Ch.4: RTS Game Design: One Point of Armor Flips a Matchup

0:00 Welcome to Learning Podcasts. RTS Game Design: How Designers Balance Speed, Range, and Armor. Picture this. The player scouts the threat, builds the textbook counter, sends it in, wins the fight. And then loses the match 3 minutes later. Yeah. The ultimate soul-crushing RTS heartbreak. Honestly, we've all been that player. My first thought is always that I picked the wrong counter, never that the counter unit was the right answer at the wrong price. Right. 3 minutes later you're watching your worker count crash and you don't know why.

0:30 That gap, between winning the fight and winning the matchup, is where unit balance actually lives. So last time we walked through counter design. Hard counters, soft counters, role overlap, cost, timing, scouting. The relationships in the roster. Yeah, that's the shape the player sees. Anti-air beats air. Siege punishes static defense. Clean relationships, easy to picture. But none of those relationships happen by accident. Somebody sat with a spreadsheet and tuned the numbers until that relationship emerged.

1:02 The counter chart is downstream. The stat surface is what the designer actually moves. And every number on a unit's stat line is in conversation with every other number, on every other unit's stat line. Oh, that's a piece I was missing. So a roster isn't a list. It's a surface. So let's name the surface. The stats the player sees on a unit's card. Speed in tiles per second. Range in tiles. HP. Armor. Damage per shot. Attack period, which is how often the unit fires. Splash radius if it has splash.

1:32 Sight radius. Whoa. That's a lot of dials for one unit. It is. And here's where it gets really interesting, because every one of those dials is in conversation with every other one. On the same unit, and across the roster. In conversation with every other one. Yeah. We all look at damage per shot and call it good. Yeah. But damage per shot is one number. Damage per shot at range 5, against a unit that can close at 4 tiles per second, is a completely different number. And there's a second layer the player doesn't see during a fight.

2:05 The hidden stats. Cost in resources. Build time. Supply or population cost. Tech prerequisite. Wait, those are stats too? Yeah. They don't show up on the card the way HP and armor do, but they're load-bearing for balance. A unit's stat line tells you what the unit does in a fight. The hidden stats tell you what it cost to put it on the map. Hmm. So the designer is tuning 12 dials, not 5. I should be thinking about the whole purchase, not just the combat numbers, all at once. And this is where the counter relationships from last time start to make sense at the math level.

2:40 A unit that hard-counters its target almost always pays a, Premium. It costs more, or it builds slower, or it needs more tech investment before it shows up. Wait, really? The counter pays a price? It has to. If the counter cost the same as its target, built in the same time, and needed the same tech, every counter would just dominate. Every player would build it. The roster would collapse into a shopping list with 1 item. So the premium is what stops the counter from becoming the dominant strategy.

3:08 Yeah, and that's actually the question I had to wrestle with myself. Why does the counter exist at all instead of taking over the roster? And the answer is, the premium is high enough that the player has to commit to a matchup before they can have it. Okay, let me put that crisply. The counter premium is the friction the roster charges a counter unit for the matchup edge it gets. It can show up as cost, build time, tech, or terrain dependency, but the property is the friction itself. Yeah. And the contrast is sharp on both sides.

3:40 A counter that costs the same as its target solves nothing. The roster gets dominated. A counter that costs 10 times its target never arrives. The roster gets nothing. Right. And I get this pushback from competitive players all the time. They'll come to me and say, the premium I pay for siege units feels arbitrary. Three times the cost, four times the build time, and the matchup window is 12 seconds. Why isn't that just bad design? Because the matchup IS the window. If siege didn't pay that premium, every player would open with siege and every other unit in the roster would stop existing.

4:14 Picture a fire alarm. If it goes off for every spark, you stop trusting it. Yeah. You just take the batteries out. Right. And if it only goes off after the building's gone, it didn't help you either. The counter premium has to land in the same useful middle. Like the kitchen smoke detector that goes off every time you sear a steak. After the 3rd dinner you take the battery out, and then the house actually burns down. The premium has to land in the middle. So, range versus speed. The kiting tradeoff.

4:43 Yeah. Let me break it down. A ranged unit, say range 7, in front of a melee unit, say speed 4 tiles per second. On paper the ranged unit wins, because it gets free shots before the melee unit closes. But the moment the melee unit closes… …the range advantage is just a memory. Right. Exactly. So the actual question isn't "does the ranged unit have more range." It's "how many free shots does it get before contact." Speed difference plus terrain decides that number, not the range stat by itself. Oh, the stress of that micro. Once you see it as a free-shots problem, you start seeing why kiting is so high-leverage.

5:27 That stutter-step dance is terrifying when you're the one clicking away trying to keep your units alive. Wait, how does armor actually work? Like, mechanically. Does it cut damage by a percentage, or what? Okay. Step one. The unit has, say, 100 HP. Step two. The attacker deals 10 damage per shot. Step three. The defender has 2 armor. Each shot now lands for 8 damage instead of 10. So armor isn't a percentage. It's a flat subtraction off every shot. Right. Which means armor is brutally non-linear.

