It can be more of a mixed bag than that, though. If your employee retention and training is good enough that you have plenty of people who wrote the engine or at least understand it really well (which doesn’t seem to be the case at CDPR since Cyberpunk’s crunch), it can be much faster to alter it than to figure out the equivalent guts of Unreal Engine. That won’t end up making a difference if you stick to well-trodden paths that lots of games from lots of studios use, but if you want to do something that Unreal doesn’t support out of the box, it can be quite hard to wrangle.
That’s only if your employee retention and training are good enough, and you don’t plan on growing your team. When adding more developers you either have to invest incredible amounts of time and money to get everyone to that level (and we’re talking years per developer), or you’ll be left without the ability to really alter it while still having to educate every new developer on your engine.
This approach simply doesn’t scale or work out in the long run.
If you’re specifically working on a game that stock UE5 can’t do, e.g. you need to make the kind of far-reaching changes that Valve had to make to Source to make Portal possible, you end up with most of those problems even if you’re doing it by modifying UE5 rather than modifying your in-house engine. You’re still ending up with a custom engine at the end of the process and still need to make tooling for it and onboard everyone, even if it ends up fairly similar to stock UE5 due to being modified UE5. It doesn’t necessarily work out any more scalable or sustainable than modifying an in-house engine once you’re making this kind of change. The outcome ends up being that games that can’t be made in close-to-stock UE5 just don’t end up getting made.
I didn’t say that they did, just that switching to UE5 can be a mixed bag rather than always unambiguously better. My original comment was pretty explicit about it not being applicable to CDPR.
It can be more of a mixed bag than that, though. If your employee retention and training is good enough that you have plenty of people who wrote the engine or at least understand it really well (which doesn’t seem to be the case at CDPR since Cyberpunk’s crunch), it can be much faster to alter it than to figure out the equivalent guts of Unreal Engine. That won’t end up making a difference if you stick to well-trodden paths that lots of games from lots of studios use, but if you want to do something that Unreal doesn’t support out of the box, it can be quite hard to wrangle.
That’s only if your employee retention and training are good enough, and you don’t plan on growing your team. When adding more developers you either have to invest incredible amounts of time and money to get everyone to that level (and we’re talking years per developer), or you’ll be left without the ability to really alter it while still having to educate every new developer on your engine.
This approach simply doesn’t scale or work out in the long run.
If you’re specifically working on a game that stock UE5 can’t do, e.g. you need to make the kind of far-reaching changes that Valve had to make to Source to make Portal possible, you end up with most of those problems even if you’re doing it by modifying UE5 rather than modifying your in-house engine. You’re still ending up with a custom engine at the end of the process and still need to make tooling for it and onboard everyone, even if it ends up fairly similar to stock UE5 due to being modified UE5. It doesn’t necessarily work out any more scalable or sustainable than modifying an in-house engine once you’re making this kind of change. The outcome ends up being that games that can’t be made in close-to-stock UE5 just don’t end up getting made.
If you continuously make games like that, sure! But CDPR doesn’t do that.
I didn’t say that they did, just that switching to UE5 can be a mixed bag rather than always unambiguously better. My original comment was pretty explicit about it not being applicable to CDPR.