about 2 years ago - not_so_gladiator - Direct link
Hi Nomads!

As one of the most important elements of Last Oasis, creatures and AI in general were one of the very first things we started with when we began planning the LOverhaul, so it felt like the right thing to dedicate our first devblog to them.

In this devblog I’d like to tackle some of the changes we’ve done to the old creatures, new creatures coming to the game, design choices we’ve made for them, some of the technical challenges along the way, and what’s in store for the future when it comes to creature behavior and design in general.

For these devblogs I will always be approaching the developers in charge of designing and implementing any changes or additions we talk about here, so we might be getting into the technical aspects of these features a lot. I’d also like to point out that some specific terminology was used in some cases that probably doesn’t mean much to you now, but as we continue to release more devblogs it will make more sense. An example would be the reference to tiers with mobs, this will be explained in-depth in a future devblog.

First part of the devblog is dedicated to the design and direction side of things, and getting into a bit more detail about Gogo and his minions, as well as showing off some of the new creatures coming to the game. The second part will be much more technical with a word from chadz, but hopefully readable still. :)

DESIGN AND DIRECTION
As you know, until now, our creatures didn’t really have a lot of interesting behaviors or gameplay related to them, with some very few exceptions, our systems simply didn’t support that, since before EA release of the game in 2020 we didn’t have enough time to polish them and create a much more satisfying experience.

Most of them were not really fine-tuned toward certain experiences, rather living in the oases mostly as resources to harvest and small challenges for players to fight for.
Due to those reasons, we had quite a lot of technical debt to clean up before getting to work on a better system in order to have a base that was easily expandable and worked properly with new mobs being added in the future, or changes to old ones.

Having a system that makes it so a designer can easily change what the mobs are doing, with precision, was a massive consideration for the rework. But before we talk about our system changes, I want to give the word to Lukasz (Klerych on Discord) and get into the design side of development for mobs, and their new variations.

While the new systems were being worked on, we were also making sure we have a proper approach for designing gameplay around mobs and AI in general.
Until recently, we didn’t have a very formalized approach for creatures and how they fit in the world of LO, but now it was pretty much a requirement as we intend to make them one of the core aspects of the game - mobs are essentially connected to all the new systems and progression coming in S5.

Hunting them, acquiring resources related to them, etc. is going to be a major part of next season’s gameplay. Our intention is to have a progression that works in a way where you always have some bigger threats to tackle and defeat (or be defeated and find another way to acquire those resources, for example through trading).

Early on in the game, the mobs you encounter will be similar to the ones you encounter now, Rupu, Phemke, and the likes, but with new twists and behaviors that connect to our new technology acquisition system and the changes to Walkers (this will all become much clearer once we showcase these changes in upcoming devblogs, for now I can only mention it lightly). An example of the progression through hunting mobs is how initially you will need to find and hunt down Phemke in order to get to the next Walker and progress forward onto larger creatures and much bigger threats.

For Season 5, the creatures in LO will be split in two important groups - there will be wild versions of the creatures, some of which either roam the oases, or guard their lairs (which will be points of interest to acquire special resources), and a much more dangerous, tougher version of them, internally called “Battle Mobs”.

These creature variations will come equipped with battle rigs, designed to put up a fight against players on their Walkers. They will be much tankier than their wild counterparts, have a variety of siege weapons available to use against you, and possess a different AI and aggro system that fits Walker combat much better than fighting versus players on foot.

I asked Lukasz for an example to give you guys of an encounter with one of our new giant creatures I teased last week, “Gogo”, also known as the Mountain Crawler, and talk a bit more about how the design process goes for when we decide where mobs are placed, how they are balanced, how they fit in the world of LO, and more.

The Wasteland is the center region of the Kali Spires map. Until now, it was mostly empty and the majority of threats were other players roaming and looking to empty your Walker storage, on S5 that will be quite the opposite.
The Wasteland is now going to be inhabited by packs of Stiletto-sized Rock Crawlers (two or three per pack) on the edges, while closer to the center will be home to the imposing Gogo, the Mountain Crawler, equipped with heaps of Rupu-operated guns on it’s rather large back.

Gogo’s design is rather simple, it’s intended to be a massive open world raid boss, one of the highest difficulty creatures in the game before you meet sandworms. Although the creature itself is not very territorial, the Rupu that made its mountainous carapace their home are, and they’ll engage anyone that tries to get too close or is foolish enough to attack their monstrous mount.
Even at its lowest tier, Gogo will be extremely tough, with one of the highest armor ratings in the game, meaning that it will take some heavy weaponry and possibly require at least a few Capital-class Walkers or several Large Walkers in order to defeat it and acquire new resources and high tier loot (among other things, defeating Gogo will be a requirement to acquire one of our Capital Walkers, the Titan).