6:00 Against a 10-damage attacker, 2 armor turns 100 HP into the equivalent of, 12 And a half hits worth of survival instead of 10. 25 percent more effective HP from 2 armor. So armor isn't damage reduction. It's effective HP multiplication. That's exactly right. And it's why armor upgrades feel so different from HP upgrades. HP gives you a flat buffer. Armor multiplies the buffer against everything, and it multiplies harder against attackers that hit for small amounts. There's a twist on top of that.

6:35 Two units can have the exact same damage per second, on paper, and play completely differently. How come? Burst versus sustained. A unit that does 20 damage every 2 seconds and a unit that does 10 damage every 1 second both deal 10 DPS. But the burst unit lands its first hit as a 20-damage spike, and the sustained unit needs two full seconds to reach the same total. Give me a concrete case. Where does that actually show up in a match? Low-HP raid. A burst unit walks into a worker line of 15-HP workers.

7:09 The 20-damage spike one-shots the first worker; the sustained unit needs two shots and 2 seconds to do the same. So by the time sustained kills its first worker, burst has killed two and walked out. Same DPS on paper, completely different game outcome. Huh. So the same DPS number means almost nothing without the attack period. Right. The DPS number is just a summary for our convenience. The attack period is the actual schedule of the violence. Which brings us to one of the weirder things about stat balance.

7:40 Stat cliffs. A one-point change to a single stat can flip an entire matchup. Hold on. Isn't a stat cliff just bad balance? Why don't designers smooth them out? Why not a gentle curve instead of a sudden drop-off where plus one armor breaks the game? Yeah, you'd think so. But it's actually more interesting than that. Stat cliffs are baked into the math because the math is non-linear. Armor scales effective HP non-linearly. Attack period scales burst non-linearly. Speed scales free-shots non-linearly.

8:12 So the cliffs exist whether the designer wants them or not. Right. The job isn't to remove them. The job is to know where they are and put the cliff at a place that creates a meaningful timing window. Yeah, I've been on the wrong side of one of those. The unit composition that should've crushed the opponent's army, the armor upgrade landed 30 seconds earlier than I scouted, and the matchup was just gone. Wait. I want to stress test this. In a 100-supply deathball fight, none of that matters. The units that lose but extract value die in 2 seconds and the opponent rolls forward. The math has to be priced for skirmishes, not big engagements.

8:50 Yeah. We've all watched a single zergling kill three workers and known we just won that game. Mine was a Reaper opener. Two SCVs dead, the opponent pulled army back home, and the matchup was over before either of us hit 10 supply. Huh. So the unit dies and the player still profits. Exactly. The opponent had to react. They had to build a defender. They had to pull army back home. They had to delay an upgrade. The trade looks bad in the combat log and good in the resource graph. Right. It's priced for the disruption it can cause before it dies.

9:23 Yeah. So stat balance has to account for that. A unit isn't just priced for the fights it wins. The cheapest harassment units can be the most balance-sensitive in the whole roster. And there's one more hidden stat that's, I think, the most underrated balance lever in the genre. Build time. Build time. Yeah, makes sense. A unit can have perfect counter stats, every number aligned with the threat, and still be the wrong answer if its build time is too long. Because the threat shows up in a 30-second window and the counter takes 90 seconds to come out.

9:55 Right. So the counter is a lottery ticket. The math says it wins. The clock says it never arrives. And the player who scouted the threat correctly still loses, because the design didn't give them an answer that fit the timing window. That would be a nightmare. The right counter, on the wrong clock. Yeah. And supply works the same way. A unit that takes three population in a maxed army costs you three other units. That's a stat. It's just not on the card. Okay, so pull all of this back up to the chapter.

10:26 The roster isn't a flat list of units with stat cards. It's a graph. Every unit's speed, range, armor, HP, DPS, attack period, cost, build time, supply, and tech prerequisite connects to every other unit's stat line. Whoa. A graph. Yeah. Move one stat by one point and ripples propagate. A speed buff on unit A becomes a kiting problem for unit B, becomes a positioning shift for unit C, becomes a build-order viability question for the whole roster. And the designer is the only one looking at the full graph.

11:02 The player only sees the edges they're touching this match. Right. Which is why balance patches are so hard, and why we'll come back to that idea later in the series. You can't just buff one unit. Every edge connects three or four nodes at once. A speed buff on the melee unit is a kiting edge, a positioning edge, and a tech-timing edge, all changing at the same time. I love that. The roster is a graph; the patch is an edge move. It's a beautiful, chaotic, fragile ecosystem. Honestly, that's the lens I want to leave you with.

11:33 Counter design tells you the shape. Unit balance is the math that produces the shape. Read every unit's stat line as a node in a graph, not a card in a list. Yeah. Once you see the graph, you stop arguing about whether unit X is overpowered and start asking which edges it's pulling on. Right. And the math only matters because the player can see what's coming. Which is the next thing we look at. The fog of war. What happens when the enemy hides behind a wall of darkness, and information itself becomes the most vital resource in the game.

12:07 Thanks for listening to Learning Podcasts.