Judge feedback

Judge feedback is anonymous and shown in a random order.

• Unfortunately I couldnt get the project working locally so my judgement comes purely from the videos. Apologies if my views aren't accurate! It looks like a good technical project, but without actually using it, it's difficult to say for sure. It's an interesting project though more types of AI and more behaviours would be awesome
• The demo could use some dev/debug on-screen print on each agent to show what is happening in the system (list of current actions, path finding, visibility spectrum etc...). As it is a tool, a documentation on how to implement it would have been nice.
• Very nice idea to make a predator/prey AI and tool chain. Like the blend of GOAP and Utility AI, to try and mirror real life. If a prey animal MUST drink, then eventually it will risk being attacked by a predator in order to get that drink! Like that you're using the graph tool in unity to control the values and it all looks quite designer friendly - not too much data for them to balance, and you've made it visual (the graphs) where you could! We'd do exactly the same in our tools to make them as useable as we possibly can! I'd be very interested to see how this scaled up with lots of AI all interacting, along with some more AI techniques (Genetic algorithms/nearal nets) to vary the characteristics of each creature, so that the world could evolve. A players action would have more consequence if they accidentally stole all of the only food a particular animal ate - leading to a collapse of the foodchain. I think you'd be really interested to see Eco on steam :) Presentation was good! The dev vlog was a great idea, and nice way to demo and talk about your project! A few hints and tips when it comes to making tools - your designer will need ways to debug what's happening - just like we need a debugger to debug the code. I think they'd find it super useful to have some sort of toggle-able overlay for all of the AI's describing their state/utility values so the designer can see directly how they're changing. It's also very important they have a way of measuring the cost of their change in system resources - e.g. how much frame time/memory their change will take, so some specific profiling would also be extremely useful to them :)
• This project is impressive but I would have liked to see more about the tool's side of the project Maybe created two different scenes with different kinds of AI and show how the same code is changeable. Some sort of either editor UI or developer UI above the AI describing what's the AI are thinking would be helpful to any programmer, designer or QA to debug and clear any issues with AI settings. I've added an example of what this could be like to the project. https://drive.google.com/file/d/1hnRwru_YyipmOIH_fDrLO-SOyjsQs6H5/view?usp=sharing When I downloaded the project I was receiving errors while trying to test when initially opening the project, there were some files/classes with the same name as some already existing class, I deleted the class that had the smaller amount of code (since I thought they were old files) and it worked fine after that. I suggest when sending a project to anyone to delete any unnecessary files such as (.vs, Libary, Logs and obj) this can cut down the size of the project and these files shouldn't be included if you were keeping a project on some sort of source controls such as Git or Perforce. Overall I was impressed with the project and it is a good idea to have some AI code that works for a variety of creatures with different values set on how they behave. I think this sort of project would be great to make work for Unity's new ECS (Entity Component System) since each creature behaviour is just data on how they would react and you could have hundreds of AI on-screen without major performance loss.
• Some interesting ideas, which show some promising AI behaviour, but difficult to assess how versatile it is. I would have preferred a standalone build that allowed for tweaking and modification of the behaviour via a runtime UI. Your vlog, while good, could have demonstrated more variety and scale (more creatures, different types of prey/predator etc.).