When it comes to design, the main decision driving process for Lukasz is mostly the immersion and the experience they can create, whether it be readying up your Walkers and going on a hunt for specific creatures, or avoiding them as you live your life in the oases. One example is our OKKAMs, they would not fit very well in Ancient City’s wetlands, but they are a great foe to battle in the ashlands region of the Sleeping Giants oasis.
The creatures that appear in each part of each biome/oasis, it should really feel like they could live there. Because, you know, they do live there.
Accessibility is also a massively important part of this. Recognizing dangerous zones as a player, knowing where to look (and avoid), for these creatures is very important. You wouldn’t want to make a base in the Kali Spires’ wasteland in the center of the oasis, with giant Gogo and packs of mountain crawlers still roaming around.

It’s also important for us to be able to bring more content, more gameplay to maps that don’t have much of that, a great example of this being the Kali Spires, which will be home to new threats and rewards for players to find and fight their way through. I already mentioned Mountain Crawlers and the smaller variations, but we’ve also made some changes to the Rupu tribes of the Spires.

To close off the design part of this devblog, I’d like to leave you with a few pictures of some of the creatures we’ve been working on:



The Papak are one of the creatures you can now expect to find on the lava lakes of the Volcanic biome. A rather odd-looking, frog-like hostile creature, Papak use their long tongues to attack nomads and their Walkers from a distance, usually leaping away when Nomads in melee range. This variation of Papak is an agile, fire-breathing, jumpy creature that chases down Nomads looking to make a Flot off obsidian harvesting near their lairs. They also have a tendency to explode when in near-death situations. There are other variations that live in different biomes and have different attacks and behaviors, such as the Mist Papak, a variation of the creature that resides in dark and damp crevices where mushrooms can be found.



The Koa are large, elephant-like creatures that peacefully roam the forests of Sleeping Giants biomes, and are usually found wherever there are large lakes, and big patches of trees. Peaceful creatures by nature, they generally avoid Nomads and their temporary settlements, but when tamed and trained by the Rupu, they become a fearsome and tough challenge to take on. With a rather large amount of space on top of them, it is the perfect place for Rupu to build their War Howdahs.

AI SYSTEM CHANGES
We’re not done talking about mobs yet, and definitely not done talking about our new Battle Mobs and the angry Rupu they carry, but for this second part of the devblog we’ll move into the more technical aspects of the AI and creature changes.

For the past few months chadz has been working on reworking pretty much the entire AI system from scratch, cleaning up some of the technical debt accumulated over time, and came up with a much more stable, expandable system.

With this system in place, a lot of our current and future work is made much more efficient, and involves less developers working on the same feature, also allowing designers to change mob attributes easily, add behaviors, and more.

Our main approach was getting an AI that works properly with Walkers, which is a very challenging thing to get into, as there aren't really any references from other games at all.
One of the first parts that got changed was the aggro system for mobs. While it sounds very easy (you hit mob, mob gets angry) - when you get into all the details it gets very complicated very quickly. You have different players with different walkers at different distances with different damage potential.
Our new aggro system might also have some complications still, but it’s looking much better and mostly ready to be tested by real players in real situations. It was also important to make it very unscripted so that it fits within our sandbox experience.

Movement work was also another difficult part of the rework (don’t talk to chadz about creatures moving backwards for now please), we’ve always had a lot of issues with that in the past, but most of it should be gone now. Mobs should be much more accurate at plotting trajectories toward Walkers and players on foot, recognize actions the player is doing, figure out the speeds at which Walkers are going (this was particularly important otherwise it’s rather easy to exploit creatures into doing specific things to cheese them, like wiggling left and right to confuse Nurr before he attacks).

We’ve also implemented a much better targeting system for mobs and their Battle Mob variations, as this was actually the basis for the entire AI rework and a massive requirement for this feature to work and be fun.

One of the key intentions with the rework is to make the creatures smarter and stronger, but stronger in a fair way. They very specifically try to actually be in a range where you can attack them with board weapons and such. They are meant to be a tough challenge, but still a fun and fair fight.

