Skyrim Lead Designer admits Bethesda shifting to Unreal would lose 'tech debt', but that 'is not the point'
Skyrim Lead Designer admits Bethesda shifting to Unreal would lose 'tech debt', but that 'is not the point'
Skyrim lead designer Bruce Nesmith explains why Bethesda switching to Unreal may not be the upgrade fans think it would be.
The problem with Unreal Engine is (and always has been) that Epic makes the engine to make the game they're currently working on. So right now it is a Fortnite engine. Previously it was a Gears of War engine. (Maybe throw Paragon in between.) It started out as the engine for Unreal Tournament.
So if you want to take that engine and start making a different type of game, it's not necessarily going to have the tools you need. It's not necessarily even going to do what you need it to do at the base engine level. Not that it couldn't, but Epic doesn't give a shit. So they give you all the source code and support for building your own version of the engine so you can add the features you need.
You want to make a vast, persistent, open world with vast dungeons you can enter and explore? Yeah you're going to have to build support for that in the engine yourself. You want to do it without loading screens? Better get deep into that engine code. You want to have vehicles or mounts? NPCs, companions, AI enemies? When they hadn't added them to Fortnite yet, totally up to you to figure out, and probably through modifying the engine. Need to make major rendering improvements? Better dig in. Problems with the art pipeline lacking features you need
Every time you touch engine code, that's new tech debt. If a new version of the engine comes out, you have to integrate the changes. The longer the project goes on, the harder that becomes. Then Epic finally comes out with the feature you built yourself (say vehicles) but its only partly the way you did it. Now you're fucked and you have to decide right there: strip out your changes, switch to theirs and redo most of your work, or, stop taking engine upgrades and integrate new features piecemeal. Now you're in tech debt hell.
Almost every developer starts off with saying, "we'll use the engine as is, no engine changes allowed!" Three months later the cynical director is having a high level meeting about allowing a major feature get implemented in engine code. But it will be alright, they tell themselves. 3-5 years later they're up to their eyeballs in tech debt of engine changes, and realizing Amazing Game 2 either needs to be built using the old version of the engine they're stuck on from 2-4 years ago, or built from the ground up on a new version of the engine.
I'd be thinking long and hard before switching to UE5 if I were Bethesda. And they have the advantage of having access to some of the best Unreal Engine developers in the world (Obsidian, The Coalition) now that they're part of Microsoft. They're also probably getting a bunch of pressure to make the change as the studios create a corps of experts.
If I were them I would be very tempted to make the necessary changes to Creation Engine, and stay far, far away from Unreal. Sacrifice a year or two and your top engine devs to overhaul the pain points of Creation Engine, keep full control of your pipelines and versioning, and make the game you want to make, not the one Epic wants you to make. You can even make awesome DLC or a smaller sequel game on the old branched engine while the overhaul takes place, and just have a small core team working out the kinks on the new system.
I guess my point is, tech debt is not the point, because there will always be tech debt. It's a much bigger decision than that.
They don't make the engine to make that game. They make the game to prove out that they didn't miss something egregious in building the engine; or, "eat their own dog food". It has gained features over a long period of time that would fit common use cases from other developers, regardless of what Epic has built.
Meanwhile, nothing will convince me that Bethesda's tech stack is worth keeping.
The overall point still stands though. No off the shelf engine will have all the features a game needs unless the game is staying within the bounds of what the engine already covers.
At this point, switching engines means a hell of a lot of work only to eventually end up exactly where they are now again.
It's a legitmate question without an easy anwser, as to whether that work is better spent moving to a new engine or improving the existing one.
Unfortunately the path Bethesda is seeming to go with is to do neither. I can't imagine making a game like Starfield and not at least trying to find a way to make more of those loading moments "invisible" to the player rather than full on "yank you out" loading screens.
They don't make the engine to make that game.
They shouldn't, if they're going to be an engine company. But anything that isn't for keeping Fortnite pulling in billions of dollars is secondary.
It has gained features over a long period of time that would fit common use cases from other developers, regardless of what Epic has built.
