There is a silver lining to the cloud of Stormgate running like garbage. This should steer other RTS developers away from UE5.
Apparently, Dawn of War 4 is running in Unity. C# is its primary scripting language.. its needed for engine integration etc. I've got more than 10,000 hours in C#. C# is very good and has been very good since its birth ~ 2001. Bill Gates might be one of the Lizard Men secretly running the world.. but damn he and M$ did an awesome job leading the creation of C#. I'm a big fan of both C# and Unity.
Unity allows you to manage 87 bazillion objects at near C++ speeds while coding in the safe and cozy C# managed code environment.
"Code in C# because life is too short to code everything in C++"
All this sounds wonderful, however, recent builds of DoW4 are running at 5FPS during heavy battles. Hopefully, King Art Games can improve things.
Sanctuary: Shattered Sun (Supreme Commander FAF like RTS with some binary terraforming) use Unity's DOTS (Data-Oriented Technology Stack) for multithreading. Programming for Unity's DOTS is quite complicated and difficult judging by what I heard, but the game was mostly bottlenecked by its graphical performance according to its developers. They will release a public demo of the game by the end of 2025, perhaps it will have some improvement in this area.
it is hard if you've never done low level programming. if you have a solid all around math/computer science background with a 4 year degree that includes stuff like compiler construction in C++ and encryption algorithms in C++ then DOTS is just another technology stack.
If you took an Avilo style "computer science" degree you'll probably find DOTS tough. If you took a CombatEx type of degree then you'll be fine. Given the level of Avilo's math education i hesitate to label his degree as a "computer science" major.
Put it this way.. once you master DOTS its like you are coding in a high-level, super-easy language like Visual Basic while getting the lightning speed execution performance of well made C++.
They’re very different, albeit overlapping disciplines, game dev is not my particular domain to be fair.
Engine optimisers aren’t going to build you a good game, they’ll facilitate a good game with good design being a good experience for end users.
I’d imagine it can’t be that much more difficult to recruit some engineers particularly au fait with C# or C++, Unity or Unreal, they don’t have to build from scratch with a whole load of systems, but may need to write bespoke optimisations for specific use cases
Stormgate’s number 1 issue is a game design one, not a game engineering one. And they made some problems for themselves in the latter category as well.
There was no need to engineer the game at such a high tick rate other than to claim you have the ‘most responsive RTS ever’, which caused a lot of performance headaches.
They’d have had the exact same problems if they’d decided to go with Unity, or any other engine.
Well no, but you’re talking about Unity being good because C# has certain advantages at ‘near C++ speeds’.
Surely if you hire some crack C++ specialists, UE can be made to be performant?
Personally I don’t think it’s super advantageous to go the UE route in RTS, for a variety of reasons. A lot of the graphical funkiness, I mean it’s just not that relevant to a genre where you’re zoomed out from the action, and high graphical fidelity makes visibility tricky
Interesting thread. Ive worked in game development with published titles for clients like EA and others.
As an RTS fan I have been thinking about this problem recently and have come across some nice opinion pieces on the subject. So, here is some food for thought:
My own opinion is that a large majority of "timeless" RTS titles (Total Annihilation, C&C, StarCraft, Warcraft III) were successful because the game engine technology brought something new and exciting to the table, and there was a sense of magic for the players (myself included). The problem I have with most UE5/Unity RTS games that I have played is that they all feel "muddy" to me, and despite differences in art style the rendering makes every game look pretty much the same.
Thanks for sharing, Myst. I agree. As someone pointed out under one of those articles, it's prolly good to rein in the technical ambition. I wonder too that if you can't build your own unique tools or engine, can you focus on the fun of the game enough, with existing tools, that it doesn't matter? I wonder if ego gets in the way of people just trying to make a thing that's really fun to play. When I think about something I would want to make, ego absolutely does get in the way. And I'm way overly ambitious. I'm not even an engineer. I don't even know one. I feel like it comes from a place of foolishness, and just not knowing what a thing takes, and focusing on the emotion of it. I personally would find it hard to slog through 8-10 years of technical problems for 99.9% of projects. I do absolutely think that you must know what you want to do, and that you must do it, and decide that failure is not an option. I think once you make that decision, nothing is too ambitious. But you must know.
