Originally Published in Prehistoric Times (2015)
"Saurian" PC Development Team
Interviewed by Dan Liebman, DansDinosaurs.com
Nick Turinetti - Project lead
Bryan Phillips - Lead animator
Tom Parker - Research lead, environment design
RJ Palmer - Concept artist, Creature design
Alex Lewko - Creature design, research
Henry Meyers - AI programmer
Q. Dinosaurs have been a favorite element in video games for years, though they are typically unrealistic and mindlessly savage antagonists. "Saurian" appears to have been developed with an entirely different foundation, one which portrays dinosaurs as animals in a complete ecosystem. How did this concept come about?
Nick - The typical portrayal of video game dinosaurs as “unrealistic and mindlessly savage antagonists” is actually one of the main reasons I began thinking about what would eventually become Saurian. I have been involved with the development of several other video games that featured dinosaurs in one capacity or another, and each one portrayed the animals as monsters that could easily have been zombies or aliens with a simple model swap. It became obvious to me that this was the limit of these projects’ goals, that there wasn’t the prospect of anyone making a game with the attention to detail to the new discoveries that have been made since Jurassic Park. Upon realizing that, I started investigating what it might take to create such a game. In the process I met several other people who were as passionate as I am about bringing dinosaurs to life in a manner that was consistent with the best available science, and Saurian is the result.
Tom - Every dinosaur game seems to have an almost set roster of Mesozoic animals that you see. While our game does feature Triceratops and Tyrannosaurus rex, we wanted to give some lesser known animals an opportunity to be featured. Our goal is to present a representation of a real Cretaceous community as it was in the past, rather than a random assortment of stock beasties. We chose the Hell Creek ecosystem specifically because we felt it had the perfect mixture of well-known and more interesting obscure animals to function as a cast for a video game.
Q. The game sounds like a paleontology lover's dream come true. How do you balance player accessibility with scientific realism? Without scaring away children, is this a potentially educational game?
Nick - As many of us are both game developers and armchair paleontologists we’d hope it does! Any attempt to recreate prehistoric life involves some level of speculation and informed guesswork to bring the past to life with the same vibrancy we see in the modern world. The All Yesterdays movement has been invaluable to opening many eyes to just how unexpected life could have been during the Mesozoic. Having said that, while we are happy to consider unconventional (but not improbable) hypothesis on their nature, one of the goals we established at the beginning of Saurian’s development was that dinosaurs do not require fictionalized abilities to provide interesting gameplay. Wherever possible, we’ve utilized the fossil record and published literature to guide our design process, including gameplay. Acheroraptor for example is not capable of “cheetah speed”; it’s legs are too short and muscular for high velocity pursuit. Instead, we’ve drawn from the work of Denver Fowler on the Raptor Prey Restraint theory of dromaeosaurid predation. By drawing our game design directly from current science, Saurian absolutely has the potential to be educational, but it isn’t our intent for it to be explicitly so. What I want people to know when they play the game is that every aspect of it has a solid, scientific justification behind it.
Q. The current plan is to allow us to play as the Tyrannosaurus, Triceratops, Pachycephalosaurus, and Acheroraptor. Is there any plan or incentive to get players to try out these different species, instead of just jumping right into the most obvious choice? What gameplay styles will be unique to each of these?
Tom - Yes, we hope that players will want to try out all of our playable animals. We aim for each genus to have very unique gameplay from each other with interesting dynamics that the others will not be able to offer. To start, Acheroraptor is much smaller that the other three playables, while Pachycephalosaurus is something of a middle-ground between it and the giant Triceratops and T. rex. This means fundamentally, each animal is interacting on a different “level” within the environment. An adult T. rex is not going to give much thought to an alvarezsaurid, while for an Acheroraptor this will be staple prey. As for deeper gameplay elements, each playable will have unique features to its play style that make them fun to play. For instance, our Acheroraptor will act as a sort of “thief” class, focusing on stealth and surprise attacks from above whereas Triceratops is something of a “tank”, a solid bruiser focused on counter-attack. Pachycephalosaurus is an interesting one because it offers the perspective of a prey animal, which is not something you see much in games where you play as non-human creatures. I don’t think I need to tell you the draw of T. rex but it will not be a cakewalk, the life of an apex predator is anything but an easy one.
Q. There are a number of non-playable species planned to appear in the game world, as well. How do you select which denizens of Hell Creek will make an appearance? Surely we won't have to worry about something biting our faces when stopping for a brief drink?
