Finding the Tao

If you go looking for definitions of the word “strategy,” you’ll find a lot of definitions that are domain-specific, that only concern themselves with a specific slice of life.

The American Heritage Dictionary, for instance:

1. The science and art of using all the forces of a nation to execute approved plans as effectively as possible during peace or war.
2. The science and art of military command as applied to the overall planning and conduct of large-scale combat operations.
3. A plan of action resulting from strategy or intended to accomplish a specific goal. synonym: plan.

“forces of a nation,” “military command.” This would be bad news for Richard Rumelt, professor of strategy and author of Good Strategy, Bad Strategy, because he mostly concerns himself with business strategy. But strategy does in fact have a place in business, so the American Heritage Dictionary is, alas, wrong.

Does strategy just mean “getting the right answer”? Probably not. Here are some problems that are not strategic:

• What is 7+7?
• Who pitched for the New York Yankees in their last World Series game?
• How many fingers am I holding up behind my back?
• Does dark matter exist?
• 8 sudoku puzzles, taking a minute’s break between them

Back to Richard Rumelt. His list of essential characteristics for a strategy are:

• a diagnosis of the situation—you need to actually have insight into the situation
• a general guiding policy, informed by the diagnosis
• coherent actions that align with the guiding policy.

I don’t think this is a bad list, and a good strategy will have those. But it doesn’t explicitly tell us why a good strategy will have those, or why 7+7 isn’t a strategic problem.

Here is my definition:

A “strategic problem” is a problem that deals with a system of connected elements—elements that affect each other.

That they affect each other makes the system complex (“complex” originally meant “braided together”).

That the system is complex makes it hard to reason about! When trying to figure out what to do, you can’t just take one element (one place, one time, one resource) into account. You have to consider all of them, because affecting one affects the others. And so, at the micro level, your decision-making is contextual. Which is hard! Figuring out how that system works, with all the interactions between the pieces, is hard!

However: once you do figure that out, the connected nature of the system means that you often need to do surprisingly little to affect parts of the system. It’s connected! Wired up! Which is why a good strategy, once formulated, is concise and simple to state. The connected nature makes it possible to make statements that apply to the whole system.

This has been a bit of a long trip, but if you think about Richard Rumelt’s definition, you’ll see that it applies to—and only to—connected, complex systems. Either the system is complex, or you will create a complex (as in: connected) system to solve the problem. Saving for retirement? A complex solution, because you are acting in the present to affect the future.

In fact, one meta strategy might be to notice that it’s hard to reason about complex systems, and to first make an attempt to de-complexify the system by changing how the parts affect each other, to make the system more understandable.

And older, less-developed version of this idea: https://www.reddit.com/r/FrostGiant/comments/l9xa0g/strategy_is_about_multiple_interacting_layers/

The phrase “Win-More” emerged from the competitive Magic community, and refers to cards or plays that only work if you’re in a winning position. It’s a perjorative, because what you really want is to turn losses into wins, and why would you want to waste a card that doesn’t help you do that?

In the abstract, fine. However, a lot of things labeled “Win-More” are only labeled as such because of functional fixedness.

Here’s an example of a card that was shared with me as an example of “Win-More”:

(In M:tG, you start with 20 life, and you generally don’t get benefits from having more, so having 30 or more life is uncommon).

The trouble is, this is only a “win-more” card in a specific context. If you’ve constructed your deck around tempo and lifegain, then it might a) be entirely plausible to have 30 or more life, and b) you still don’t have a path to victory, or even board control. In that scenario, Serra Ascendant takes a local advantage (not a global one, because then it would be win-more) and leverages it to put out a cheap 5/5. You could be losing (perhaps your opponent is constructing a combo that will scoff at your measly 30 life), and Serra Ascendant could be a threat that demands attention.

Here’s another one, the first answer given when someone asked for an example of a win-more card

This is declared a “win-more” card because

This card assumes you can deal consistent combat damage to your opponent. If you can your creatures get a little bigger.

Why is it a win more card? If you are already able to consistently deal combat damage to an opponent then you are better off running removal, burn, or another creature than running this card.

In some random defensive draft game where both players can and want to block everything, this is correct analysis. However, literally all you need is a creature with flying to invalidate it. If you have a 1/1 flying Artificer’s Assistant, and an opponent has a 2/4 Acridian, then you are currently losing, so anything that changes that state certainly can’t be declared “Win-More,” because you weren’t winning in the first place. And Curse of the Stalked Prey does change that.

