Chronologically, here are the RTS games I've either completed the campaign(s) of, played a bunch of multiplayer of, or at least sunk a reasonable amount of time into before dropping:
Dune 2 (1992)
I played this one well after it was released, completing the Ordos and Atreides campaigns, IIRC, while on a Dune kick after I had finished Emperor Battle for Dune and wanted more Dune. In retrospect, both this game (1992) and Emperor (2001) are lacking compared to RTS games released between those years, with only the former having the good excuse of being the first in the (mouse driven RTS) genre.
And, IMO, as the first, it laid a shaky and flawed foundation for all RTS games to come (as far as single player is concerned), the early popularity of which, looking back, completely rode on the novelty of said mouse control, hiding the fact that the problem of how to make an interesting game from it, given AI limitations, would never really be solved, despite some promising initial attempts.
Warcraft (1994)
One of the first PC games (along with Doom and Bomberman) I remember playing, and the very first RTS I played. I finished both Human and Orc campaigns back in the day, and upon revisit as an adult.
Unlike Dune 2 whose gameplay specifics I found completely forgettable, here I remember particulars in both the large and the small. In the small, for example, I remember managing match-ups between the human archer (more range, less damage) against the orc spearman (vice versa), or economizing spell caster MP between buffs, recon and summons, a genre first, AFAIK. In the large, I recall positioning carefully constructed defensive formations at key choke points (usually bridges). Larger still, I remember being very careful with the order in which I exploited mines, and also being somewhat careful not to expose my flanks through wood chopping. The latter mechanic also featured in the later Age of Empires games, where it found rare but indeed deliberate use in multiplayer, However, like many other potentially strategically or tactically interesting mechanics throughout the genre, it would sit firmly in the background during single player play, both in the case of Warcraft, and later games that featured it.
Compared to Dune 2, the player could now select multiple units at a time with the mouse, although they'd still have to press a keyboard button followed by a mouse click to get them to move.
I like the art and aesthetic of the first Warcraft the most out of all games in the series.
Command & Conquer (1995)
The second RTS I played, and one that made a huge impression on me due to it being (as interactive media), in part, an allegory for the break up of Yugoslavia. It really captured the zeitgeist of the 90s unipolar moment and in particular its relationship to the consequently lost and confused "third world". Later entries in the C&C series would have less geopolitical focus and move to sci-fi and parody, with the only two exceptions being the first Red Alert and Generals, the latter of which I have not actually played. Anyway, I completed both the GDI and Nod campaigns, both back in the day, and upon revisit as an adult.
The game introduced four, maybe five, important innovations.
First, the convenience of clicking just once to issue a move order, along with, IIRC, arbitrarily large unit selections.
Second, the choice of alternate levels during the campaign.
Third, missions that put the player in charge of a small composition of troops and vehicles with the only way to gain more by reaching scripted points on the map. Actually, Warcraft had these as well, IIRC, but they were much more frequent in C&C, and unlike Warcraft, in some of these the player would reach a base at some point, (albeit sometimes without a construction yard) gaining the ability to (at least) construct units mid-mission. This is, arguably, single player RTS at its finest i.e. when it's not RTS at all, but RTT and the focus is on using limited resources to assault hand placed enemy fortifications, with perhaps some choice in reinforcements and a brief base building skirmish to finish things off. The commando missions, which are Westwood's (surprisingly competent) attempt at a mouse driven clone of the Mega Drive Rambo 3 game, showed, through a level design extreme, the similarity between good RTS single player and action design. Full circle, given the lineage from Thunder Force to Dune 2 via Herzog Zwei.
Fourth, a varied air game consisting of callable air strikes, controllable flying units, and infantry transports.
Finally, tiberium. A toxic, infinite, and modal resource. Modal in the sense that although never depleted, it is harvested faster than it replenishes, so once the accumulated stock disappears, income slows to a trickle. The subtlety of these resource mechanics were lost to the single player, as would be the case with most resource related mechanics in all other RTS games which I've played since. The issue is the AI either cheats so its supply lines don't matter, or it doesn't cheat and its supply lines matter but its a cakewalk regardless as it's not cheating.
As for multiplayer, the asymmetric GDI vs Nod match-up was too imbalanced, tenuously fixed 25+ years later in OpenRA. Tiberian Dawn is my favorite OpenRA mod and, IMO, underrated.
