Comments section is pretty clear who read the article and who is only reacting to the title.
Helldivers 2 has weeks long campaigns where the entire playerbase has to defend certain planets. The "GM" keeps a close eye on progress of this event and manually adjusts it to make sure the campaign doesn't complete too soon.
There isn't some poor motherfucker staring at a screen watching the progress of each individual match spawning more bad guys whenever they feel like it.
This continuous evolution has actually seen Joel getting up at unsociable hours to sort out situations when the Helldivers 2 team realised the game wasn’t as balanced as it could be. “There have been some sudden moments where maybe one planet was too easy or one was too hard and [Joel] had to get up in the middle of the night to give the Automatons a bit of reinforcement so the players don’t take [the planet] too quickly,” Pilestedt said.
This isn't "The Director from L4D", this is the human touch of direct influence of game balance globally without the need to release a patch or even a hotfix, and way beyond just how many bots drop at a time. If you want to compare it to something, it might be The Wizard from Oz, pulling the cranks to drive the facade.
If The Director wasn't dumping enough zombies in general, Valve would have to patch The Director to make it do so because it's restricted to its coding limitations. A human Game Master can run it only limited to the variables the devs give him. Open new planets for plundering, change the weather, accuracy of your calldowns, the options are as limited as imagination and time.
Saying this, I don't know why they would leave this all up to Joel, you'd think there'd be a team of 3 to bounce ideas off or something.
This continuous evolution has actually seen Joel getting up at unsociable hours to sort out situations when the Helldivers 2 team realised the game wasn’t as balanced as it could be. “There have been some sudden moments where maybe one planet was too easy or one was too hard and [Joel] had to get up in the middle of the night to give the Automatons a bit of reinforcement so the players don’t take [the planet] too quickly,” Pilestedt said.
That’s honestly messed up and embarrassing for their developers. Writing a decent director shouldn’t be that hard.
This is the most absolutely buck wild and inefficient version of this idea I’ve ever seen, ever. Director AI’s have existed over a decade, and have already been used to solve problems exactly like these. If I was Joel and was woken up at 2am to go drop a couple more bugs on Leedle III I’d tell my entire management team to eat shit and do it themselves. What the fuck. Who came up with this? This is half baked as hell
I agree who the fuck thought this was a good idea? What happens if Joel suddenly kicks the bucket or is injured? Why is this not an automated system that has the option of being manually adjusted? This sounds like a downgrade from L4D.
No this is a reasonable approach. Arrowhead are a a rather small company of around 100 people and automating things is easier said than done. They also never anticipated the game to be as successful as it is so at the time it probably wasn’t high on the priority list. Now they pay Joel overtime (I hope) and can think about how to implement an automated script to adjust the game.
It sounds like he also writes the events. The article gives the example of one part ending faster than expected, so he added a bit about the forces pivoting to mine the planets. A director AI can’t make stuff up from scratch. (ML-AI could, but it wouldn’t be very good.)
Was there a level of “meta” in the Director of L4D? My understanding was that handled the flow of an individual match, but there wasn’t influences of the matches on other matches.
Since this is about the overall campaign that encompasses a level above the individual matches.
Not from what I recall. It is neat how each HD2 mission contributes to the state of the in game universe. Just reminded me of the Director from L4D taken to the next level.
As for Control 2 and project Condor, both were announced back in November 2022. They are, however, likely still some significant way off, with Control 2 confirmed to still be in the the “proof-of-concept” stage last October
Sounds like we are still a ways out from Control 2.
While I did like the gameplay and environment of Control I really don’t think the story was all that memorable. I don’t know how they would out do the original.
eurogamer.net
Gorące