Gained and lost. Very basic things necessary to make all the new features work with anything "not Fortnite" were missing when UE5 was released. It absolutely released as an engine for making Fortnite type games and everything else was/is an afterthought. You either had to make atrocious work arounds, engine changes, or wait for stuff to be fixed/added, delaying your project.
Meanwhile, nothing will convince me that Bethesda's tech stack is worth keeping.
Do you have inside knowledge? UE5 isn't the be-all end-all of game engines. Not everyone should switch to it. And frankly, as gamers and devs, we desperately need a good competitor to show up soon. Epic is gaining way too much control over our experiences.
You just reminded me of the plague that was UE3 for MMOs in the 10s, they couldn't have many players on screen, and so much texture streaming it's unreal (heh).
Thank you for being one of the few people in this thread with any sense.
Have you done game dev? Have you ever used Ue5?
Yes. AA/AAA dev with UE3, UE4, UE5 and several proprietary engines. I've even contributed to Unreal Engine code with bug fixes.
But arguing over whether or not you should use this engine or that engine, the engine is in service to the game. Is the game good? I donât care what the engine is. The gameâs good! Letâs play the game.
That's the problem, the games aren't good anymore!
Yeah they can focus more time on making the game good, if they don't have to focus their energy on a bad engine.
That engine terrifies me at this point. I'll admit I don't have any knowledge on the inner workings of a game engine, but it feels like it has to be held together with band-aids and gum
That understates the importance of a studio having respect for the customers. Yes being a good game is the most important thing but if ES 6 released tomorrow on the same engine as theyâve been using, there would a community uproar.
But why? As the lead designer in the article states, if the game is good who cares what engine they're using. The creation engine isn't holding Bethesda back. Just imagine if Starfield had released on Unreal instead of Creation engine? Would fewer loading screens and better facial animations have saved Starfield? I don't think so. The engine was not the issue with Starfield, the piss poor game design was the issue. Unreal engine isn't going to solve boring perks, boring quests and a bland world.
If TES6 comes out on creation engine 2 or 3 or whatever, and it's the next big thing like Skyrim, nobody is going to give a shit that it's the same engine. People might actually be angry if it's not on the Creation engine because that would mean modding is going to take a huge hit. Every current Bethesda game modder would have to learn how to mod Unreal engine and I can near guarantee it's going to be a lot harder than modding Creation engine.
How long are they planning to be hamstrung by the tech debt they've accrued? Sooner or later they're going to have to do something about it, surely?
Their games all look the same, in that it's always obvious that it's a Bethesda-engine game (whatever they're calling it this week). They're always janky, usually at least a console generation behind their contemporaries, and they always feel held together with duct tape and prayers.
Playing their games is an exercise in sighing and trying to ignore the jank. Everything always feels like it's wheezing along and trying to do anything beyond the obviously intended actions is likely to cause instability in the quest scripting.
I'm reminded of how Deus Ex players would try something only to find that the game was built to take that into account and allow for it. It's the opposite of how it feels playing Bethesda's games.
I wonder how much ship-of-theseus work can be done. I doubt they made things terribly modular but hopefully certain functions and groups of them are replaceable for some improvement. But, yeah, not ideal at all.
If CEOs only knew how much time and money they lose to tech debt, they would dedicate their careers to fighting it.
That sounds like a problem for next year's CEO.
Long-term tech debt is also really just part of the problem, the same thing occurs in shorter time intervals too when you e.g. push fixing a bug from the time before release to the time after or even just from the time when one developer is working on that particular feature to after the time when it is merged into the shared code base.
If I can't put a skull in a basket and balance it on a cheese wheel, then trick an NPC to walk on it, thus yeeting them into the stratosphere, is it even really a Bethesda game?
Watch Bethesda miraculously over performs the unfathomable hype around next elder scrolls game. Puts out what is unanimously considered the perfect performing and engineered video game that all peoples of the internet love. Give it one week after creation kit is released for it and someone will have made a mod that injects all the old physics bugs/features into the perfect game and will be the #1 downloaded mod for the rest of its existence đ¤Ŗ
And gain new tech debt in the process!
