What’s a “death ball”?
It results because of math. To defeat 100 units of equal strength requires *more* than 100 units (!!!).
Take for example 100 Zealots (Starcraft / Starcraft 2 melee unit). If you send 1-Zealot vs the 100 Zealots, it will die before it deals much damage.
In fact, even if you send 10 Zealots vs 100 Zealots, the 10-Zealots will almost certainly die before dealing any lasting damage. (All Zealots have a "shield" that regenerates. You probably won't even get past the shield).
As such, 100 Zealots can perpetually kill 10 Zealots over-and-over again. Your 100 Zealots can kill 1000+ Zealots (as long as they only attack 10-at-a-time).
----------------
Same thing with Axis-and-allies infantry / tanks, or many other strategy games. The "strength" of a ball of units is the *square* of their size.
That is to say: 100-Zealots is 100x stronger than 10-Zealots (!!), not 10x stronger as you may initially assume. As such, both players end up building a bigger-and-bigger death ball as a primary tactic (which is in fact, unfun and stale).
In Starcraft: Brood War, your 100-Zealots could only be grouped into groups of 12, meaning you need 18 actions to attack-move the group. In Starcraft 2: your 100-Zealots can all be in a single group, so you only need 2-actions (one selection + one A-move) to attack-move the group.
Lanchester's Square Law applies to Starcraft, AoE2, Axis and Allies, etc. etc. The assumptions therein are "ranged units all within fire of each other"
Lanchester's Linear Law applies to Risk, Hearts of Iron, and "bottleneck" situations in Starcraft/AoE2, etc. etc. The Linear law applies when units are no longer within reach of each other, and are instead seen as reinforcements to a limited sized "front" where combat takes place.
In Risk: only 3 units can ever be attacking, and only 2-units can ever be defending. This is the conditions of the linear law.
In Starcraft / AoE2, a "bottleneck" may force a limited number of melee units to fight, leading to a rare situation of the linear law in action.
Hearts of Iron "saturated frontline" concept also gets into the linear-law. Once the frontline is saturated, additional reinforcements do nothing to change the battle.
--------
Take those assumptions, craft a rather straightforward differential equation out of them, and solve. Bam, you've arrived at Lanchester's Laws (though Lanchester was the tactical genius who figured this out over a hundred years ago).
Modern board games almost always come back to Lanchester's laws. Lanchester wrote and published those differential equations so that generals could create wargames after all, to train their armies / commanders.
But it still means attack-moving 108 units (aka: 9 groups, all labeled 1, 2, 3, 4, 5, 6, 7, 8, 9) requires 18-actions.
That is: 1 (select group 1), then a, then left click on the attack-area. Then 2 (select group 2), a, left click on attack area. Etc. etc.
That's a difference of 18-apm (Starcraft Brood War) vs 2-apm (Starcraft 2). Furthermore, that "Group 1" in SC2 could have 150-units or 200-units (or 400-zerglings, lol) in it. They pretty much made death-balls have 10% of the APM requirement in SC2 compared to SC:Brood War.
I'm all for making the game easier and more accessible: but the death-ball tactic was a little bit "too obvious" and kind of unfun to watch, compared to the tightly coordinated groups of 12 that SC: Brood War was famous for. The groups were too big, numerical advantage was the biggest advantage in SC 2 IMO.
Nuke and Psi Storm exist in both games, but Plague, Irradiate, Ensnare, Stasis Fields are all AoE effects... as well as AoE damage sources (Tanks, Archons, Valkyries, Devourers, and kinda-sorta Mutalisks).
-------
Lets put it this way: if you really wanted to, you could just have 12-units per group in SC2. But no one would do that, the APM-advantage of having a big group (and manually splitting in the rare cases where a split is needed) is far superior than using 9 groups to split up 108 units.