While there is still lots of work to be done to improve these systems, we have built a good, proper base that will help make the work on both the old and the new creatures coming to the game much easier. We’re talking a day of work for one or two developers, compared to a week of work for several developers in order to implement a new creature, or make changes to their movement or behaviors, etc.

We’re not gonna get it right on the first time, nearly guaranteed. But this time there’s a difference, the technical debt has been cleaned up, and the whole thing is expandable.
It’s always been so painful for us to see suggestions but us not really being able to implement it in an efficient way. This time people can actually bring suggestions and we can properly work on them without spending a lot of resources, thanks to these new systems. We can actually respond to feedback without having to work for periods of time that wouldn’t be acceptable to us, or involving developers whom are already busy with other work.

Any improvements we do to one mob, it’ll be an improvement to all mobs, which will eventually work towards an actual living world. Because of many mistakes we did in the early development, that just wasn’t possible before because it was a very rigid system. A lot of that has been cleaned up now.

I also wanted to touch on the differences between creatures like Rupu, and something much more complex like Nurr, Okkam, etc. so I asked him to provide some insight into the technical differences between them.

The big difference between these two is collisions.
Humanoids have very few interconnections, you don’t have stuff like one humanoid pushing another humanoid that pushes another humanoid. They don’t push themselves around. But Walkers and physical mobs, with all kinds of different shapes, need a different approach to collisions. For example Nurr being long and thin, Okkam is a big four legged creature with empty space between the legs, etc. And then you throw the Walkers in there, environmental obstacles, and you try to make that work together.

At the very least they need to collide and react accordingly without getting sent into low Earth orbit immediately after touching each other. Other games have quite straightforward approaches to their mobs. Each mob checks if it has an obstacle, and if not, it’s allowed to move. We can’t really do that because we have so many physical actors. We need a system that’s very robust with all kinds of mechanics.

The big difficulty here is making all the physics interactions work. If you can avoid it in your game, don’t make physical creatures. It’s a world of pain. In our case, we can’t really avoid it, unless we wanted to make our creatures very very simple, almost boring and non-interactable.

You’ve probably seen it in other games where vehicles suddenly launch into the air with intense velocities. That’s usually a result of a physical object (vehicle) suddenly trapped in a completely static object (environment or non-physical mob), therefore only the physical object will violently go very fast to resolve that collision.
Of course it’s happened to us a lot, but we’ve come to a point where it’s so much better because all objects have physics, therefore the systems know how to resolve it.

For the most part, the changes to these systems are a lot less visual than other changes in LOverhaul, they’re not something you can properly look at and see the work done in the background and all the things that make it tick, but we hope the improvements toward the gameplay around our mobs will showcase the work on the background.

I know this is a rather long post, and I would’ve loved to do more visual content for it, but I also wanted to get it out to you guys quickly and talk about the technical aspects of it, not just design and showcasing new creatures. I’ll leave a small FAQ below with condensed information:

Battle Mobs, what are they?
Battle Mobs are essentially boss-type creature variations. Rather than the wild versions of Last Oasis mobs, these have been tamed by the rupu tribes. They come with big (and small) howdah mounts on top of them which can have a wide variety of siege weapons as well as armor plating. Some small ones use mostly short range personal weapons like javelins and the like, but the larger ones will come equipped with larger weaponry such as rupu versions of Catapults.

With these creature changes, are there any coming for the Sandworms?
Yes. Our current Worm is balanced for a rather small engagement, 5-15 experienced players can make short work of the Worm and not incur on many losses in terms of Walkers.
To us, the Worm is part of the game’s endgame gameplay, and we want to reflect this with more balance changes for it.
The Worm will become much more of a final boss battle in Season 5. There will be no certainty of returning from a Worm Sighting event map alive and with all your battle walkers ready to fight, none whatsoever. Our intention continues to be that the dangers in the map would be so harsh as to intentionally make it not worth it to fight each other in the early stages, and only really think about betrayal after the loot is within reach.
We’ve got a lot of ideas for new Worm attacks and mechanics, and while they might not be possible to get implemented during S5, we know we can balance the current Worm much better even without them.

On a final note, next week we’ll continue with our second devblog, this time dedicated to changes to siege weapon design, how they affect gameplay related to hunting mobs, and how it ties into our new raiding balance changes. We’ll be showcasing some changes to the fundamentals of siege weapons, including some additions to the controlled weapon arsenal. Because if we are to show you new mobs, we also need to show you what kind of guns you’ll be shooting at them with.

—Neon





Recent Last Oasis Posts