Starfield
Design Goals
Overview
After receiving considerable praise and positive feedback for my work on weapon tuning in Nuclear Winter, I was entrusted with overseeing weapon design for Starfield. My primary objective was to enhance the handling and overall feel of weapons. In previous Bethesda titles, I felt like the moment-to-moment combat experience needed improvement, and I was determined to address this issue.
I distinctly remember a moment at Quakecon while manning the booth for Nuclear Winter. Someone came up to me after playing and said "why does combat always suck in Bethesda games. Nuclear Winter is better but still not great." All I could do was shrug and reassure him we were actively working on it. A few years later I would get my chance on Bethesda's next biggest game, Starfield.
Over the production of Starfield, I had control over weapon damage, recoil, accuracy, mag size, reload speed, ADS speed, weight, weapon bash damage, and the creation of all the mods and their stats for the impressive array of fifty weapons we shipped with. I also created the weapons used as quest rewards, as well as the named weapons you could purchase from vendors.
I distinctly remember a moment at Quakecon while manning the booth for Nuclear Winter. Someone came up to me after playing and said "why does combat always suck in Bethesda games. Nuclear Winter is better but still not great." All I could do was shrug and reassure him we were actively working on it. A few years later I would get my chance on Bethesda's next biggest game, Starfield.
Over the production of Starfield, I had control over weapon damage, recoil, accuracy, mag size, reload speed, ADS speed, weight, weapon bash damage, and the creation of all the mods and their stats for the impressive array of fifty weapons we shipped with. I also created the weapons used as quest rewards, as well as the named weapons you could purchase from vendors.
Main goals
Make gunplay feel more realistic and dynamic
I wanted our weapons to feel less "gamey" and err on the side of realism when possible. This meant upping the falloff range of bullets, increasing recoil, and making combat more deadly. I was not a fan of bullet-sponge enemies and made sure that enemies died in three to seven hits at first. This would last for most of the game production until we upped both player and enemy health near the end so that skills and powers could play more of an impact.
Make gunplay flow better and reduce clunkiness
A key concern when I started work was that players couldn't interrupt a lot of actions. This made the game feel clunky and non-responsive to me. So with that in mind, my guiding principle became "prioritize player intent." This meant that if the player wanted to do something - such as start sprinting while reloading - it would interrupt the reload to sprint instead of finishing the reload.
Give Each Manufacturer's weapons their own personality
With fifty base weapons in Starfield, it would be easy for some to fall by the wayside. To combat that, I wanted to make sure they each had their own unique flavor with their personality tied to their manufacturer.
- Old Earth weapons were less advanced and had the fewest mods, but they were extremely fun to shoot.
- Allied Armaments were heavier, but more powerful.
- Va'Ruun were mystical-looking and extremely powerful, but slow firing.
- Arboron were slick and precise. They also had the single EM-only weapon.
- Kore Kinetics were expensive monster houses of damage, but they ate through ammo like crazy.
- CombaTech had faster fire rates, but did less damage. They also had access to the most mods.
- Laredo were the slowest firing, but the most accurate. They also were less advanced, so didn't have recon sights.
Make mods feel more important
I wanted players to see a mod on a weapon and to think, "I really want that!" To achieve this, I needed to make mods worthwhile and unexpected. This meant introducing new types of mods to the game (such as ammo variants). Additionally, I implemented a novel system where weapons were generated with random mods instead of in previous titles where weapons came with static mod sets. This helped foster a sense of discovery for players as they naturally found new mods during regular gameplay.
Challenges
No long saves until a few months before release
One lasting issue we had throughout production was getting saves working properly. Saves would regularly break and become unusable upon updating to the next build. Because of this, players had to start out with preset builds and saves each playtest, meaning fewer weapons and mods were tested. This also led to relatively little organic feedback at higher levels.
Most feedback was focused on other aspects of the game
There were so many different and important systems in the game that weapons were rarely the focus of a playtest, even though players used weapons nearly every playtest. This meant not enough feedback on the weapons themselves.
The only feedback I would get was when I made sweeping changes, like adjusting accuracy and recoil to make the game feel more realistic - which playtesters both hated and loved. When I got weapons into a state I was happy with and rarely saw anymore feedback on them, I attributed it to the "no feedback is good feedback" philosophy. Looking back, I should have put more emphasis on getting weapon specific playtests set up instead of gathering feedback from general playtests.
The only feedback I would get was when I made sweeping changes, like adjusting accuracy and recoil to make the game feel more realistic - which playtesters both hated and loved. When I got weapons into a state I was happy with and rarely saw anymore feedback on them, I attributed it to the "no feedback is good feedback" philosophy. Looking back, I should have put more emphasis on getting weapon specific playtests set up instead of gathering feedback from general playtests.
A relatively small team
Throughout the project, the weapons group was only a producer, a handful of artists, and a programmer on it part-time. It would be a constant struggle throughout development to get extra people on the team for more than a few days of dedicated time.
It took a year took get another part-time programmer. Until then, I was heavily restricted in creating anything outside of what the current tools could do. This meant I had to come up with work-around solutions to the plethora of the problems that appeared.
It took a year took get another part-time programmer. Until then, I was heavily restricted in creating anything outside of what the current tools could do. This meant I had to come up with work-around solutions to the plethora of the problems that appeared.
