Hi All,
I wanted to give a quick update on my progress with the Deerhound/Wolves/Deer update.
The Testing Branch has been unusually silent on this update due to the fact that I need to deliver these three creatures as a unit as they all work together. This means I have needed to be 3x the work before I could show the results, which is what is taking extra time, but I am getting close to having something ready for the testing branch.
So here is a high level account of what I have been up to over the last month to keep you all in the loop. I apologize for it taking this long. The systems are very complex and I have been heads down working instead of communicating as much as I normally do. These are exciting systems to work on.
What ended up happening was that my initial work on the Deerhounds was not producing the results I wanted gameplay wise in the first week. Without the wolves and deer, the deerhounds did not have enough of a purpose to make them interesting on their own. So I paused on the Deerhounds and began work on the wolves, which were meant to be a bit of a surprise, but since this update is taking a while, I decided to announce them directly, now. Wolves are coming.
The wolves were very complex because they have pack behavior. They needed to have essentially a shared brain sometimes, but other times do their own thing. The wolves also are both fearful or man but also will defend their kills and pups, which makes them have more AI states than the usual see man and flee mechanic of bunnies, rats etc...
Wolves are now almost ready, and I will continue to add behaviors to them as the testing branch progresses. They can hunt as a pack and keep themselves alive cooperatively. Over time I will have them defend kills and their pups as well. I will also start on the stealing sheep, goats etc. soon (preferring the young and old ones)
Once I started working on the wolves, I needed them to have a main prey item, and that is where the deer come in. Happily, the deer's herds and wolf packs are very similar AI wise. Deer are less complex, but also need to generally move as a unit. Wolves tend to use deer numbers as an indicator of when they should enter the map as well. They flow in with the deer for the most part, but will also stay around if they find other easy targets like sheep (this is on the todo list right now)
I have done quite a bit of research on medieval wolf/human interaction and some of it was quite brutal. The 1300's were around the height of the wolf hunting times. In some places taxes were even paid in wolf tongues as a method to keep their numbers down. How the Clanfolk deal with the wolves will evolve over time. I am not that interested in collecting tongues or having a yearly puppy hunt (Which was a thing) There was lots of use of poisons, traps, nets, and though there were rules about hunting most other animals at the time, with wolves it was anything goes. So they are a very rich historical element, but yeah, I think I will pass on the puppy hunts :) It is very interesting reading though.
So once I had the deer spawning, I decided to upgrade the world spawner for the animals to take seasons into consideration for both the animal numbers as well as when the babies will show up. Right now just the deer and wolves are hooked up to this more advanced system, but I will continue to add more creatures to it over time. This means in late spring and early summer would be the only time Fawns or Wolf Pups would appear. It is a neat thing to see when they arrive.
So in review, week 1 was getting deerhounds in there, but it ended up being mostly a design dead end as they behaved like large versions of cats. I want each animal to have a unique purpose, and the deerhounds needed one... so I paused them and started on wolves. Week 2 and a bit was wolves. This time was mostly figuring out how to handle pack hunting and figuring out how wolves would come and go from the maps. I also wanted wolves to be threatening but not a chore. It needs to be ok to manage them vs killing them, if desired. Deerhounds were to be the answer to managing the wolves as well.
The second half of week 3 and a bit of week 4 was getting the deer working while also upgrading the world spawner, but the bulk of week four was circling back to the deerhounds now with a new tool in my pocket, the wolf pack.
It occurred to me that an owner and their dogs works much like a pack and they can work as a team. This solved a large problem that I was having in that I just did not know how I was going to tell a deerhound to go kill a deer. You can't just give a dog a job and expect them to do it without their owner being there...
So then I thought, ok there would be a special hunting case for when you had dogs and that would trigger a person to come along as a guide (Which was going to be really confusing/lame) But packs were perfect.
How a pack works is that the pack detects one of the members going on a hunt, and the rest of the pack decides to help if they can. So now that is exactly how it works with a hunter. There is no new interface needed. If a hunter has a dog, that dog knows they are hunting and comes along to help in their own way. And boy o boy do they help.
Deerhunting with a bow is an all day affair that can span the whole map as the hunter chases and misses, spooking the deer. On the other hand, Hunting a deer with a bow AND a deerhound is a very efficient prospect. Multiple hounds are even better. So I am very happy with this. Prior to firearms, everyone wanted a deerhound on the homestead, and now I know why.
