Game Engines: Part 2 - When in-house engines go wrong
A series of unfortunate events
For part two of the series I’m going to be digging into some of the known horror stories involving in-house engines.
The reason I’m doing this, is to paint a picture as to how difficult of a balancing act engines can be, especially at the AAA level.
The hindsight we can garner from these horror stories tells us heaps about what actually went wrong. Rather than blind criticism from players about ‘bad engines’ or 'lazy developers’ we can actually understand the complexity involved around the development of a game and where our criticisms should really lie.
Frostbite - EA
Nothing exemplifies an engine horror story like EA’s Frostbite.
Explained in detail via probably one of the greatest pieces of investigative journalism, How BioWare's Anthem Went Wrong, Frostbite has been at the core of multiple bad game launches for EA.
EA made Frostbite to be a first person shooter engine, primarily for their Battlefield series of games. However, EA in their infinite greed and wisdom, then forced all of their studios to use the engine for their games, this included RPG’s such as Dragon Age and Mass Effect, sports titles like FIFA and NHL, and racing games such as Need for Speed.
Because the engine wasn’t set out to support other genres of games, this caused a lot of problems as tools and systems didn’t exist to support features that weren’t in a FPS. Save game systems and levelling up systems for example.
Now, when other multi-billion dollar companies, make their own engines, they resource internal teams whose job it is to expand, maintain, and document said engines. Meaning that should features be missing, or a team needs help there are dedicated people who can do that.
However, EA did not provide ample enough support for the engine and so studios were left to figure it out for themselves, and where features were missing, cobbling them together on top of an engine that didn’t want to work that way.
“Frostbite is like an in-house engine with all the problems that entails—it’s poorly documented, hacked together, and so on—with all the problems of an externally sourced engine,” said one former BioWare employee. “Nobody you actually work with designed it, so you don’t know why this thing works the way it does, why this is named the way it is.” (source)
Fifty-four games under EA were built using Frostbite, only ten were Battlefield, the game the engine was built for. This led to increased development times, heaps more testing and bug fixing, and multiple bad launches of titles such as Anthem, Mass Effect Andromeda, and Battlefield 2042.
Given that its quite common for AAA studios to opt to create their own engines given the scale they operate at, you can’t fault EA for choosing to develop an engine in-house, and forcing their studios to use it.
You can, however, fault EA for how they built and supported the engine and the studios using it. The lesson here is, if you are going to try and streamline development and reduce costs by making your own engine, don’t hamper development and cause overspend by not supporting it properly.
Crystal Tools - Square Enix
Like many big studios, Square wanted parity across the plethora of games they were making. With so many Final Fantasy titles in development and each team making their own tools Square decided to build their own engine, originally called the White Engine, but later renamed to Crystal Tools, because apparently they wanted it to reflect the flexibility of crystals (are crystals flexible?)
From 2007 to 2013 Square sunk all their resources in the Crystal Tools and released all three Thirteen titles, Fourteen (the MMO), and Dragon Quest X on the platform.
Whilst, EA tripped up because they under supported the teams using the engine, Square did the opposite.
The engine was forced to grow to meet the needs of the studios’ various projects and consequently failed to keep up, impacting the development of various games.
In the case of Final Fantasy XIII, the artists were making assets before the engine had been modified to take them, when the engine finally caught up, those assets were incompatible and had to be remade.
As Final Fantasy XIII suffered more and more delays it was given priority and all other projects reliant on the engine had to be paused. The game's producer later remarked that upgrading the engine at the same time as a major release may have been a mistake.
Despite this, Square Enix went on to create another proprietary game engine known as Luminous Engine.
However, learning from their mistakes, they did not force all teams to use it and only a handful of titles were published on the Luminous Engine, notable Fifteen. The wildly popular Final Fantasy 7 Remake titles, for example, use Unreal Engine.
Tiger Engine - Destiny
Many things hampered the development of Destiny and their sequel but unfortunately the prime suspect was often their Engine, the Tiger Engine.
Destiny was pitched as a ten year franchise with live service elements. If I was asked to make an engine to support ten years of live service gaming, my primary goal would be to create something that was flexible and allowed for quick iteration and prototyping.
Despite it’s name, the Tiger engine was anything but fast or flexible. One of the biggest complaints about development of the first Destiny title was that the engine was so resource intensive, that even small changes made to parts of the game, required an overnight process to render, some developers described this as grueling.
Given how thematically disjointed and shallow the game was on launch has led many to speculate that the slower than expected development process led to major story and gameplay elements being cut for the launch of the game.
The sequel wasn’t any strange to development issues. Destiny 2 did allow the studio to address issues with the engine and speed up the cumbersome development process. However, the development tools continued to be at odds with the intent of the game. It took time for the studio to get up to speed with releasing regular content and unfortunately hit a point in the game where it had too much content and the testing loop was getting too laborious on the studio and they were forced to remove content in order to keep adding more. This is at odds with other similar MMO style games like WoW and FF14 where they have continuously added content without removing any for well over a decade. The difference there is that those games don’t strive for the visual fidelity that Destiny does, and it may actually be that it’s currently impossible to develop an engine that is going to have a high visual quality and be continuously adding content.
More examples
Despite there being many successful in house engines such as Decimal, Anvil, RAGE, LORE etc there are many big studios at the moment pivoting away from their own in-house tooling.
After the repeated issues with Frostbite, newer titles such as Apex Legends and Star Wars Jedi: Survivor went in their own direction with Jedi adopting Unreal and Apex using Source.
Despite several successful AAA games, CD Projekt Red has announced that their future Witcher and Cyberpunk games will no longer be using their in-house RED engine and will be instead switching to Unreal. Developers commented that RED was unable to meet the visual fidelity of their aspirations and was found to be quite buggy.
Bethesda confirmed that in order to develop Starfield and the next Elder Scrolls titles significant upgrades needed to be made against their in-house Creation Engine. The studio has been infamous for their engines buggy nature and unique visual elements, mainly involving characters, and after Fallout 4 and 76 it appears that the patience of gamers to tolerate such issues has run dry, and so an upgrade was sorely needed. However, the engine’s upgrades took a lot longer than expected.
Halo Infinites lacklustre launch, and failure to revitalise the series had been put down to poor management and their engine (feels like Frostbite all over again). Despite the aging Halo engine which dates back to as far as the early 2000’s getting significant upgrades for the development of Infinite it was described by many developers and notoriously difficult to use. The development relied heavily on using contract employees who most of them had no previous experience with engine causing considerable overheads. At this point in time it is believed that the studio 343 Industries is abandoning the engine in favour of Unreal.
I really could go on for hundreds if not thousands of hours more about the fun complexities of in-house development engines. My goal is not to paint a picture that all in-house engines are doomed, especially when their are so many working examples to pull from, but to highlight how extremely complex these things can be and the ripple effects they can have on the quality of game at launch. In my next piece i’m going to be diving into the world of commercial engines. With so many studios switching to Unreal Engine, there is a lot to unpack there.
If you have enjoyed this piece or any of my other pieces of writing feel free to share I your channels!