Tom - Deciding which animals will be included in the game has been a long and ever evolving process. To start with we picked the specific layers of Hell Creek we wanted to set our game in (the formation spans over a million years). We chose a zone at the very top of the formation called Hell Creek III, which spans the last 300,000 years of the Cretaceous. We then had to pick through the literature to determine which animals were present at the chosen time. This gave us our candidates. It then becomes a process of deciding which animals are most common and important to the ecosystem. We would also exclude animals not noticeable from the perspective of our playable animals, for example small fish or tiny aquatic salamanders. That said we did want to include enough animals to make our ecosystem really feel alive, and we have a sizable roster planned for the final game.
Nick - Actually, I would be very careful when and where you decide to take a drink in Hell Creek, there are several species of crocodilian and other aquatic menaces to be wary of... One of the things we quickly learned when researching the Hell Creek Formation is just how much is known about the entire ecosystem. While the dinosaur fauna is not as diverse as some other times and places, there are extremely well known animals like Tyrannosaurus, Triceratops and Ankylosaurus, as well as some obscure and downright weird animals as well, like Anzu and our recently revealed alvarezsaurid. Unfortunately the floodplains of Hell Creek were not terribly good at preserving complete skeletons of small animals, so while we have a good idea of what was living there, there usually isn’t sufficient material to get a good idea of the animal’s appearance. So in addition to selecting animals that are relevant to our playable dinosaurs, we also focused on animals we could reconstruct with some degree of accuracy, either because we have enough bits or we know they were very similar to a better known relative.
Q. Have there been any similar simulation titles that influenced the design, and what resources and people have proven to be most helpful along the way?
Bryan -There have been a couple simulation titles that have been influential in Saurian’s design, including the museum exhibit Be The Dinosaur simulation, Odell Down Under, Sanctuary Woods’ Wolf and Lion, and BBC’s Walking with Dinosaurs’ “Big Al Game”
Nick - We’ve been exceedingly lucky to work with numerous talented and very knowledgeable people, both on the technical side and on the conceptual and research side of Saurian’s development. We’ve gotten amazing support from Nick Canafax of Rival{Theory}, the developers of an awesome AI system called RAIN that provides the framework for our dinosaur’s behavior. We’ve also had great feedback from several paleontologists, including ceratopsian and stratigraphy expert Denver Fowler, locomotion expert Dr. John Hutchinson and Hell Creek/Cretaceous mammal expert Dr. Gregory Wilson. I also want to thank the members of the Hell Creek forum and Skype group (http://z13.invisionfree.com/Hell_Creek/index.php?) for their major support in securing many valuable references as well as being a sounding board during our early development. Thanks Stygians!
Q. What are some of the biggest hurdles you've faced in developing "Saurian" so far?
Nick - We’ve had a few hurdles over the course of our development. A perennial one is lack of funds. None of us are paid or working full time on the game, Saurian development happens in our spare time. We’re students, have full or part time jobs and have commitments just like anybody else, and all things considered I think we’ve done an amazing amount with our limited time and resources. Having said that, we are working towards a crowd-funding campaign to try and free up key team members to work on the game full time, and I’m very confident that once we have everything ready to go, we’ll secure what funds we need to bring the game to the public. Another major hurdle has been finding references for more obscure parts of the Hell Creek ecosystem. Flora has been particularly vexing. Fossil leaves and other parts of plants from the Hell Creek ecosystem were first discovered and named back in the late 1800s, and many leaves have been assigned to living plant groups based primarily on shape or general form. As we develop better tools to classify and analyze fossil plants, it has become apparent that most of the plants in Hell Creek became extinct 66 million years ago at the end of the Cretaceous and either left no descendents or are only somewhat close to living groups. We’ve managed to piece together our interpretation of the Hell Creek flora with hundreds of hours of research. If anybody knows a paleobotanist with some experience in this area, send them our way, please!
Henry - Maintaining a creative democracy with a group of volunteers over Skype, bound together solely by their passions for dinosaurs, games, and science. We’re doing remarkably well on that front, though; it’s been really inspiring. Also AI, but I address that in another answer.
Q. What type of mood and atmosphere are you hoping to convey to the player, and how do you go about this?
Tom - Most of all, we would like to invoke a feeling of chilling tension, everywhere in this environment should feel beautiful and tranquil yet full of danger, with the possibility of a 6 ton predator bursting forth from the trees at any moment. Bryan, our animator, calls it a “dark paradise”. Classic palaeoart that we think captures this mood include the art of Doug Henderson and Phil Tippett’s 1984 stop motion masterpiece Prehistoric Beast.