Learning how to lead on the fly
Transitioning to the weapons group became a pivotal moment in my career - I was the only designer on weapons. Before this game I had not been a lead. I had been the only level designer or the primary system designer on a project before, but I was never in a position where I was setting up and leading meetings, helping mentor people, doling out tasks, reviewing work, and giving final say on major design decisions. The initial atmosphere within the team was more of a loosely connected group tackling JIRA tasks, with no clear leadership apart from the exceptional Dane Olds, who later became our Weapon Art Lead.
I jumped into the work headfirst and began writing up big proposals about how I wanted to improve gunplay. Presenting my work to the team, I had hoped they'd rejoice in awe at my game design prowess, but I received only blank stares and was quickly met with apathy. This led to some frustration on my part. I wondered what I was even doing there. The sentiment seemed to be that change was difficult, if not impossible, as we were often advised to "make it like Fallout 4," presumably due to time constraints.
After realizing this, I understood that I was approaching this the wrong way. I attempted to increase morale by getting people hyped about what we could still do. I started daily meetings with the weapons team to look over every single weapon we had and led discussions on what we wanted it to be like, not just what I wanted. Despite taking over a month to get through our vast array of weaponry, the team loved it. I strongly believe that these regular meetings are what led the team to creating such great weapons.
I jumped into the work headfirst and began writing up big proposals about how I wanted to improve gunplay. Presenting my work to the team, I had hoped they'd rejoice in awe at my game design prowess, but I received only blank stares and was quickly met with apathy. This led to some frustration on my part. I wondered what I was even doing there. The sentiment seemed to be that change was difficult, if not impossible, as we were often advised to "make it like Fallout 4," presumably due to time constraints.
After realizing this, I understood that I was approaching this the wrong way. I attempted to increase morale by getting people hyped about what we could still do. I started daily meetings with the weapons team to look over every single weapon we had and led discussions on what we wanted it to be like, not just what I wanted. Despite taking over a month to get through our vast array of weaponry, the team loved it. I strongly believe that these regular meetings are what led the team to creating such great weapons.
Lessons Learned
Making mods more impactful meant the base weapon needed room to grow
To make weapon mods feel like they were more impactful (also making them more attractive for crafting), the weapon base stats needed room to grow. My goal was to make mods give tangible bonuses so that players would notice the effect that mod had on their weapon. I originally made guns very accurate, which people loved and hated. But that meant that mods, such as the compensators and ergonomic grips, didn't make any kind of noticeable change unless you were firing at more than 50 meters. I ended up lowering accuracy across the board, which people mostly hated but quickly learned to be okay with once they got to see mods appearing on weapons.
Don't make every weapon drop with a mod
I set up modded weapons to drop random mods in different groupings based on the level of the AI or area. This meant low level weapons only had a small chance to drop with a mod or two, but higher level ones ended up dropping often with five or more mods on them. Todd Howard asked to bring the chance of getting a modded weapon from a drop down to 10% from the initial 90% I set it at.
At first I disagreed, but over time I realized that was the right call. If you always got a modded weapon from drops, then there'd be little point to go craft more mods. This meant that modded weapon drops also felt more special. In the end, I believe we didn't go far enough and should have dropped the chance down to 5%.
At first I disagreed, but over time I realized that was the right call. If you always got a modded weapon from drops, then there'd be little point to go craft more mods. This meant that modded weapon drops also felt more special. In the end, I believe we didn't go far enough and should have dropped the chance down to 5%.
Make decisions a more collaborative process
When I joined the weapons team, I rushed into massive overhauls and proposals too quickly, dropping them out of nowhere on the team in our weekly meetings. I needed to approach these topics more gently and not expect people to just jump onboard with whatever I wanted to do because I was the only designer. It was much easier to get people interested when they were also involved in the process, so I set out to make the weapons team highly collaborative. Everyone working on weapons was encouraged to voice their opinions on design, which not only solved many issues we had but made the work more fun.
This also occurred with some big weapon changes I wanted to do. I didn't want to wait for permission from above for everything I was doing, especially when those decisions could get bottlenecked for weeks. Not wanting to wait around, I realized that if I did the work and recorded video of it, displaying what I wanted to accomplish to the team, then that'd get more traction and give those above something more tangible to look at. Sending these out to the team meant we could design more collaboratively. Others would voice their opinions, and we'd iterate and make it better - thus increasing the chance it'd get approved. Sometimes this didn't work and I'd have to revert my work, but more often than not, it helped get things moving by creating comparisons to look at.
This also occurred with some big weapon changes I wanted to do. I didn't want to wait for permission from above for everything I was doing, especially when those decisions could get bottlenecked for weeks. Not wanting to wait around, I realized that if I did the work and recorded video of it, displaying what I wanted to accomplish to the team, then that'd get more traction and give those above something more tangible to look at. Sending these out to the team meant we could design more collaboratively. Others would voice their opinions, and we'd iterate and make it better - thus increasing the chance it'd get approved. Sometimes this didn't work and I'd have to revert my work, but more often than not, it helped get things moving by creating comparisons to look at.
Reception
Overall, the reception has been phenomenal. Players and critics alike have said that this has been the best combat in a Bethesda game ever! They've loved the weapons and the gunplay and I'm incredibly proud of the work that the team did on these.