Generally, the simpler and more direct the combat, the less desirable “Win-More,” conditional value mechanics become. However, most games have some sense of space and locality where elements are shielded or at least distant from immediate competitive pressure. You don’t need to win against the entire world all the time—you just need to win in some niche, however small, and find a way to convert or leverage that into globally applicable value.

Sometimes that value isn’t even direct, obvious value, it’s just negative value for opponents. If you have a happy little farm that can a) grow, and b) churn out suicide bombers—well, the farm is clearly a “Win-More” element (if we to hold it to the standards of competitive pressures), because it’s conditional and assuming that the farm isn’t invaded or whatever. Nor is the farm churning out obvious scalable, global value—you can’t do much with suicide bombers other than, uh, suicide bomb. But the farm could conceivably grow if it just a) kept growing and b) kept sending martyrs at enemies. It would never be a military superpower, just a happy little farm, with (coincidentally) a lot of dead enemies.

In one sense this is a “win-more” card: it can’t stand up to any kind of burn, it only does damage to the opponent if it isn’t blocked. But that misses the wider context, which is that it’s cheap, will kill the opponent if not handled, and (can be) painful to deal with. You don’t get to choose what penalty your opponent pays (dead creature, spent burn spell, lost life), but they have to pay something. If you can reliably a)force them to pay, and b) make the penalty greater than what you paid, then this is a path to victory. Literally built on the back of iterated “win-more!”

(This is basically the Zerg gameplan)

In fact, you can combine “Win-More” (the correct word is simply “conditional”) elements to provide each other with their necessary conditions:

Everyone here can only provide a place to sit because someone else is providing a place to sit. And yet it works.

Roles are bad in competitive play. They limit your thinking and cause you to misread the game situation. They exaggerate the importance of trivia, and gloss over what is actually important.

HotS is a game where you try and destroy the enemy core before the enemy team destroys yours. There are microgoals on the way that will help you achieve that—killing forts, defending forts, killing minions, getting mercenaries, and getting objectives.

None of that requires an archetypal medieval fantasy RPG party. One can imagine this being achieved by:

• a barbarian horde
• a crack squad of U.S. marines
• a band of wizards
• an army of ethereal ghosts
• a bunch of shifty, evasive elves
• a swarm of giant insects
• some rootin’ tootin’ cowboys
• a bunch of mindless zombies
• teleporting aliens
• Godzilla
• A bunch of self-healing rangers

There are a lot of different ways to fight. Tank-dps-healer covers only a tiny fraction of them, and many of those not very well.

“It’s cool that there are lots of ways to play, but why not just stick to what I know?” There are two problems with this:

• tank/dps/healer isn’t a flexible composition—it’s just a familiar one. It’s fine to stick with what you know if it works—but it has to work
• not to put too fine a point on it, but you probably don’t know tank/dps/healer. There’s a lot of depth required to play it well

In the end, the tank-dps-healer paradigm makes things more complicated than they need to be, limits our vision, and creates a lot of distractions on our learning path. So while we might have tanks, or dps, or healers in our comps, we don’t think in those terms. There are simpler, cleaner ways, that help us actually win games and have more fun.

When I began this I said that roles are bad in competitive play. This is true, but that doesn’t mean they’re all bad. Roles serve a social purpose in helping strangers queue up for a game without screaming at each other. However, on a competitive team, we can and do put away the training wheels.

This is a good question—but it is out of the scope of this post and would distract from the main point. There are simple, clean foundations we can build our understanding of HotS on. But to understand those well, we need to sweep away the clutter first.

(I’m writing this Phoenix Rising Amethyst, my team, but anyone should feel free to use it)

In competitive games, risk is bad. We want to eliminate random chance as much as we can.

The reasoning goes: we are trying to be good at the game.

• If we are better than our opponents, we want to get the benefits of that. We don’t want to lose to luck!
• But we don’t want to win by luck either, because it gives us bad feedback and encourages us to repeat risky strategies. But these are unsustainable by definition.

Being a team game, HotS has an additional reason risk is bad: it creates division in the team. The most common source of “team” errors is when one part of the team thinks a risk is worth it and commits, and the other part holds back.