Tech debt accumulates over time. Starting fresh is the opposite of that
Starting fresh also leaves you with nothing
Chances are, you start fresh, start copying/ adapting existing code. Leaving you with much the same issues as you were trying to get away from
Not another one, unreal is so bad for mods, please just overhaul your own engine, take the time, another year for Fallout 5 is nothing.
Fallout 5? We already had two fallout games since the last TES game. Drag wants TES!
fallout 5 is most likely gonna be todd howard's last game ever. it wont come until after TES 6
It's cute that you think changing engine weight only add one year to a project
Do you think Total War will need 2 years to fix their engine then? They are also doing that after it finally caught up with them in Pharaoh.
It better be a major point, their current engine is preventing their games from meaningfully competing now. Their 20 year old engine, makes 20 year old games with a mediocre coat of paint.
Their engine is not hamstringing them. Plenty of good games have shipped with Gamebryo/Creation engine, without massive numbers of bugs.
The problem is that Bethesda doesn't give a shit about fixing anything - they ship bugs that have been in previous games, that users have outright identified and fixed for said previous games. They apply the exact same we-don't-give-a-shit attitude to their engine.
Also every engine is "20 years old", Source2 has some code from GoldSrc and Quake Engine, because if the code works perfectly then you don't just rewrite it for no reason. You rewrite parts of the engine - the parts that are holding you back in some way. And Bethesda has been modifying and extending their engine.
But, ignoring all of that, suppose the engine really was the issue: it takes 5ish years to write an engine from scratch. Starfield was in development for 8 years. Skyrim released 13 years ago. Skyrim also released 2 years ago, and a couple of times in between those periods too. Bethesda could have rewritten their entire engine from scratch if they wanted to, in that time.
The problem is that Bethesda just doesn't give a shit about quality, they chose their engine development choice by development choice. The problem is that Bethesda.
I know it would have the same issues as the Unreal Engine - all the training, engine building, and systems integration it'd take to get a game released, but I think it'd be cool if Bethesda were to make an Elder Scrolls game on their ID Tech engine. That codebase is pretty celebrated.
I would have absolutely zero interest in an Unreal Elder Scrolls or Fallout game.
There's already hundreds, maybe thousands of indie games trying to be that very thing available on steam today and they all suck.
This indie games are also made by amateurs without the ability to make something with the breadth and scope of an actual Elder Scrolls game. They're not bad because the tech sucks; they're bad because they're not well designed in the first place.
Check out The Outer Worlds for a competently made game of similar scope and mechanics that's also on Unreal and doesn't suck.
Is the game good? I donât care what the engine is. The gameâs good! Letâs play the game. Well, that's the issue Bruce, isn't it? The games are not as good as you think they are.
Nesmith explained that the studio canât focus on the needs of modders when creating a core game. Maybe you should wonder why so many of those mods are corrections, bug fixes and extensions to the engine though. Maybe that's a hint?
Dont they fucking own IDTech?
I mean they already own the Quake engine, or ID Tech 8 or whatever, just do it on that.
I'm gonna assume that this is because they are Microsoft now. They use a ton contractors on everything these days instead of actual employees but the catch is contractor can only stay there for 18 months and then have to leave or get hired so any experience on the project gets thrown away. Moving to unreal allows them to bring on a wider group of dev contractors so hopefully the ramp up isn't as long. This is just a theory but the same thing happened to Halo since that's moving to unreal to. If they just retained actual talent this wouldn't be a problem and the games wouldnt suck as much even if engine was less advanced.
I hope they don't go with unreal, the level of broken-ness of every game on UE5 really worries me
I don't know anything about UE5, but many game developers seem to be putting out unfinished shit right now. I have to wonder if it's the engine or just the studios.
Yepp it's about Biden doing his best. That is what this is about.
I assume you commented in the wrong post?
ooh wasn't there a post going around yesterday about a bug that actually puts comments on the wrong threads?
Nope I was just referencing the interview with Joe Biden where he said "the important thing is that I gave it my best. This is what this is about."
No dude, that isn't the point what so ever.
Where I read here that changing the engine "would be better, but that is not the point".
Yes dude that is the point.
For me the parallel was that people could improve the process but don't because of ego/comfort/not understanding the problem.
Wasn't a jab at Biden just an association on my part.