No sharding?

We have not received permission from Depraved to share the quoted post.

It was also made on UE3 and was barely even optimized to run those kinds of fights. Yes, there was definitely lag, but considering smth like WoW that allegedly broke servers when too many people were in one place ("too many" meaning a few hundred afaik) - L2 was a god damn miracle of dev engineering back then.

And iirc Steven (or someone else) said that UE5 is much better for mmo-sized fights than the previous versions. And I think I heard smth about them make good progress on their netcode too? Not sure about that one though. Oh, and UE5 is just optimized much better overall.

In other words, there's a high chance that we'll see low amounts of lag during big gatherings.

We have not received permission from worddog to share the quoted post.

This vid has been posted in some other thread too, but still. This is probably 2-3 hundred people (maybe a bit more). All in one location (pretty much instanced) in a game from 2003.
https://www.youtube.com/watch?v=SgBQ1Vvx5HI
Here's another 300+ (144 on just one side) people now in not an instanced location
https://www.youtube.com/watch?v=yLrccSBRG_w

Netcode, optimization, player hardware - all have become much better since those days. And Intrepid are pushing them even further (allegedly). You need shards on poorly optimized games that weren't built for this shit. BDO should've been a great representation of that, but their engine is a bit fucky afaik, though even then, I'm fairly sure their sieges happen on just one server so they still do it.

We have not received permission from worddog to share the quoted post.

I was giving an example of what was possible to have in one location nearly 20 years ago and said that tech has gone a long way since then.

Here's 1300 bodies animating a fight in one place (1:08:00)
https://youtu.be/HB7gWTpeB08?t=4083

And Intrepid is still working on bettering this. Again, tech has come a long way and good devs can push it even further. And afaik there's some potential server (netcode?) trickery that lets Intrepid to have seamless splits of a location onto several processing servers, so that the overall load is shared across them. But I'm not well-versed in that and don't remember where I heard that from, so take that with a huge bag of salt.