Warcraft 2 (1995)
I played it when fresh, and also revisited it as an adult, finishing all campaigns, including those in Beyond the Dark Portal.
A good sequel, best in its refinements to the original Warcraft, from which it doesn't deviate much, and also innovations lifted from C&C (better mouse control, flying units and more fixed resource missions). Its own more ambitious innovation - the naval aspect - on the other hand, feels disjointed and the game might have been better without it.
Red Alert (1996)
Played it fresh, but never revisited. Finished both campaigns, but not the expansion's.
Like Warcraft 2, it added gimmicky, but less disjointed (in single player), naval play. Unlike Warcraft 2, Red Alert was more of a variation than a refinement on its predecessor. The most important variation on paper being that resource fields no longer automatically damaged harassers, but again, this didn't change much at all in single player, while multiplayer, which I did attempt over LAN with friends, remained broken for the asymmetric match-up.
Age of Empires (1997)
Played fresh and revisited a few years later in my late teens around the time Rome Total War was released, although I don't remember if I completed all the campaigns (I definitely completed Egyptian and Greek campaigns, and at least part of the Japanese campaign).
An RTS with all the features from the games above, now "standard" in 1997, but also adding more opportunity for efficient play through economic micromanagement and strategy. This is due to having 4 resources to gather, 1 of them infinite, the rest finite, and the ability to exchange one for the other at unfavorable rates via a "market" building, or with other players. Again, I only found this truly mattered when playing the sequel in multiplayer (I never played the first game multiplayer), and could otherwise be completely ignored in single, as after a certain point throughout the course of any given scenario, the player would be free to gather resources at their leisure. I suppose though, the market could act as a crutch for players who managed to consume all their finite resources, but didn't want to restart the scenario.
Another innovation was the wonder, ruin and artifact victory conditions. I don't know if these precede timed base defense missions, which definitely featured the following year in Starcraft, or if some other game did those earlier, but they are very similar - the difference being the former are player triggerable. Of course, if an opponent builds the wonder (or collects the artifacts etc.), before the human player does (rare for the AI), it would turn into a timed offense mission - an obvious way to add welcome challenge to RTS single player, but for some reason generally hated by PC gamers.
In this chronology, Age of Empires is the first game where the terrain was height mapped (cliffs in C&C and RA were just impassable tiles, and Warcraft 1 and 2 didn't even bother with an attempt at the illusion), but actually, it was released a month after Total Annihilation, the legendary engine of which simulated a fully 3D battle space, albeit with terrain also sensibly limited to a height map (so no tunnels or the like).
Age of Empires was also very symmetric, a step back towards Dune 2 and Warcraft 1 for 1997, with "civilizations" being only minor variants of each other. Perhaps a multiplayer centric concession in the game's design, though I wouldn't know how effective, as my experience with Age of Empires multiplayer would be reserved for the sequel.
Starcraft (1998)
Played it fresh, but didn't complete the campaign, instead spent a lot of time in online multiplayer. I revisited the game for single player as an adult, finishing the the Terran and Zerg campaigns, but stopping somewhere towards the end of the Protoss campaign.
Starcraft completely embraced Warcraft's tactical mechanics, including unit selection limits, while at the same time trivializing one (mostly unrealized) aspect of the resource game. Minerals (formerly lumber) now exist as structures of size proportional to their wealth, usually arranged in arcs around other affordances for a base, rather than as forests of player choppable tiles of trees. A minor loss of a mechanic that never lived up to its potential, more than made up for by trading the two tier, disjointed, water and land terrain system of Warcraft 2, for a superior, two height, but entirely land based one. Also, in comparison to Warcraft's 2 warring sides, there are now 3 different species, all playing very differently to each other, with "equivalent" units varying wildly, instead of only subtly.
My own, controversial opinion here is that Warcraft "micro", made even more pedantic in Starcraft due to smaller tile sizes and increased pace, does not lead to the most enjoyable tactical play. Or rather, I should say I enjoy the tactical play in later RTS titles more. That said, I very much appreciate the high skill ceiling Starcraft has, and the local optimum in game design which Starcraft is.
Age of Empires 2 (1999)
Played it fresh, completing the tutorial and Joan of Arc campaigns, while also sporadically playing multiplayer (LAN) with friends. Never revisited, although I could and maybe should, especially for the internet multiplayer, which is very active at time of writing.