Nick - Playing the game in it’s current state, I already feel that Saurian’s Hell Creek is a place where you are never completely safe, and that’s a feeling I want to extend to others as well. There’s always something out there that requires a player to be alert to their surroundings. Being alone in a meadow and not being able to see or know where a predator might be leaves me uncomfortable. Being chased by a T. rex is scary, but watching it move through a forest clearing, knowing it’s looking for you and being unable to do anything about it is much scarier.
Q. The character models in the game appear highly contemporary, with vivid color patterns. What sort of role does aesthetic play in a simulation of the ancient natural world, and what sort of visual treats can we expect?
RJ - Have you ever noticed how more limitations can yield a more creative result? That’s exactly the mentality behind my design approach for the animals. I like being handed a set of rules and looking for the ways to push a design as far I can. I cannot express how often I’ve had to ask “is this possible?” but that often points me in the direction of a more unique design. All this science has pretty much ruined my enjoyment of almost all other dinosaur properties because I cannot unsee all that I’ve learned. That’s okay though because realistic dinosaurs are cool as hell. So yeah, expect some visually exciting but grounded dinosaurs.
Alex - It’s really pretty important, and there’s a lot of different things that I consider when working on any design. Probably the biggest one is what the animal was like and how it would fit into the world. That often has a very strong influence on where I might consider going with a design in the first place. I also tend to be very considerate of the animals’ forms, and can make patterns that will serve to compliment them. In addition, having been with the project for a very long time, a theme that has always been prevalent since the beginning is that our creatures should be unique, distinct and recognizable. As a result, I’m constantly trying to come up with designs that feel fresh and new while still respecting current knowledge. We didn’t want the animals to look too terribly dull either, so in many cases I wound up trying to strike a balance with something that has a little bit of flare without going too far with it. The other thing is, we have a variety of people working on this project and naturally everybody has their own idea of what the animals should look like. So, I’m always very mindful of what sorts of things people want to see and always try to explore ideas that incorporate those things. I think in the end you’re going to have a pretty diverse bestiary filled with a wide variety of stunning creatures.
Q. Animation is crucial to convey realism in a digital character. While footage of birds might be very helpful for some species, many of these creatures have no obvious extant analogue. For example, ceratopsids are currently believed to have kept their front limbs slightly sprawled, and digits facing away from the body. How do you tackle the biomechanical issues, and what specific challenges did you encounter here?
Bryan - While we used as much data as possible to influence the animations, we actually went through a long period of trying different things for each gait with the Triceratops. As an example, we tried sprinting animations based on an elephant’s run and a gallop. Sometimes these were small changes and other times this meant scrapping the animation in favor of starting fresh until we found something that looks right without straying from what is actually known about these animals. We also consulted with several paleontologists, Denver Fowler gave input on the Triceratops and John Hutchison gave input on the Tyrannosaurus, both were very helpful in creating what we hope are the most accurate set of animations for those animals possible. On top of that we have our own living Dinosaur on the team, an Emu named Gerry.
Nick - We’ve also assembled a sizeable library of scientific papers that deal specifically with things like the range of motion in different joints, the natural articulation of limb bones and vertebrae, and what footprints and trackways can tell us about how dinosaurs moved. Dr. Phil Senter in particular has produced quite a few papers that nicely summarizes forearm anatomy for major dinosaur groups and these have been very helpful as well. I think the biggest challenge is that, even though it’s extremely helpful, the scientific literature isn’t a blueprint. For many of these animals we might be the first people to attempt to create virtual representations of how they moved, and it takes a lot of tinkering to generate something plausible.
Q. One could easily be overwhelmed with the number of possible interactions between species, from mating to fighting. How do these interactions work in "Saurian", and will we have to execute complex courtship dances to pass on our genes?
Henry - These interactions will work differently from species to species, so it’s hard to generalize. Certainly, intra-species interactions will be more nuanced than inter-species interactions for the most part. The AI for each species takes influence from ecologically analogous extant animals for whatever the fossil record can’t allude to, so their mating rituals will vary in complexity just as they do in nature. I think it’s safe to hint that you can expect Anzu to throw down some funky dance moves, though.
Q. While many games make violence a central focus, "Saurian" is sure to take a more naturalistic approach. How will combat work in the game, and what sets it apart from combat in other games?
