Despite respectable review scores and reasonably good sales, Final Fantasy XIII is widely derided, for many good reasons and a few bad ones. Defenders of the game often point to the battle system as the game's saving grace. I found myself sharing the sentiment when I played through the game recently. That might be seen as damning by faint praise, given that so much of the game was attacked, but the combat system really is quite strong, and manages to incorporate principles that are familiar from hard-core RPGs in a dynamic and accessible way. While it probably won't satisfy players who cut their teeth on those older systems, FFXIII represents an interesting fusion of principles.
The three main features of FFXIII's battle system are its flexible roles, turn-continuous structure, and stagger gauge. The first two continue long-standing traditions in the series: the active time battle system that has been present in some form in many entries, and the job system that has shown up repeatedly since its first use in FFIII. These have received only minor tweaks, mostly to the charge-up/cool-down aspects of ATBS that sometimes made FFXII's battles into dull, stately waltzes. The stagger, by contrast, seems to come from the Shin Megami Tensei lineage, resembling the "Press Turn" system from the more recent incarnations of the Persona series.
In the classical appearances of the Press Turn system (Nocturne and Digital Devil Saga), exploiting an opponent's weakness gives the player an additional action. As it appears in Persona 3 and 4, this system allows the characters to perform an "all-out attack" for heavy damage if every enemy in a group has been hit with its weakness. Similarly, FFXIII's stagger system rewards players that effectively use magical damage and status effects by putting enemies into a state where all attacks do more damage, and enemy actions can often be completely prevented. In fact, some enemies are almost completely invulnerable until they are staggered.
Of course, Megaten games are practically synonymous with hard-core difficulty, which would not be acceptable for a mass-market game like a Final Fantasy. To accommodate the broader player base, FFXIII makes two key concessions. Most critically, and unlike Press Turn, the player characters cannot themselves be staggered—the player doesn't need to manage this additional level of risk. The game also obviates the need for the player to remember elemental weaknesses and specifically design parties to address them. Any attack can increase the gauge, even if it does no damage, and the AI remembers vulnerabilities.
That doesn't mean that party design isn't important. Unless the player engages in some ludicrous grinding, each of the six characters will only be useful in three of the six possible roles. While the Sentinel role is disposable, it's essential for the three-person party to have at least one person competent in all the other roles. In addition, party members cannot switch their roles independently, but must change en masse to a new set of roles, called a "Paradigm", from a deck of six options defined by the player.
This puts the player in a unique position. RPG battle systems can broadly be divided into two kinds: "Actor" systems in which the player directly affects the success of an attack, and "Director" systems in which the player tells the character to attack and success or failure is determined mathematically. FFXIII puts the player at yet another remove, (let's call it a "Producer") where he chooses the "cast", the "roles", and the "plot" (the target) and lets things play out as they will. Even the party leader under the player's control is usually most effective when set to auto-battle, where the AI chooses the attack.
Unfortunately, the Producer's tools are limited. There is no way to set up a new Paradigm on the fly, and consequently no way to swap in characters from the reserve during battle, as was possible in FFX and FFXII. This proves to be a significant irritation in event battles where the game determines your party for you, as it tends to make bad choices in both the contents and arrangement of the Paradigm deck, with the result that these battles must often be restarted in order to have a realistic chance of winning.
Additionally, the game doesn't remember the Paradigm deck you used with a particular party (except for those event battles), so changing the lineup involves the tedious task of building the deck again. This makes picking a party and sticking with it a more attractive option to the player than constantly experimenting. Hence, most people play Fang/Lightning/Hope from the first moment they can and never look back. While there are slight variations in skill that make some characters more useful in particular fights, the process of switching them in and back out again is more trouble than it's worth. And, the game can't let those differences become too significant, because absent an escape option the only way to rework a strategy is to either lose a battle or quit it.
This leaves the battle system situated somewhat poorly in the fiction. With half the party unable to join any given fight, and a system that makes it inconvenient to even swap them in from the menu, those three characters essentially become irrelevant in play (a problem that becomes even more acute late in the game when the differences smooth over). Without making a tremendous adjustment to the system, this could have at least been alleviated by making a four-item "party deck" available to the player, each with its own attached set of paradigms. This at least would have allowed the battles to feel like the whole party was actually involved, but it also would have allowed the designers to give characters significantly different sets of abilities, even within their roles, without inducing a string of "Game Over" screens.
The fiction also feels wobbly because if the party leader dies it's automatic Game Over. This strains credulity in a game that allows quick revives with a Phoenix Down; there's no reason not to just automatically deduct one from your inventory and let the battle keep going. More to the point, however, if a game includes this mechanic, then it probably should avoid insta-death attacks. This, for instance, is the record of my battles against the first form of the game's final boss, Orphan.