Now, by 1999, the fundamentals of RTS single player were fully established (mostly by the original C&C, really) as a number of (usually low challenge) mini-games embedded in what was essentially a multiplayer format. Due to the rising accessibility of online and LAN multiplayer (where an RTS game's mechanics could fully shine) the PC gaming culture's aversion to non social challenge altogether, and the game industry wide emphasis on "immersion" and story-telling starting around this time, players would never again see any real improvement to single player RTS that would make games in the genre more challenging, which is just another way of saying tactically or strategically interesting. Instead, as a sort of trickle down effect from multiplayer design concerns, or for their own sake, later titles would offer improvements to moment to moment tactical play, that could make for a more enjoyable single player experience as well.
One such example is Age of Empires 2's unit formations. Even though most players, myself included, barely ever make use of all the available options, the behavior of units in the default formation, whereby they march in a well formed square with ranged units behind infantry, alone is enough as a partial solution to Starcraft's "problem" (feature for some) of mandatory unit positioning pedantry.
Tiberian Sun (1999)
I bought this not long after it was released, but it sat in its box till a few years later. In fact, I played it after Red Alert 2, which shared the same excellent engine. I completed both Nod and GDI campaigns, and at least part of Firestorm.
The engine featured deformable height maps, projectiles, tunnels (and overpasses), and a host of other potentially interesting things, all relegated firmly to the background, possibly going completely unnoticed in single player. Nevertheless, as later RTS games would prove, enough cool things in the background can improve the general experience.
A more immediately palpable first (in this chronology) in Tiberian Sun was the addition of veterancy, a (mostly psychological, as far as single player is concerned) incentive for players to keep their units alive.
Homeworld (1999)
I remember being absolutely charmed by this game BitD, but due to life circumstances, I never finished the campaign, saving it for a future revisit. However, upon revisit, the LARPer in me was long dead, and even e.g. my music taste changed - I see Samuel Barber as a hack now. As a result, much of the game's charm over me had dissipated. That's not to say I found the game no longer aesthetically appealing, as it very much so is. The backgrounds, the ships designs, the visual effects, the voice work, and most of the music are all incredibly well done. It's just that over time I've put less value on this sort of thing, and more value on the game in the game, so to speak.
Anyway, even when you put close to zero value on immersion and story-telling, Homeworld still remains a very interesting RTS mechanically. Full of potential, but hilariously broken in single player, and possibly multiplayer as well. The player just needs to max scouts and put them in sphere formation to trivialize ~90% of what the campaign throws at them. Alternatively salvage everything with salvage corvettes.
As a result, I dropped the campaign second time through as well, despite wanting to find some redeeming perspective on the play. I guess I could always try playing with a laundry list of self imposed restrictions, or modding out the most obvious broken strategies. Anyway, it's a real shame the game's innovative and still rarely seen mechanics, such as 3D unit positioning and formations (but also things like simulated ballistics and fuel), find zero game built around them, just LARP.
Red Alert 2 (2000)
Played fresh, but never revisited. I must have completed both campaigns given the amount of time I sunk into this game, though I can't say for sure. I also remember playing a bunch of skirmish in lieu of human vs human multiplayer, as I only managed to organize LAN with friends a handful of times, again due to life circumstances.
Starcraft allowed some player constructed buildings (e.g. Terran bunkers) to be garrisoned as fire bases, as did Age of Empires 2, but this is the first game I played where you could garrison buildings which were just part of the environment.
Also, it kept the naval aspect, lost in Starcraft and Tiberian Sun, and did a good job of it, as controlling the seas meant invulnerable naval artillery ships blanketing large swaths of land too.
Classic C&C distinguished itself from Blizzard's *craft series namely by the simplicity of its units' actions - there was no unit selection specific menu of special abilities. Instead, some units were "deployable", something Red Alert 2 used to great effect even for literally the most basic troops (allied GIs).
Emperor Battle for Dune (2001)
Played fresh. The second RTS game with completely 3D graphics I played, after Homeworld. Completed all three campaigns.
Unlike Homeworld, despite the graphics being 3D, the game was spatially even simpler than Tiberian Sun, lacking any height mapping, opting instead for 2 height levels like Starcraft, IIRC.