But even without the team aspect—say you were shotcaller, and your team was perfectly obedient—risk avoidance is basic to competitive games. This is probably best explained by famed Starcraft player, commentator, and educator Day9, in his excellent article “The Marginal Advantage.” It’s a delight and you should read the whole thing (it’s not too long), but this is worth highlighting:

The marginal advantage embodies the notion that one cannot, and should not, try to “win big.” In a competitive setting, the strong player knows that his best opponents are unlikely to make many exploitable mistakes. As a result, the strong player knows that he must be content to play with just the slightest edge, an edge which is the equivalent to the marginal advantage. More importantly, a one-sided match ultimately carries as much weight as an epic struggle. After all, the match results only in a win or a loss; there are no “degrees” of winning. Therefore, at any given point in a game, the player must focus on making decisions that minimize his probability of losing the advantage, rather than on decisions that maximize his probability of gaining a greater advantage. In short, it is much more important to the expert player to not lose than it is to win big. Consequently, a regular winner plays to extend his lead in a very gradual, but very consistent manner.

The idea is that a small advantage we are certain of is better than a large advantage that might materialize. We can build on a small advantage; we can’t build on a dice roll.

Internalizing this single concept, and being guided by it, will do more for your game than anything else.

The way the community talks sometimes, you can simply a-move the enemy team if you’re up a talent tier or have a numbers advantage. Little attention is paid to how to fight—probably because it can be complicated. It is complicated, but we don’t have to tackle it all at first. We can start simply: at the team level.

Stand and Shoot: A comp with good ranged autoattack damage and good space control—say, with a Raynor, a Zul’jin, an Uther, and a Johanna. This comp wants to play defensively. It is more concerned with “keeping the damage alive” than “focusing” anyone down. This comp should mostly be standing still or kiting backwards in fights, moving forward only when the (ranged, auto) dps have nothing to shoot.

Pounce/Brawl: This comp wants to get in close and fight. Fights tend to be simple, but getting those fights tends to be hard: there is often one enemy ranged hero that you don’t manage to lock down, and that can make your “pounce” into a “suicide.” This comp likes flanks and does not like chokes. Closing the distance and escalating the fight ups the risk/unpredictability of the situation (and remember, risk is bad); compensate for this with more careful play to set up better fights.

Kite And Backstab: This comp wants to pull the enemy apart, then murder someone alone. Any comp with Lunara and Valeera is obviously this, but even a comp with Raynor and Tyrael might be best played this way. This comp can be a little hard to build intuition around, but believe it or not it is how most comps should play.

Stand And Shoot:

• The tank plays too far forward, taking unretaliated damage that otherwise wouldn’t have happened.
• DPS play too aggressively, taking damage
• Focusing on “focusing” someone down, rather than getting good overall trades, e.g. focus-firing someone, but clumping so you eat a lot of AoE
• Not being mindful of the possibility of getting flanked

Pounce/Brawl:

• Staying within enemy attack range without getting anything done. This comp should be in or out, not in-between.
• Attacking head-on, particularly when clumped, is usually a mistake. This comp likes flanks!
• Staying in a fight for too long, particularly if there’s a ranged hero you haven’t managed to pounce on
• Not being mindful of flanks, particularly on your healer.

Kite/Flank

• Overcommitting to a fight, particularly against multiple opponents, or when there are enemies nearby, or when you hero just isn’t that good at fighting.
• Clumping up. This comp wants to expand to bait enemies apart.
• Not macroing enough. The enemy isn’t (necessarily) stupid and needs a reason to split up—often part of that reason is macro.

Different heroes are geared toward fights of different sizes.

Illidan and Zeratul, for instance, are fantastic for small skirmishes. Their high mobility and damage let them dodge attacks while putting out their own.

However, they’re not quite as great in larger fights. They can certainly be useful. But they’re generally less favored.

Kael’thas, on the other hand, can be very effective in large fights, with strong AoE attacks. He pays a price in low health/mobility, but it’s a small price to pay for all the AoE.

In small fights, however, the AoE has much fewer targets, and he’s still just as slow and squishy.

In general:

• short range heroes are better in small fights
• AoE is better in large fights
• single-target damage is better in small fights
• healing is better in small fights
• mobility is better in small fights
• escape mobility is at least no worse in large fights than in small fights.
• ranged heroes are better in large fights

It’s trivial to come up with heroes that fall in two categories—Nova, for instance, is both single-target (good for small fights) and ranged (good for large fights). Figuring out those nuances is another topic. But this is a good start.

It sounds obvious: There are powerful AoE abilities in the game, so bunching up is risky. Hence, don’t bunch up. Everyone knows this.

Yet people do it all the time, and get punished for it. Why?

At root, it’s lack of imagination.

Here’s why people bunch up: something good happens in a fight. They move toward the enemy to capitalize.

