Devs share the most memorable things they've had to cut from games

Making games is often as much about refinement as creation.

For as many polished finished products make their way to market, there are thousands of ideas that never make the cut, or end up so iterated on and mutated by the time they’re finished that they’re unrecognizable from their original form.

So with an eye towards visiting the video game version of the Island of Misfit Toys, Gamasutra reached out to some game developers and asked them about some of the most memorable ideas they left on the cutting room floor, and why they had to leave them behind.

Mitch Gitelman, Harebrained Schemes

Of course, a huge number of features tumble into the vast chasm between ambition and practicality. As Mitch Gitelman explains, that gap was expanded when creators starting selling their projects on crowdfunding platforms, and had no real precedent for scaling expectations.

“During our Kickstarter campaign for Shadowrun Returns in 2012, Harebrained Schemes learned about scope control the hard way,” he says.

“Our original goals for the game were relatively modest but once we smashed through our initial funding goal in only 28 hours, all bets were off. As one of the earliest video game projects on Kickstarter, we had little research to fall back on. We were caught with our pants down and the world watching, improvising a series of exciting stretch goals that included the addition of a second city to adventure in if we raised one million dollars. The original tabletop RPG game was set in Seattle 2053 and we told backers that if we were lucky enough to reach such a high funding level, they would be able to vote for the additional city.”

Promising to nearly double the scope of their game was a massive reach, but luckily for Harebrained’s fans, a short-term loss became a long-term gain.

“Fast forward several ‘How the hell do we make this game?’ months later, and we realized that there was no way we would be able to deliver on our promise and we were forced to cut it. However, we promised that post-launch, we would take the additional city they voted for and use it as the setting for a full second campaign AND that our Backers would get that campaign for free," continues Gitelman. 

"While this was a painful cut, it led to Shadowrun Dragonfall - a far larger and richer experience than a second city in Shadowrun Returns. Dragonfall was a fantastic opportunity to build on the promise of Shadowrun Returns and take our narrative ability to the next level. While giving away forty thousand units of the game to backers may not have been the smartest business decision in the world, it proved that we were a studio with integrity and set us up for future Kickstarter successes culminating with Battletech, the most ambitious title in Harebrained Schemes’ history.”

Brenda Romero, Romero Games

Sometimes what gets left behind isn’t some novel mechanic or over-ambitious system. Sometimes it’s a wholly bespoke character, a creation that any fiction writer can attest can feel more alive and precious than any collection of art or dialogue would suggest, something Brenda Romero can sympathize after her time developing for Sir-Tech.

“After we finished Wizardry 8,” Romero told Gamasutra, “I really missed a few NPCs when the game shipped. I spent so much time writing, refining, and testing them, living in their heads, that to have no reason to do that anymore was a bit like a kid moving out of the house. I particularly missed Rodan Lewarx, Saxx and Balbrak. That I remember these characters nearly 20 years later should tell you something.”

Ryan Hewer, Little Red Dog Games

Another recurring theme amongst these lost ideas are features that had to be trimmed because of a cold, pragmatic lack of time or resources, an experience Ryan Hewer bumped into working on Deep Sixed.

“We had originally intended for Deep Sixed to have greater user customization – particularly in the form of the AI avatar that accompanies you throughout the game. Our hope was that at a particular point in the story, the player would select from a series of predetermined personalities, each with different avatars and voice actors behind them – and that choice went beyond the superficial experience to actually having gameplay implications. Each AI would handle the procedurally-generated situations differently – and over time, the player experience would be vastly different depending on the avatar chosen. In the end, while this was a desirable mechanic, we simply didn't have the resources to re-record hundreds of lines with additional actors.”

That customization element wasn’t the only thing the team ended up ditching during development to streamline their process.

“We also had hoped to open the game up with some modding tools so that players could design their own missions, import creature models and develop their own story-arcs; again, this would have added significant time to development – and we had already pushed back our launch already. Ultimately, we're proud of what we released. A developer needs to know when to let an idea go – no matter how good it is – if it's an anchor weighing down production.”

