EPTGame design is a process in which the designers create the game, and then the players play it (duh). This means that designers have an idea of gameplay that they think would be fun, and then create situations that will generate that gameplay. The issue is where the job of the designer stops and the job of the player begins. A designer can create a system in which units receive very specific move and attack commands from the player, and all the designer has to do is program how those commands work, or the designer do a great deal of work creating a reasonably effective attack move command, and all the player has to do is enact that general command and watch as the game unfolds. The advantages of having the designer manage that particular aspect of gameplay is that the game can be designed with that issue held constant, making it easier to design the game as a whole. It can also take the more tedious aspects of gameplay and take them out of the hands of the player, allowing the player to focus on more interesting tasks.
But the disadvantage is one of the more often-forgot aspects of gaming suffers: skill. If your opponent has a general attack-move command and you have a general attack-move command, then your attacks will be roughly equal, and you'll have to find somewhere else to get ahead. But if you can micromanage your attack and force it to come out more in your favor, then you can work on commanding your units particularly well and shift the game in your advantage. Similarly, if your army control is lacking, then you'll find yourself behind more often, and you'll have something to work on.
Various games vary as to how much they put in the hands of the individual. Some games prefer a vast and complex system in which everything is in your hands; other games prefer a series of quicktime events, and the game functions more as a movie than as gameplay. Put everything in the hands of the designer, and you're just watching a CGI that some game developer made; put everything in the hands of the player, and you're just asking them to design their own game. Games also vary in what aspects they find important to put in the hands of the player; Starcraft puts a great deal of emphasis on managing your economy efficiently and effectively, while DotA actually has most of the armies generated by the computer, and the only part controlled by the player is the heroes.
Obviously there are a number of solutions as to the problem of where the developer manages things and where the player manages them, and different solutions are better for different kinds of games. A good principle is whatever aspect of your game is supposed to be the most fun should be where the player has to work to make things happen. The parts of the game that seem more boring should be the parts where the designer takes more responsibility. The idea here is that the designer wants to draw the player's attention to the best parts of the game, so reward the player for paying attention to and spending actions on the fun parts, but if a part of the game is more boring, then forcing them to spend actions on that part of the game is forcing the player to do the part of the game that isn't as fun.
This means that the appeal of your game is directly related to how you draw the line between developer and player. In an RTS, a great deal of the draw is in finding effective strategies, fleshing them out, working out the economy aspects associated with them, and executing them effectively. This means that putting the player in charge of designing strategies is effective. But sometimes it is tempting on the part of the game designer to build the strategies for the player, rather than letting them work it out themselves. An example of this is the marine-marauder-medivac; there is a very strong "forced synergy" between these units, such that the player is pretty much forced to build an MMM composition if they want to make use of these units. It doesn't take a genius to look at the terran unit pool and see that these units clearly cover each other's weaknesses nicely, such that very little strategizing is actually done on the part of the player.
Many of the recent HotS unit announcements reflect this same phenomenon. The Warhound, for instance, practically has a big red "Use Me Against Mech!" sticker on it. This is not necessarily bad; the warhound is, hopefully, going to make mech viable TvP, and I personally am quite excited for HotS. But most of the units being added have very clear "Use Me This Way" stickers on them. Protoss needs a harassment unit; use the Oracle! Anti-air? Use the tempest! Siege breaker? Use the swarm host! This takes strategy out of the hands of the player, and puts it in the hands of the designer, and naturally the point of designing the game is so the player can have fun.
Other HotS units are done in a more interesting fashion. The replicator's purpose is... what? Well, that's not exactly obvious. That's not to say the unit is worthless; replicating enemy units can indeed be a powerful ability. It just gives the player a great deal more freedom as to how the game should be played. Similarly, the shredder has an unclear place in gameplay. It's obviously a very powerful unit, but as for how it should be used, who knows? Maybe to secure counter-attack paths? Maybe for drop harass? I've used the thing in the HotS custom map someone designed, and the thing is extremely powerful, at least TvZ, but the investment isn't really worth it if you just use it for defense, and it's too hard to make it useful if you do a straight-up attack with it to the side of your army. Attack with it, and it'll die before it can plant itself and start emitting radiation. It has a sticker like "Use Me For Board Control", but board control is an elusive and confusing enough quantity that interesting strategizing and gameplay can be generated by figuring out the intricacies of such a unit.
A unit customization model does the same thing, but to a greater degree. Rather than developers telling their players how to play the game, and the players executing as best they can, developers lay the groundwork for interesting play, and players figure out how the play should work.
