Camelot Unchained now boasts a record 5100 bots and humans in a single battle

    
42
It's not quite so much of a model after all.

Camelot Unchained’s City State Entertainment has posted its monthly newsletter today, and it’s big news indeed: The studio’s just hit 5100 ARCs (autonomous remote clients – its combination of real players and bots simulating real players) in its test. Just a month ago, the company hit 3500 bots and humans on a single battlefield.

As we’ve previously covered, CSE does not actually expect to host this many players in a single battle; the Kickstarter promise was 500, though studio boss Mark Jacobs has intimated he thinks the final count will be somewhere between 1000 and 2000 by launch. The idea here is to push the system as hard as possible in testing to see just how much the game can get away with as the devs add more NPCs, siege engines, and VFX with a lot of people, but without requiring sky-high PC requirements.

The rest of the newsletter is long as usual, with a long entry on what the studio calls “characters 2.0,” an overhaul of the player characters now trickling out to testers; a look at zone design as enhanced by player builders embedded in beta one; and a run-down of updates to the ability builder system, which will ultimately allow players to craft their own skills and spells in-game.

newest oldest most liked
Subscribe to:
Reader
styopa

Is 5100 bots a record? Because I’m pretty sure Archeage broke that regularly.

Reader
PanagiotisLial1

Got the joke, but also makes you wonder on WoW bot across all servers, lol

Reader
Eliandal

I liked Archeage, I really did, but dang, that was good ;)!

Reader
Mark Jacobs

That indeed was a good one Styopa.

I should probably have them saying “Want to by g&&D?” right? :)

Reader
psytic

They should hire some former A Net folks. This game would be fantastic with some of their ex world building and art asset members.

Reader
Randy Savage

With that many bots, Mark Jacobs could be president!

Reader
Mark Jacobs

Hehe, I’d need a few more than that but thanks for the thought! :)

Reader
Loyal Patron
Patreon Donor
Kickstarter Donor
Dean Greenhoe

Just reading between the lines here… but what I envision is not 5,000 players but 5,000 physics enabled elements that all have to tracked and drawn during a possible battle scenario. Players, npc’s, plants, birds, siege, boulders, rubble, bodies, wagons, traps, smoke and any number of game elements. The more that can be handled the more realistic the battle.

Just my humble thoughts.

Reader
Mark Jacobs

Humble doesn’t mean wrong. :)

And that’s it in a nutshell. We aren’t planning to having 5K people in a battle but if we can render/network 4K animated objects along with the VFX/SFX going off, then we can handle 500 players in a keep siege or open-field battle.

The key for us is launching with an engine that can support the players we are expected to have in the game. We test this way now so we don’t find ourselves in a situation where we fail at launch because the engine couldn’t handle the stress. These types are great at finding bugs in the core systems/code and because we’ve been running them we’ve found/fixed/fixing problems that would have occurred later in the development process/launch.

As I and the other people who were in that test were ecstatic with what they saw. One of my fave bits was when one of the players asked if we supported VR yet. I said yes, told them the magic incantation on how to invoke it and they then ran through the crowds in VR. They were, to put it mildly, shocked at seeing 2K+ ARCs/players running around. Now, we have no plans for a VR version of CU, but it does show what our engine is capable of in terms of maintaining the rendering speed you need to support VR.

Thanks Dean, as always.

Reader
Oleg Chebeneev

This becomes a better meme than ship sales in SC

Reader
Bruno Brito

Where is this a meme? This is actually incredible.

Reader
Mark Jacobs

Bruno, it’s OC just being OC doing his usual drive-by snarky comment about CU. He knows the importance of this kind of testing but that’s not the point.

Next week, when our characters 2.0 update should be integrated into the main production branch, I’m looking forward to seeing how things look then. From what I’ve seen, it’s a nice improvement in quality and in rendering speed with better LODing.

Reader
Bruno Brito

Oh, i’m sure between Project Ascension and SC, we can find new crashgrabs for him to praise.

