Lionel's Blog

I'm the head of engineering at Tegus, a financial research startup. I write about politics, psychology, and software.

I've worked on a few open-source projects and infrequently blog.

Recent blog posts:

Predicting the Success of Pair Programming Interviews

The Computer Boys Take Over by Nathan Ensmenger

Who by Geoff Smart and Randy Street

RSS Feed

Game Theory in Heroes of the Storm

05 December 2015

I’ve been playing a lot of Heroes of the Storm lately and really enjoying it. On its surface HotS is all about fast reflexes and good teamwork. If you dig deeper, however, good play also involves mastering some fundamental concepts that aren’t immediately related to landing your skillshots or comboing your heroics. In this article I’ll lay out a few of these concepts, as I see them. I’ll assume the reader has some experience with the game – if you haven’t played you might want to try it before reading (it’s free).

Strategy and execution

Imagine you no longer had control of the moment to moment execution of your hero’s actions. You could generally direct your hero to “push out the bottom lane”, “join that team fight and focus on Morales” or maybe “dive to steal the enemy boss”, but wouldn’t be casting individual spells or controlling movement.

How would you decide what to do? I find this thought experiment helpful because it forces me to separate questions of execution, or carrying out a play correctly, from questions of strategy – deciding which play to make. At low level games, consistent execution is enough to win. But at higher level games, you have to think more carefully about your strategic decisions.

Risk and reward

How do you decide what to do? The simplest way to decide is to weigh the risks and rewards of different strategies. Some actions, such as pushing out a minion wave when the other team are on the other side of the map, are low risk. But they also don’t bring your team much advantage. Other strategies are high risk but also bring great rewards. Which should you prefer?

As in economics, in HotS good actions give more rewards for less risk. Typically, though, we’re faced with choices where the ratio between risk and reward is more or less constant. You can push to destroy a keep and risk being wiped or you can get a bruiser camp on the safety of your side of the map. You can dive to steal the boss or go kill your own boss and trade. Because the ratio of risk to reward is relatively even between these choices, one isn’t obviously better than the other.

Deciding on a strategy often boils down to this: given that you usually won’t be disproportionally rewarded for taking on risk, where do you want to be on the risk-reward curve?

This sounds like a question of personal preference, but it isn’t, or at least not entirely. Two critical factors influence where you should prefer to be. They are: who is ahead? And who is better?

Being ahead

Many decisions in games revolve around “being ahead”. Being in the lead is a tricky concept to define, but it can be thought of as something like this: you’re ahead if you will win the game unless its trajectory changes.

Notice that being ahead isn’t a question of relative power between teams. Being ahead is about inevitability. It’s about which team has time on their side. The team that is behind must do something to change the flow of the game, or else they will lose. The team that is ahead only has to defend their lead.

That’s why being ahead impacts preferences along the risk-reward curve. If you’re in the lead, you should prefer less risky plays, even if they have correspondingly less rewards. You want to reduce variance in the game – eliminate surprises and lock down your lead. On the other hand, if you’re behind, you must make high risk, high reward plays. If you don’t shake the game up, you’re going to lose. The farther behind you are, the more risk you’re forced to take on if you want to win the game.

People misunderstand this dynamic all the time. Throws are very, very common in Heroes of the Storm. Any time a team with a significant lead wipes at core trying to end the game at level 15, they’ve made this mistake. They took a 50% chance to win the game immediately over an 80% chance to win the game 5 minutes later.

It’s worth pointing out that while throws are dramatic, obvious moments of misjudgement, incorrect play when losing is probably equally common. Conservative play when behind is essentially giving up on any opportunity to regain the lead. Pushing minion waves around little by little isn’t going to win you the game if you’re 2 levels behind – your opponents will just do the same thing and maintain their lead. In the long run, you’ll lose. When behind, you have to take more risks.

Being better

If your opponent is better than you, should you prefer a short game or a long game?

The answer is that you should prefer a short game. Games like HotS are, roughly speaking, a series of decisions made by players, with a bit of random luck and accident thrown in. A better player makes good decisions at a higher rate than a less skilled player; over a longer game, the better player has more and more opportunity to bring his superior skill to bear. Thus, as the game goes on, the better player is increasingly favored to win. (Another way of putting this is that the less skilled player’s best shot at a win is a game where variance plays a large role.)

Compared to, say, StarCraft 2, it’s hard for teams in Heroes of the Storm to control the length of the game per se. However, the “who is better” question still has a general impact on risk-reward preference. Essentially, the more skilled team starts the game out ahead in the sense I discussed above. They have inevitability on their side; the longer the game goes on, the more likely it is that their superior skills will lead to a win. Because of this, teams that perceive themselves to be better than their opponents should be more conservative, all other things being equal.

Conclusion

If you’ve gotten this far you’ll have a working understanding of risk and reward in Heroes of the Storm. There are a lot more game theory concepts at play in HotS - resource utilization and value versus tempo come to mind - which I might cover in another post if people are interested.

comments powered by Disqus