1) Staggered Orphan twice, halfway through the second one he insta-killed Fang, Game Over.
2) Got through first stagger and started healing when Orphan insta-killed Fang, Game Over.
3) Staggered Orphan twice but started them too damaged to go all-out, because he insta-killed Hope twice. After second stagger ended, he insta-killed Fang, Game Over.
4) I put otherwise worthless "Death"-resistant gear on Fang. Orphan insta-killed Lightning once and Hope twice. Just as he reached his second stagger, he insta-killed Fang, Game Over.
5) Fang was still geared up, but I didn't shift to medics soon enough, they focused on Esuna rather than healing and one of his big attacks got me. Game Over, my fault.
6) Finally won.
Aside from demonstrating the needless frustration caused by essentially unavoidable Game Overs, this chain of events brings up a few other niggling problems. The first is that FFXIII will happily ignore the stagger gauge if you're doing too well in a battle. Orphan, for instance, never saw his stagger period run out: in all of these battles he would consistently come out of stagger with about a third of the timer remaining and smack my party with an enormously powerful attack. I had similar experiences with various Behemoths—even when I managed to stagger them they would often change forms halfway through, healing themselves completely. And of course, there was the Vetala, which erects an almost perfect defense before the stagger gauge runs out. This speaks to an inconsistent view of what the "Stagger" status means. For some creatures, it means complete incapacitation, for others (including Orphan's final form) it means susceptibility to the debilitating "launch" attack, and for yet others, like Orphan's first form, it seems to be a minor inconvenience that can be voided at will.
The other flaw that struck me was that the medic AI could be pretty terrible in a fight with a hard-hitting enemy, especially one that inflicts negative statuses. In these situations the AI would favor clearing statuses over healing, and particularly over healing the one character who cannot be allowed to die (the party leader). The AI that produces the auto-battle also tended to make poor choices: in particular, it was in love with Fang's inefficient Blitz attack. Of course, for the party leader the player can override the AI, but in critical situations the time lost by doing so outweighs a single inefficient attack.
I also didn't much care for the battles as visual experiences. This speaks to some of the game's larger problems, which I'll probably come back to. The battles sometimes had a tendency to become visually jumbled, especially when multiple spell effects were firing at the same time. However, the larger problem is that this didn't really matter, because most of the screen was useless. It displays the field, but the arrangement of the characters and enemies on the field is usually not important, and is out of the player's control anyway. The health bar of the active target and its status effects appear there, but again represent only a tiny and relatively inconsequential amount of data. For most enemies the stagger gauge is more important than the life bar, anyway. Unlike FFXII, you can't see which characters enemy units are targeting, though this would be useful to know.
Finally, you can see the damage you're dealing, but that's wasting your time anyway, as these numbers rapidly become ridiculously large while whittling mere slivers off the even more humongous enemy life bars. Many monsters in the game have so many HP that the number really ought to be relayed in scientific notation, in contrast to the characters, who won't break 10,000 HP before the game ends. Yet, the characters typically deal out damage many times the value of their own life bars in a single blow, making them the world's most dominant offensive force at the same time they are its most delicate flowers. If the enemies are going to have these enormous life bars there's no reason not to give the players the same, at least to make it seem like they're operating in the same universe with the rest of the game.
The most serious problem with the battle system, however, is that the concessions made to the broader base make it less interesting to play. Without the need to manage a stagger gauge, the Sentinel role is essentially useless. Without the need to tailor specific strategies using the particular strengths of individual characters, the same party can roll through the entire game. Without the need to remember the weaknesses of particular enemies, the player's task largely devolves to swapping between the paradigm with three offensive roles and the paradigm with two offensive roles and a Medic, with perhaps an occasional Saboteur or Synergist paradigm for flavor. What the player does in his Producer position is far less interesting than it could be, because with the exception of enemies with more than 5.0×105 HP, the combat often just isn't that challenging.
That isn't to say that the battle system can't be satisfying. When Final Fantasy XIII chooses to challenge the player, for instance in the early parts of the Gran Pulse phase of the game, the combat system is exciting and demands some creativity. Had Square Enix given the player the additional flexibility to handle more demanding combat, the battle system could have been an unskippable masterpiece. This would not have saved the rest of the game, but those flaws can wait for another post.
Currently Sparky works as a scientist in Rhode Island, and works gaming in between experiments and literature reviews. As a writer, he hopes to develop a critical voice that contributes to a more sophisticated and interesting culture of discourse about games. He is still waiting for a console port of Betrayal at Krondor.