Like the first C&C, the campaign proceeds by selecting territories on a map, but it's more dynamic i.e. procedurally generated, than the branching paths through handcrafted levels in Tiberian Dawn. An interesting concept, but as others have noted in other threads, feels a bit like the campaign consists of just playing skirmish games i.e. multiplayer against bots. Not entirely true, as I think three or four levels in every campaign were actually hand crafted, but the point stands.
Also, like Homeworld, the game was broken by a few units the AI had no answer for, and the levels weren't designed around, as they weren't really designed in the first place.
Empire Earth (2001)
Played fresh, finishing all the campaigns.
Highly forgettable, and broken, skin over AoE2, which I played more as a lulzy sandbox, back when such things were novel.
Warcraft 3 (2002)
Played fresh, finishing only the first, human campaign. I was interested in the multiplayer, but barely anyone was playing the base game at the time, rather various mods (of which DotA would become king, effectively killing the RTS genre).
Heavy emphasis on hero units and special abilities - an increasingly salient aspect of the original Warcraft design with each iteration in the *craft series. Also heavy emphasis on fighting neutral units guarding loot. Come to think of it, the game might have borrowed somewhat from Heroes of Might and Magic.
High level multiplayer matches, in this case, seem no different to how I remember the single player, which was all about getting a powerful hero to lead a small, but well balanced composition of units to micro its way through the enemy, even in maps where you could build a base and additional units - a nod to the superiority of tactics over strategy in single player RTS. A departure from other RTS games, even Starcraft, where a large part of the enjoyment and frenzied play comes from having to manage multiple fronts not visible from the same camera location.
Age of Mythology (2002)
I had this on my backlog since close to release, but only got around to it in the mid 2010s, when I completed the campaign. I imagine Age of Mythology is absolutely fantastic multiplayer, but by the time I was finally putting time into the game, there was no convenient and easy way to play a match online.
My favorite game in the Age of series. Replaces stone with favor, which is used to cast spells (basically artillery, buffs, or summons), and also adds a lot of very welcome variety between what are now called "cultures". My favorite are the Norse, whose ox carts (replacing the usual mining camps, lumber camps etc.) keep their economy highly mobile.
Interestingly, the developers chose to allow map designers to heavily constrain base building by only allowing players to build towns over predefined "settlement" spots, IMO a welcome bit of opportunity for design in an otherwise highly free form genre.
Rise of Nations (2003)
I short-listed this game after looking for classics that I might have have missed BitD, and was quite eager to spend time with it. However, I dropped it after spending about 10 hours in the Alexander campaign and also playing a few skirmish matches.
The problem I had was that the moment to moment play, fundamentally, felt like a casual city builder game for phones. It was all about clicking on things when a certain bar (resource number) filled up (reached some required cost), in a certain optimal order for the "strategy" you were pursuing. Very little of the play was actually positional, I found, and certainly not at the tactical level.
On the plus side, battles in the (interestingly, but ultimately pointlessly, dynamic) campaign were timed, meaning the player had to actually have a decent strategy. A welcome challenge, but not enough to make up for the rather boring and busy-worky way in which you'd execute the strategy.
I must say though, I feel as though these negative initial impressions could be overcome with more experience, or spending some time in multiplayer. I might yet return to Rise of Nations.
Dawn of War (2004)
Played for the first time last week, completing the space marine campaign. Haven't played multiplayer yet.
The second game appearing in this list where a single unit of control consists of multiple individuals, but the first to be really built around "squads" (Rise of Nations' infantry units also consist of multiple individuals, specifically 3). Firstly, squads are customizable weapons platforms, and how you equip your various squads forms is large part of your strategy. Secondly, only squads, not vehicles or heroes, can capture "points" on the map, which are also resources. Lastly, squads are self microing in combat. The player simply needs to be careful with where they position their squads in the large (well, medium really), and maybe retreat them from unfavorable melee engagements, but otherwise there's little pedantry to be had.
The campaign in Dawn of War is very basic - almost all missions are simply skirmish games where the AI is given a head start with some fortifications and a few pre-built bases, and proceeds no differently from a Warcraft 1 game where first the player must fortify against AI attack vectors, then throw resources at the enemy till inevitable victory. Not very appealing, but better than it sounds due to the strength of the core tactical play, and the squad customization aspect of the strategic play. Also, the AI attacks from many directions, so early on it can indeed feel quite frenzied mounting a defense to win some breathing room.
