The State of Heroes of the Storm A.I

Hello, and good day to you all! Find below a few words from Kevin Gu regarding the state of Heroes of the Storm A.I.

My name is Kevin Gu, and I am a Senior Designer on the Heroes team. I have been working alongside Jeff Beaudoin, our Lead Gameplay Engineer, on our game’s A.I. We wanted to take a moment to update you all on where we are and how our three major A.I. initiatives are progressing.

Reconciliation of the Goal and Tactical Systems

There are two systems under the hood of Heroes that work together to drive agent behavior: the Goal system and the Tactical system. The Goal system is responsible for determining where A.I. agents should move, while the Tactical system provides agents with awareness of their surroundings so that they use and target their abilities intelligently. Last year, in an effort to improve agent in-combat positioning, we made it possible for the Tactical system to override the Goal system. Although we believe the changes were ultimately a good iteration, they did expose several underlying issues that we are addressing.

Moving forward we will be reverting some of changes made in the last major overhaul and restructuring the Tactical A.I. behavior tree. We are reinforcing the intended separation of the Goal and Tactical systems, and will be paying closer attention to make sure that any future changes do not disturb this delicate balance of power. This will reduce the frequency of agents being pulled in two directions at once, most commonly observed as agents rapidly bouncing back and forth in place. This work will also allow us to more quickly diagnose the root cause of misbehaving agents.

Laning and Defense

Our second priority is to improve A.I. agents’ understanding of “laning” and how they approach pushing and defending those lanes. In the most recent patch, we added logic that allows the Goal system to understand that lanes pushed closer to a core are usually more valuable to defend. Our next step is to include units within each lane to further modify how valuable lanes are to defending agents. Both pieces will ensure that the A.I. will better understand how to defend their lanes.

Goal Staging

Our third priority is to improve the ability for agents to prepare or “stage” for their assigned Goals. Our internal Tactical A.I. tree can now better determine if all assigned agents to a goal have arrived. This allows agents to intuit if they should wait in shrubs because their team members are not yet near their collective objective, and subsequently helps agents work together.

Thank You

We want to extend a big thank you to those who have shared feedback on A.I. We hear you! We are hard at work improving the A.I., and plan to bring the items mentioned above and more into the next major Heroes patch. Again, thank you for your continued patience, and we’ll see you in the Nexus!

3 Likes

Nice

Still, I would like A.I to be used to handle team comps balance on QM queues better.
With the amount of data available it should not be dificult to balance comps around - with NO queue time increase, many many times a random 10 people QM round could have a way more balanced round if the system had switched 1 or 2 people around.

It would be nice if AI could play overall more defensive. AI dying is usually more of a problem than AI not contributing.

Especially before teamfights if they could sense that their team is grouping up and try to stay in the middle of that group rather than ahead of it. Ideally you never want your AI to initiate a fight.

First Issue: You made it VERY hard to comment, if that was not intentional than something done goofed, if it was intentional, because that is the feeling it leaves, well someone new what comes.

Second Issue: The AI is plainly horrible. No two ways about it. The Elite AI feels Just like Beginner AI, you can hardly feel the difference, and both are worse than simply bad.

Third Issue: AI is NOT JUST the Mercs and Heroes. What about a Nova’s clone AI. It has been reported as a bug for way over a year now. By this point in time i guess it’s a feature of the New and Improved AI. The Clones are simply mentally challenged and we have the game-plays galore to prove it. If you’d show us how the Clone ‘thinks’ (semi code-decision-tree) maybe we could work around it…

Fourth Issue: Yes, AI learning can drastically improve AI Heroes, OR a bad programmed AI, perchance with two decision ways, can lead to an AI that walk forth and back and does nothing… Which is always fun. Back in school my teacher would have told me for that project: “Six. Sit.”

5th Issue: I am really unsure, whether the New AI was introduced with the idea to give us a better AI or to punish the team with the leaver back in the day, cause that was what it felt like. Please take heed, never introduce something untested, which IS what you did. Or someone at a Very Critical point in testing failed and said, yeah lets release, I am proud of what I created. “One mans treasure is another mans garbage”, not the correct saying but think along the lines of hubris and ones own ideas and projects are always the best. This is what the contemporary AI feels like to most who remember the older one. Or maybe its just me…

Final Note: You apparently created a nice SC2 AI, although its kinda super-human. You have the Data, you have millions of game plays, heck conspiracy theory thinks some players are actually bots in a behavioral study experiment while simultaneously training you AI in a real game, so why is the AI SO Bad (AISOB).

Truly Final Note: Maybe someday someone will fix the AI, and we wont have to trees colliding in the AI tree making it walk back and forth doing nothing, but meanwhile it would be nice to fix MM in QM. You have the big data, unless artificial grind is truly a thing you want, because most games (if the player is in the same condition) are either you completely stomp or get stomped. It is unbelievable that you can in several game 1v5 wipe team and the next several you lose 1v1. This is not balance. And Yes, There are feeders, which you can’t carry. There is NO 50% force win / lose, but there is bad forced MM, which often decides the game beforehand, which is either intentional or nobody thought of MM especially in QM - which if i look at play-histories for most players is The Most Played game-mode. Tough there might be bias that you mainly see players in you match that also often play in QM.

Sorry for the wall of Text and the Bad English, dyslexia makes it even harder. But you really had to go through many hoops to comment, especially compared to post where it asked you to comment, this had build frustration :(, besides the AI had me riled for months (Nova Ex-Main, used to accept the clones, now they are worse).