Henry - One thing that sets us apart is that combat is not necessarily central to gameplay. You will only be pressured to fight if you’re playing as a carnivore and can’t find anything to scavenge. If you play as an herbivore, it may be possible (though unlikely) for you to avoid combat altogether. Combat mechanics will vary from species to species, especially depending on what scientific literature has to say about them. For example, you’ll see Fowler’s Raptor Prey Restraint theory at play with the Acheroraptor, but you may see something totally unexpected from animals like Anzu for whom we have no analogous literature to draw from.
Nick - Another goal for Saurian’s combat system is to try and make it feel as natural and immersive as possible. It is our belief that the more direct control a player has over their actions, the more engaged they are in gameplay and to that end we’re avoiding things like quick-time events or scripted combat situations. Living predators try to incapacitate prey as quickly as possible to minimize the risk of injury, and many have developed either specialized anatomy or specific techniques for combat. As dinosaurs faced these same pressures in the Mesozoic, we want Saurian to challenge players in the same way; finding the best way to deal with attacking challenging prey or defending themselves against specialized attackers.
Q. There is tendency for dinosaurs in media to sound repetitive and unconvincing, but many people are surprised by the complexity and variety of vocalizations real animals are capable of. With this in mind, how are the sounds of these animals going to be created?
Dillon “Jheuloh” Gotham - We are very much aware of the weird and wonderful vocalizations animals produce! To some extent, the voice of the animals for Saurian depends on what species is being worked with. Not just in terms of what sounds “convincing” but also in terms of what “character” should be sold. T. rex and Triceratops, for example, should sound just as imposing and as charismatic as they look! For many of Hell Creek’s animals however, there is more freedom and less expectation, and through them we may showcase the many and varied voices in nature.
Q. Will the AI provide the primary gameplay experience, or will scripted events become important as well?
Tom - Scripted events will not be present in Saurian in any capacity. Our aim is for a completely emergent narrative, constructed from the AI, dynamic events such as weather, and the ways in which the player interacts with these aspects.
Henry - AI is going to be one of the driving forces behind emergent gameplay and narrative in Saurian (in addition to its survival mechanics). You will never experience any sort of scripted narrative sequence. This is, in part, due to the AI architecture we're experimenting with--a type of machine learning called reinforcement learning, inspired by behaviorist psychology. Some variant of this architecture has been used by a few games in the past (Creatures, Black & White), but never with this intention or scope. Basically, rather than programming elaborate state machines for each dinosaur, we simply give them a handful of behaviors, the ability to sense their internal status and environment, and let them loose in the world; over time, they learn their own survival strategies based on their goal to stay alive and healthy. In other words: we can never fully control our AI dinosaurs (only guide and influence them), but they can potentially be far more complex than we could have ever hand-coded them to be. They have already surprised me by figuring out strategies I hadn’t anticipated or even intended to be possible. It's really exciting, but also very daunting. Not only is this use of learning AI largely unprecedented in game design, but reinforcement learning is still an active field of academic AI research. As a result, there are many questions that I can't answer until I've done more research, development, and testing; even then, I may not be able to answer all of them. Which isn’t the most PR-savvy thing you can say as a team member for your pre-Kickstarter indie game, but it's important to be transparent and offer assurance that, if they don't work out, the AI will be redone more traditionally. There has been a rift growing between videogame and academic AI over the last 20 years or so, and I am enthusiastic about the idea of working to bridge the two. All this means that Saurian's AI could potentially be some of the most behaviorally complex in videogame history, and really help us depict dinosaurs as complex animals with their own internal and external motivations, rather than their traditional videogame role as bullet sponges. I think it's a risk worth taking.
Q. "Saurian" will utilize inclement weather, such as storms and floods, to provide an exceptionally dangerous environment. Will dangerous weather conditions occur randomly, or is there a predictable seasonal pattern to some extent? What sort of defenses does the player have against these hazards?
Tom - There certainly will be a predictable seasonal weather pattern present in the game. The literature states that Hell Creek had two seasons, a hot wet season and a dry cool season. The former is the period in which storms predominantly occur. Additionally, the transition period between the end of the dry season and the start of the wet season is extremely prone to fires due to lightning strikes on the still relatively dry environment. Saurian will reflect all of this data in its seasonal weather patterns. Players will have to learn their environment in order to figure out strategies for surviving these catastrophes.
Q. Is it too early to say what the minimum PC specs might be? If the title is successful, is there any chance of console ports in the future?
Tom - I’m afraid it is definitely too early to tell what the specs for the final game will be but optimization is a major focus of development. The way Microsoft, Sony, etc. handle indie development would mean that our title would have to be extremely successful for us to even consider console ports. The game is being developed with the potential to be played on Macs as well as PCs.