The above mentioned core tactical play consists of "deep strikes" (i.e. air drops), various infantry weapons, morale management, jump pack assault squads, artillery, melee tanks, anti-tank tanks and Blizzard like special abilities, made easier to manage through the attach feature (where you can integrate a hero or special unit into a squad).
Company of Heroes (2005)
My favorite RTS game. I got a copy of the first stand-alone expansion - Opposing Fronts - bundled together with a graphics card I purchased in 2008. So really I only played it by complete accident, at a point where I had given up on the genre. I went straight to multiplayer, only completing the Normandy campaign a few days ago on a revisit.
Why is it my favorite? The moment to moment play emphasizing broad tactical strokes via vehicle physics, arcs of fire, turret movement speeds, and squad autonomy. Micro consists of reversing vehicles with shift clicks, and satisfying use of timed grenades. It is very much so an improved Dawn of War, so much so that it really doesn't replace the former, as there's a bit of magic to DoW's comparative simplicity.
I particularly like the strategic play where, like in DoW, resources are tied to capturable points, but now it's all resource types, and they belong to territories, which must be connected to your base for the player to actually receive those resources. This opens up a whole dimension of play which successfully abstracts the operational concept of supply line ambush and security, especially so in multiplayer, but also serves to make AI harassment more potent in single player.
The campaign is slightly more varied than Dawn of War's, but ultimately amounts to the same sort of systematic fortress assault missions (unfortunately) backed by infinite resources and reinforcements, with the exception of a few maps where you're playing defense. Happily, a few maps offer "medals" for managing to complete objectives under a par time, or in a certain order, adding much welcome but ultimately still low, challenge. Unlike Dawn of War, missions employ heavy restrictions on available reinforcements, beyond the usual gradual opening up of the tech tree from early missions to later ones.
Supreme Commander (2007)
I tried to get into Supreme Commander a year or two ago, but didn't find it interesting. Invariably, at some point during a match I'd be playing completely zoomed out, directing resources to waypointed production of units, into what were essentially Dragon Ball Z like beam battles made out of metal. A more apt analogy might be a plate spinning game. You'd spin one plate by directing unit production into a particular axis of attack, but then notice another plate was about to tip over, i.e. your opponent was about to overwhelm your forces on another axis, so you'd spin that plate too, and so on and so forth.
Anyway, playing Supreme Commander put me off playing Total Annihilation too, which I understand is very similar, although I would be happy to learn there are enough differences to warrant giving the latter a try anyway.
Sins of a Solar Empire (2008)
Played fresh. This game, along with World in Conflict, was hyped by one of the writers at Anandtech, a site I read for PC hardware news and reviews. The design (or, more cynically, the marketing) for both this game and Rise of Nations emphasized the 4X aspect and, I must say, from my experience with both these games, I've learned this is a major red flag. I spent about 10 hours with the game, waiting for it to "click", but it never did.
Just like RoN, the play here felt like a mobage city builder, but with Warcraft 3-like special ability centric combat thrown in. Not to my taste, and at this point I simply stopped playing newer RTS.
--------------------------------------------------------------------------------
ADDENDUM: Some games "on my radar":
Z (1996)
I played the demo BitD, but it played too differently from other RTS games for my pre-pubescent brain to handle. However, I shortlisted it after I discovered it's a proto-CoH, and indeed after vetting it for a few missions, that's exactly what it is. Unit construction is resources is position, a design equivalence which removes, in my limited experience, the systematic approach to all single player RTS levels with base building i.e. fortify against AI attack routes, then concentrate wave after wave of forces against their base in complete safety. The result is highly dynamic play, even against an AI, something to be commended.
KKnD (1997)
Considered by many as the best C&C clone, which is no bad thing, IMO. I played through a few missions together with a friend on their machine BitD, and did like it, but forsook it in lieu of Starcraft and Age of Empires at home.
Homeworld Cataclysm (2000)
Homeworld, to me, has infinite potential, but is unfortunately broken beyond repair, as explained above. Perhaps this stand-alone expansion fixes it? Certainly everyone who recommends it emphasizes that it improves Homeworld's gameplay. On the other hand, the only complaints I hear against it are about non-game, interactive media aspects I care little for. But even there, such opinion is hardly unanimous. I know
jebsmoker is quite a fan of Catalcysm.