Random MMO fan
Reader
Random MMO fan

Glad to see their server can already handle so many clients, I just wish the rest of game development would progress a little bit faster.

Reader
Mark Jacobs

It actually has recently as we’ve been talking about in the updates. We’re still adding gameplay-focused engineers so I expect it will move along even faster.

Reader
Tamanous

The higher the stress test number, the higher the functional player limit is. Good to hear, as it means normal game play numbers have plenty of room to scale with the resource limits.

Reader
Mark Jacobs

Bingo, QFT, +1, etc.

Reader
rafael12104

Man, this bodes well for an average size battle. And that’s all I want. I don’t need a 5000 plus BR. Heh.

Reader
Mark Jacobs

Correct. As per my answer above, we are doing this kind of testing to expose problems with the systems involved as well as to point to areas that need optimization.

Though it would be amazing to be part of even a 4K battle.

xpsync
Reader
xpsync

I wonder what happens at 5101?

Reader
IronSalamander8 .

The universe explodes! :o

Reader
PanagiotisLial1

No, “only” the server! :p

Reader
rafael12104

We break the time-space continuum and go back in time?

Reader
Mark Jacobs

I don’t remember the exact number but somewhere soon after that the engine had enough and told us it was done for the evening. :)

The good news was that particular test showed us we fixed the problem that prevented us from large-scale testing on the weekend and also pointed to one of our mini-servers (physics) that might need some attention from us.

Reader
Bruno Brito

You should make it when everytime the engine tells you it’s done, it plays Despacito.

camren_rooke
Reader
camren_rooke

This is welcome news considering the problems I have with ESO pvp are as follows:

1. A queue to get into the pvp alliance war at prime times can take over an hour to get through.
2. Lag is so bad that you literally cant fire off skills or ultimates cause there is so much going on. Yes, the skill bar is greyed out.
3. There is so much lag at times, the pvp realm will crash you out and if you are lucky you are back in pvp realm when you come back in, if you arent lucky you are back in Tamriel with another hour plus wait.

Number three has seriously pissed off folks who just log off.

Here’s hoping for better from CU!

Random MMO fan
Reader
Random MMO fan

Yes, when I tried ESO I noticed major performance issues in PvP. I never crashed but was pretty disappointing seeing all the lag and my CPU only showing a single core being used most of the time, so I hope CU will have much better optimization for both server and client side.

Reader
Mark Jacobs

Without throwing any shade about ESO, problems like those and Tier 4 in WAR, are why we are doing things this way. And frankly, we’ve only begun to stress test the system. By the fall, I want us to have an internal build that will throw these numbers against the servers 24×7 for days/weeks. We could do that right now via the cloud but running 5K ARCs is rather expensive per hour since they are run off separate AWS instances. I’m looking forward to that.

Reader
Bruno Brito

Tier 4 War even in RoR is pirate ship meta.

Please, PLEASE, balance AoE.

Reader
Loyal Patron
Patreon Donor
Armsbend

Go for 10,000! Then 20,000! Then go for 40,000! Then keep going!

Reader
Tamanous

With the recent AO fresh server news, your statement reminds me of when Funcom was developing AO. They stated they were giving up on small server shards and shooting for one 10k+ prop server. Back in the day this was unheard of.

Sadly, they couldn’t manage it and split the servers into 2. It still had one of the roughest MMORPG launches in the history of the genre.

I suspect CU player caps will be set entirely based on game design reasons with reason.

Reader
Mark Jacobs

Tamanous , you are correct about the game design reasons for our server size. This testing is just for a battle but then we will see how many people we can support concurrently in one shard.We are not going for a mega-server because I don’t think that is a good idea for most games but we will certainly be higher than say Dark Age was and less than 10K. What we want to have is a number that when full allows people to play and have fun and not feel like there are too many people on their shard to have fun. Tough balancing act.