On September 02 2025 04:06 JimmyJRaynor wrote: Are there any well running RTS games made with UE5?
On September 02 2025 09:31 WombaT wrote: Well no, but you’re talking about Unity being good because C# has certain advantages at ‘near C++ speeds’.
Surely if you hire some crack C++ specialists, UE can be made to be performant?
Personally I don’t think it’s super advantageous to go the UE route in RTS, for a variety of reasons. A lot of the graphical funkiness, I mean it’s just not that relevant to a genre where you’re zoomed out from the action, and high graphical fidelity makes visibility tricky
I think Frost Giant's first move was the biggest mistake. I think trying to make a large scale RTS with 100s of fighting units in UE5 was a big mistake. I think they overestimated themselves. They talked in such giant grandiose terms they talked themselves into thinking they were some kind of giant.... in reality they had yet to release a single piece of software.
Tim Morten keeps making Bobby Kotick look smarter and smarter with each passing month.
Well no, but you’re talking about Unity being good because C# has certain advantages at ‘near C++ speeds’.
Unity has the potential to work well in a large scale RTS scenario. UE5 does not have that potential.
Meanwhile, BAR sets the unit cap per player at a default of 2000 (which can be adjusted with custom settings), and retains performance.
My favorite thing about BAR is the "Flanking Damage" mechanic. If 2 units are attacking a unit from opposite sides, that unit takes twice as much damage from each attacker. In practice this radically increases the possibile outcomes of each engagement.
On September 04 2025 07:43 ThunderJunk wrote: Meanwhile, BAR sets the unit cap per player at a default of 2000 (which can be adjusted with custom settings), and retains performance.
My favorite thing about BAR is the "Flanking Damage" mechanic. If 2 units are attacking a unit from opposite sides, that unit takes twice as much damage from each attacker. In practice this radically increases the possibile outcomes of each engagement.
Yeah, BAR is crazy. 100 player game, every projectile is a physical object taking into account elevation, distance, flanking etc. and it runs well.
In general, you can't trust AAA studios when they say something is "too hard"
Solo dev built a better game in 2 years than AAA studios can do in 5 years with hundreds of devs and unlimited budget.
On September 04 2025 07:43 ThunderJunk wrote: Meanwhile, BAR sets the unit cap per player at a default of 2000 (which can be adjusted with custom settings), and retains performance.
My favorite thing about BAR is the "Flanking Damage" mechanic. If 2 units are attacking a unit from opposite sides, that unit takes twice as much damage from each attacker. In practice this radically increases the possibile outcomes of each engagement.
Yeah, BAR is crazy. 100 player game, every projectile is a physical object taking into account elevation, distance, flanking etc. and it runs well.
In general, you can't trust AAA studios when they say something is "too hard"
On September 04 2025 07:43 ThunderJunk wrote: Meanwhile, BAR sets the unit cap per player at a default of 2000 (which can be adjusted with custom settings), and retains performance.
My favorite thing about BAR is the "Flanking Damage" mechanic. If 2 units are attacking a unit from opposite sides, that unit takes twice as much damage from each attacker. In practice this radically increases the possibile outcomes of each engagement.
Does it work well in practice? Sounds a cool mechanic but I can see it being a bit wonky at the same time depending on the thresholds.
I could see it being a bit chaotic positionally, if you’re slightly off you’re not getting your flank damage, versus being slightly differently positioned by a few ‘pixels’ and it kicks in.
Do you feel they’ve addressed that kind of concern?
Does anyone remember World in Conflict? That game felt so innovative in the way it handled multiplayer (8v8 battles where everyone picks a "class" that gives them a certain type of units to handle). Basically turned Battlefield into a team RTS (including completely destructible terrain/buildings). I'm not saying it was perfect at all, but definitely I think an interesting way of handling RTS multiplayer that might thrive in today's market (team-based games with defined roles for each player). I also loved the way that they set WASD to move the camera around lol.