This new hunting behavior also required new AI behavior for the humans to deal with new situations, like a downed animal. By now you can probably see where the month has gone. The first time a hunter took down a deer with the dogs, the dogs backed off and then the hunter stood 10 paces back and took pot shots at the deer with the bow, sometimes a dozen times, until it was dead... So yeah, there is a new "put down" logic to avoid that silliness. Once the dogs down the deer, the hunter will finish it off by walking up to it, and then take the carcass home to the butcher block. A hunt is a pretty detailed series of AI tasks, but it seems to work well and looks pretty good.
Now I am getting into the last few nice to haves before the testing branch goes live. So I spent some time getting realistic wolf and deer sounds, as well as deerhound audio, but I ended up recording my own dog (standard poodle) which sounded oddly similar to the deerhounds when I pitched the audio down 5%. The up side of recording Max (my dog) is that I have virtually unlimited sounds I can record based on his moods, and then I can get these into the game for the deerhounds. Also since all the rest of the Clanfolk human audio are family members, I figured the dogs should be too.
I also need to make things like dog beds and do some testing of the day to day life with the dogs. Hunting is tested in sandbox settings, but when it comes down to day to day life there are different problems that appear. Here is an example...
So it turned out that I solved the problems with doors and cats with cats being liquid (as everyone knows) but huge deerhounds do not fit under doors... So how does the deerhound get into the house to sleep at the foot of the bed. Well another half day went to keys inheritance. Deerhounds will inherit the key settings of their masters. This means we can be assured that they will remain loyal companions, near their owner. So far it seems to work quite well.
There is also the question of setting the master of a dog. Since the dogs act more like fuzzy slobbering tools, they need to be able to be assigned. Dogs will pick their human like cats do, but this will also be able to be overridden in the case of dogs to prevent frustration with none of the dogs linking with the one hunter. This I still have to set up, but will likely have working before the first testing branch goes live. It is all these little things that eat up time.
Another fun new system is carcasses. When cats kill a bunny, the bunny vanishes... When a pack of wolves kill a deer, the deer does not vanish. The carcass is now a temporary feeding and defense territory for wolves. Human want to collect carcasses, so I need to stop them by default, else the human would wander into a very bad situation as well.
Right now Wolves will not attack humans, but eventually they will. I am keeping the interactions between dogs/wolves/deer for the first testing branch release, and will slowly add the human interactions because it will add a whole new layer of AI complexity. Over time the wolves eat the carcass, but if a human decides to steal it, they can. The resulting hides and meat will be lowered based on how fed upon the carcass is. The act of feeding also spoils the carcass, so you need to get to them quickly.
I also had to upgrade the AI rate for fleeing and hunting animals. The tick rate for thinking of non livestock is once every 2 seconds, which is fine for wandering around and pooping, but during a chase it lead to constant pauses and misses. I upgraded the task system to allow it to bump up the fleeing and hunting animal AI to human levels to keep it snappy. This is another behind the scenes upgrade that was cool, took time, but isn't easily apparent. Cats are likely ultra deadly now and will need a re-tune.
Another thing was an upgrade to the threat assessment system for long range travel. This system still is not perfect, but now with wolves being a thing, I needed to prevent deer from picking greener pastures that were full of wolves. So this new system detects the livability of end positions when doing a long range wander to a new position with a pack or a herd. (Wolves also do not want to move in next door to people) As they wander, they also continually assess if their path is taking them past somewhere nasty and reevaluate. The wolves introduced a lot more and mobile threat to the environment, so the wander and flee tasks needed an upgrade.
Wolf howls are just cool, and lead me to upgrade the ambient sound system to understand time of day as well. This means that the wolf packs will begin to howl in earnest in the evenings, and roosters will crow a lot more in the mornings. In working on the audio system, I also decided to finally give Foxes a voice. They still need to be moved into the world spawner and will get some hunting AI of their own, but they sound like foxes now :)
The sound environment is a lot richer now with the wolves and red deer. Both of these animals have long range calls and fill out the environment well. The deerhounds playfully bark and the foxes yip as they go about their business. I am going to continue working on this as well as atmospheric audio is very important to me, but it will be ongoing as the testing branch progresses. I have happy, sad, confused, hurt, angry sounds for the dogs that I would like to get hooked up to the right behaviors. I think it will add a lot of personality to the dogs as they follow their owner around all day.
Testing branch wise, I am hoping for mid next week to have a version live to test out. This version will not be near feature complete, but it will be playable and will not break your progress, which is my main concern. Once that version goes live, I expect to go back to testing branch updates every few days again as I make fixes and we talk about ideas for this new system.
Thanks for your patience everyone. This system has taken longer than expected, but I am doing my best to make it worth the wait :)