a “Hi” from the guys there that had obviously ignored many puppies all week long.

Carnage Controls

Controlling Carnage: Effective Side-Scroller Controls on Mobile by Kayla Kinnunen from Roundhouse Interactive.

This was a well laid out and practical talk by Kayla covering the progression of controls in the Warhammer 40K based mobile title Carnage. The talk included clear diagrams such as the “comfortable reach distances” for thumbs at the edges of both screens on an iPad, and how this was used to make sure that commonly used controls were more easy to hit than uncommon (yet still important) controls. The talk also covered the importance of seeing real players playing your game and learning from this. Finding out that buttons, which have to be precisely hit, are problem as peoples thumbs tend to move around on touch surfaces without tactile feedback to guide them. In answer to this hit “zones” were introduced instead. Where the visual button acts merely as a guide as to where the zone is, rather than the precise place on the touch screen that needs to be touched. Overall it was a practical and well laid out talk about the processes behind developing a mobile control method via an ergonomic, game mechanic, and player focused approach.

No More Excuses, Your Guide to Accessible Design by Tara Voelker from Ready at Dawn. 

Tara’s talk was a nice way to finish out GDC with Tara laying out some guidelines to follow if you want to make your game more broadly accessible to a wider range of people with different capabilities. This is important because, well, think about other humans huh? But if that isn’t enough, consider that one in four people will have a (at least) temporary impairment in their life (So would you like to play games your whole life? That could be you with the impairment), and that according to a survey by Popcap impaired gamers tend to play longer and be more dedicated consumers. Thankfully my iPad mostly behaved, so here is the list of most of the things you should do:

And with that, my GDC was over. As I have already said it was a fantastic time and I am well aware of how lucky and privileged I was to attend. It was a well organised conference, with excellent speakers (so many to see on the vault too once they are up!), and filled with wonderful people. I want to go again, of course, but even if I don’t it will be a memory that sticks with me.

Thanks for reading these impressions. I hope that they were useful. As I said on day one they have at least been good for me to refresh and reinforce my memory.