Andrew Trese, Trese Brothers Games

Some of the most fun emergent gameplay surfaces when you start blending procedural systems and get to the witness the unpredictable results. Unfortunately, sometimes some of those unforeseen interactions end up terminally breaking your game, as Andrew Trese discovered.

“In our space opera RPG, Star Traders: Frontiers, we have built a massive, branching narrative from story characters spread across the an ever-changing galaxy,” Trese says.

“Underpinning that, we have a character simulator that drives relationships between all the characters and conflicts – even allowing them to assassinate each other. In the first meeting between these two layers (story and simulation), the non-story characters started killing the story characters. You'd be working your way down a long story arc and return to your homeworld – only to discover the news that Zette Faen, your critical story character, had been assassinated by a Prince from halfway across the galaxy. It was nearly malicious – as if the simulation really wanted to be the top dog and was killing the story characters on purpose!  In the end, to keep the story arcs workable, we had to remove the ability for the simulated characters to murder the story characters – at least while the story arcs are active!”

Keith O'Conor, Romero Games

Sometimes, painful cuts are necessitated when one member of a team finds themselves working in a vacuum, out of touch with the people designing the other elements of a game. If a system doesn’t integrate well with the whole, it runs the risk of being eliminated entirely, a painful lesson Keith O’Conor learned during his time at Radical Entertainment working on Prototype.

“The designers wanted something that gave a 'hazy street after a battle' look to support the explosions and chaos that appeared throughout the game,” O’Conor remembers. “I spent some time working on an effect that was good-looking, relatively cheap, and integrated well with the particle effects system.”

While polishing the game and looking for ways to optimize and improve performance, O’Conor made a disheartening discovery.

“There was some overhead associated with the fog effect that cost GPU time even when there was no fog on screen, so I went back to examine what could be done about it. To my dismay, I realized that the system had hardly been used anywhere in the game. The tools were lacking and the effect in general was difficult to use, so it had largely been ignored. It was too close to shipping to do anything about it, so with a heavy heart I cut the feature completely in order to gain back a fraction of a millisecond of GPU time. I learned a valuable lesson that day about the necessity of providing good tools support, and the importance of getting artist feedback on the design of any visual effect right from the beginning!”

Carlos Carrasco, Weird & Wry

Of course, there are also times when what sounds on paper like a lot of (explosive) fun turns out, when implemented, to be a terrible idea.

“We had a huge meteor rain of fire of destruction that landed on the player’s space station with little warning – usually in an expensive and busy area,” Carlos Carrasco says of their game Spatials: Galactology. “For a full minute, the fury of the cosmos unleashed a stream of flaming, exploding rocks on the poor station. This was part of a random events-focused beta release. Who doesn't like being bombed from orbit by meteors – determined just by random chance? Our players, it turns out.”

Kyle Creamer, Trykon Studios

And then there are ideas that really are solid and functional, that seem to fit a game’s vision for a large part of its development, but then at some late stage end up getting trimmed because they don’t quite fit with the end product. Kyle Creamer recounts an instance of this in Omnicube, around the the concept of ice walls that would shatter on impact.

“This was a great way to create a temporary "one-use" wall,” Creamer points out. “You could use it to position a block once, but only once. It created a lot of complications – because now, in addition to the static features of each level, the player would need to figure out how to utilize the ice walls to temporarily position blocks.”

In the end, however, the ice walls were removed “for a couple of reasons: scope and aesthetic. We initially designed this feature when we were considering having a separate art aesthetic for each set of levels; each group of 15 levels would have a theme, such as ‘fire’ or ‘ice.’ Eventually, though, we decided to go for a more unified aesthetic for our game; this helped us cut down on art costs and development time – so our ice blocks just didn't really fit any more. I was sad to have to remove the ice walls – but ultimately, this choice helped us get our game finished and ready to ship, and that was the most important thing.”