Multiple people moving toward anything are going to be brought closer together:

Worse, they might have to go through a choke to get closer, making them bunch up much faster.

So the short version is: people bunch up because they’re trying to hurt the enemy team.

“But….we’re in a teamfight. What else is there to do?”

Lots of things! You felt safe to move forward—presumably you could move backwards, instead, and go get a camp, or push a lane, or walk around to set up a flank. You could stay right where you are and see how the enemy reacts to whatever just happened to them.

Lots of things, anyway, that aren’t moving right into an AoE.

Commit to not bunching up, and other options will start to suggest themselves naturally.

Sometimes, for whatever reason, it’s not a good idea to fight. Perhaps you have a defensive comp that can’t attack effectively. Perhaps your comp isn’t good at 5v5’s—maybe it has a lot of melee assassins. Whatever. If you take a fight, you’ll lose. What do you do?

Well, there’s a sort of rock-paper-scissors equilibrium:

• splitting beats grouping, because you can push multiple lanes at once
• ganking beats splitting, because you can have a numbers advantage
• safe macro play beats ganking

If you’re familiar with “Rush>Boom>Turtle” of RTS games, “Rushing” is equivalent to ganking, “Booming” is equivalent to split pushing, and “grouping” is equivalent to turtling.

So: when they’re grouped, you split. When they’re split (and you can see them on the minimap), you gank them. When you can’t see them, on the minimap, you play carefully.

It’s common to talk about the “map objective.” The temples on Sky Temple, the seeds on Garden of Terror, etc.

It’s all wrong. There’s only one objective: the core.

Of course, you can’t just right-click the core at the beginning of the match: that won’t work. Rather it’s useful to go after some intermediate, achievable goals, that will make it easier later to destroy the core.

That’s where map objectives come in, right?

Nope. I’m talking about forts and keeps. Killing those gives you catapults, and opens the map up for you. Taking objectives doesn’t give you anything lasting—unless it’s in the form of destroying enemy forts or keeps.

Why do we want to win the objective? Because it makes it easier to destroy enemy forts and keeps. If you can do that without the objective, so much the better!

For many, the most counterintuitive consequence of this is that there are many times you don’t want to go the objective when it’s up. This sounds like a paradox—it’s the objective, isn’t it? But that’s the error of calling it “the objective.” A better term is “map gimmick.” It changes the landscape of the map, but the objective is always the same: the core, and before that, forts and keeps.

There’s a common pattern I’ve noticed in games.

An example: the teams are both just poking weakly at each other. The Stitches on one team hooks the enemy healer, and his entire team dogpiles on him and kills him. Now it’s 4v5, so they run forward—

right into “The Kael’thas Zone.” They eat a ton of damage, lose one or two, and then the rest of the (low-health) team gets chased down and dies. The enemy team pushes down two forts and a keep.

All because the team succeeded (with the healer pick), and then didn’t know when to quit.

I think this is why it can be so hard to climb, even if you’re really good. You can spend the whole game creating advantages for your team—but often this only gives them more confidence to play badly.

If this post only gets one thing across: quit while you’re ahead.

A lot of winning in HotS comes down to simple waveclear efficiency. Even if you’re “losing” in one lane, you can completely make up for it in the other lanes.

You can look at how different characters perform differently re: waveclear very simply: every character has needs (input) and performance (output).

For instance, Probius has incredible waveclear performance, with his exploding Warp Rifts. But he is squishy and very vulnerable to being flanked, and so needs safety, either through protection (other heroes nearby) or position (staying near his own forts and towers).

Tanks and bruisers (“warriors” for short) are particularly interesting in this model. A lot of warriors actually have quite good waveclear, if you’re just looking at raw damage output—usually through short-range AoE of some sort. However, they have quite strict requirements to actually use it—chiefly that they have to get close to the wave! In an uncontested lane this isn’t a big deal, but in a contested lane, it means giving the enemy a chance to “poke” and use ranged abilities on them, without fear of retaliation.

Contrast this with “mages” like Kael’thas or Tassadar, or even just characters with a ranged AoE ability like Tychus. They also do AoE, which is good for waveclear, but they can do it at range—which means they don’t have to get close. This makes them stronger than short range heroes at waveclear in contested lanes.

This leads us to a simple rule of thumb:

To maximize effective waveclear, put your short-ranged heroes in uncontested lanes, and your long-ranged heroes in contested lanes.

We’ll stop there—most of the reasons not to do this are measures to deal with bad teammates. But we don’t have bad teammates.