I feel like 1v1 is just too off-putting to most gamers nowadays, but a team-focused RTS game could possibly find success.
Of course I think indie RTS game's biggest problem is just that most of them are way too derivative and nostalgic (big problem with a lot of the indie game market). Show me something new, not just worst versions of StarCraft or CnC.
On September 04 2025 00:42 JimmyJRaynor wrote: Unity has the potential to work well in a large scale RTS scenario. UE5 does not have that potential.
Tempest Rising uses UE5. It's an RTS with 200 pop cap just like SC2.
you can't play 2v2 with a pop cap of 200 on a machine that costs less than $3000. and when you have a $3000 machine player 2v2 you can heat your home in january with your computer. they have a 2v2 with 100 pop cap mode that can be played on a reasonable machine.
there is no 3v3 or 3v3AI.
UE5 can not run decent scale RTS beyond 1v1 games.
On September 04 2025 07:43 ThunderJunk wrote: Meanwhile, BAR sets the unit cap per player at a default of 2000 (which can be adjusted with custom settings), and retains performance.
My favorite thing about BAR is the "Flanking Damage" mechanic. If 2 units are attacking a unit from opposite sides, that unit takes twice as much damage from each attacker. In practice this radically increases the possibile outcomes of each engagement.
Does it work well in practice? Sounds a cool mechanic but I can see it being a bit wonky at the same time depending on the thresholds.
I could see it being a bit chaotic positionally, if you’re slightly off you’re not getting your flank damage, versus being slightly differently positioned by a few ‘pixels’ and it kicks in.
Do you feel they’ve addressed that kind of concern?
It's continuous. There's no cutoff or breakpoint values. The closer to 180º the attackers are, the closer to 100% extra damage they do. It's very smooth.
I'm not sure about is how flanking damage is applied from bombers. I would guess 150% because they're hitting from above, which is a 90º angle. I don't know whether flanking damage is increased by attackers attacking from elevation. Like, if there's a marginal damage increase if two attackers are attacking at 90º laterally, and also one of them is on a hill, so it ends up being like 100º. I suspect that is indeed how it works, though!
In practice, it feels like every battle is microable, and you're always trying to find ways to get damage from multiple sources from multiple angles. For example, you could have a small cloud of ticks surround a big strong enemy unit, and also have an actual damage dealer attacking it - then that unit will go down way quicker because the ticks activate the flanking damage from all sides, so everything is multiplied by 2.
There's definitely ample chaos in BAR. The fact that you can't just count how many shots it takes to kill something makes it far less predictable, and forces a lot more emergent decision making.
Other cool things: 1) Units that die leave metal behind, which can be reclaimed or in some cases resurrected. The wreckage itself acts like destructible walls that block pathing both of units and attack projectiles. This too leads to emergent decision-making during fights. 2) Metal deposits aren't neatly concentrated in base locations. They're spread all over the map. So raiding and runbys are a defining feature of gameplay in just about every match. 3) Windspeed is variable and partially random, so macro build orders can't be the same every game. Windspeed values are map-dependent. Some maps have more wind, some have less. (If you want macro build orders to be the same every game, you can always play on maps that don't have wind.) 4) Navy and Hovercrafts.
It's a completely different system than Starcraft or any of the Starcraft clones. It's something very special.
The original creator of the RTS engine Spring (Beyond All Reason uses a fork called Recoil) later joined Massive Entertainment and became one of the programmers of World in Conflict.
I think wind is a bit more interesting in Total Annihilation, as it does not only affect energy generation, but also projectile trajectory like in Team17's Worms series.
The original creator of the RTS engine Spring (Beyond All Reason uses a fork called Recoil) later joined Massive Entertainment and became one of the programmers of World in Conflict.
I think wind is a bit more interesting in Total Annihilation, as it does not only affect energy generation, but also projectile trajectory like in Team17's Worms series.
Hard disagree. That's a bad application of RNG imo
EPT