And sure, stronger players probably split their death-balls into smaller groupings of maybe 30 units or so (you don't want to just A-move all over the place, a degree of flexibility is useful), and only occassionally merges everyone together into a deathball. (Fast and loose is good: you don't want to dedicate an entire army to a battle, especially if they might lose).
But no one is making groups of ~12 in SC2. Its too small.
In a made up but typical RTS, one swordsman beats an archer, ten archers are an even match for ten swordsmen, and fifty archers beat infinity swordsmen. In a 10v8 archer battle, the bigger side survives with about 3 archers, but in a 100v80 archer battle, the bigger side survives with over 50. This game has archer death balls.
So, a few hundred riflemen or dozens and dozens of tanks.
The death part of it comes from the numerical advantage, if an enemy unit walks into range of many times its number it'll get blown to hell before it can inflict meaningful damage, even if it is a technologically superior or inherently stronger unit.
If your knights are +2 armor and +2 damage, you'll probably want a mono-culture of knights. Futhermore, a singular upgrade (aka: Cavalier research) immediately applies to all knights.
Your +2 armor / +2 damage Cavaliers are probably superior to your +0 / +0 archers from the Feudal age, even if those archers are theoretically useful against some units.
Strictly speaking: its not a monoculture: Scouts + Knights is the common combination for cavalry civs, because Knights lose vs Monks/Archers, while Scouts win vs Monks/Archers. Furthermore, Scouts share the +2 armor / +2 damage upgrade with Knights. But still, the upgrade system almost "forces" a monoculture (or close to a monoculture): its far more efficient to only research one or two units, rather than researching many, many different kinds of units.
I'm sure the other games have their own terminology for the evolution of this "tactic".
One of the more interesting applications of this term I've seen is in TW WH2 (Total War: Warhammer 2) where a faction like Vampire Counts gets a lot of utility out of death-balling even trash tier units like skeleton warriors - so that opponent infantry and cavalry gets mired in the mass of unit collisions to make it easier to get a clean spell targeting off. I think the term is pretty well established at this point as evidenced by the derivative usages that are now popping up.
That is to say: pathfinding was fast enough that it clearly wasn't a problem in C&C. But they clearly wanted to nerf infantry-based death balls. And boy oh boy, are infantry-death balls nerfed.
An occasional infantry unit still finds its uses in C&C. They're incredibly cheap after all, and tanks have a difficult job killing infantry with their tank-cannons. So as long as you have a wall of tanks in front of your infantry, their rockets can contribute heavily in combat.
Advance Wars has a bit of a rocket-troop deathball problem however. Rocket troops are so cheap and effective, that large masses of rockets overpower any possible opposition.
Its clear to me that C&C had designed the "instant kill squish" effect to mitigate this problem. Maybe in playtesting, they tried it out and realized that rocket troops are just too powerful. A singular unit that truly does everything at very, very cheap prices.
But also if you play with anywhere near balanced resources between the sides you will crush the AI. It's so dumb.
The crazy thing is that to get a good score (that coveted S rank) you have to abuse the absolute ineptitude of the AI to do a commando raid on its capital. The game designers balanced the game against abusing the AI stupidity. Strategy and tactics are mostly there to delay the AI and it's 50x production advantage while you rush an APC around the edge of the map.
Advance Wars was a single player game (more or less). Fighting an overpowered foe and beating them with your brain was the whole appeal of a strategy game. Overcoming horrible odds was the point, both gameplay wise and narrative wise in pretty much every mission. It was super fun. The secret of strategy games is that they're much more fun this way. Focusing on multiplayer is what drained the RTS genre of its appeal.
The "mech. infantry spam" almost made you C-rank (or worse) each map. Mech. Infantry was too slow to beat any mission. But if you wanted a reliable win-path (and didn't care about the time it took to win), Mech. Infantry spam was the way to go.
Adv. Wars "balanced" it out in their own way: not by nerfing units, but by simply encouraging the S-rank screen, taunting the player to beat the map faster (which meant building less-efficient, but faster, higher-cost units like tanks)