Part One: The Various Types of RPGs and Diablo 2

The Original Problem of RPGs

From an academic perspective, the best thing about RPG history is that it has a clear beginning. The first RPG was definitely one of the early forms of Dungeons & Dragons (hereafter D&D except at the beginning of a sentence), which were developed from 1968 through 1970. The development of D&D explains why there are so many different kinds of RPGs today, and why the audience for one type might not like another. In the previous book in this series, Reverse Design: Final Fantasy 7, I laid out a fairly comprehensive history of the path from D&D to modern RPGs. If you’re looking for a longer explanation of RPG design history, I recommend that book and the many great books it cites, especially the work of Jon Peterson and Shannon Appelcline. For this book, however, we only need to look at two events which happened relatively early in the history of RPGs: the birth of the specialization style, and the birth of the action RPG. These two events occurred in reaction to D&D and created the majority of the design material which would eventually make up Diablo 2.

Dungeons & Dragons presents a huge problem for RPG designers: it does almost everything that an RPG can do. By the early 1980s, D&D already included hundreds (if not thousands) of important RPG design ideas. And these were not small ideas, either. Most of the structural systems which we would recognize in our modern RPGs had already seen a few iterations by the time the books comprising Advanced D&D (AD&D) came out. Ideas like random enemy spawns, randomized loot, lighting levels in dungeons, ability cooldowns, difficulty settings, and item creation all existed in D&D long before any videogame designer attempted to use them.[1][2][3] The staggering completeness of D&D meant that designers had a hard time making an RPG that felt mechanically original. Many of the first RPGs to emulate D&D only changed the setting and aesthetics. There are many games which are more or less D&D in space, or in steampunk settings, etc. Some of those games are classics, but they didn’t really change the scope or methods of the RPG genre. One game took the opposite strategy, however, and focused only on a small subset of D&D’s mechanics. That game was Rogue, which spawned the roguelike genre.

Rogue abandons all the parts of an RPG which do not directly relate to dungeon crawling. There are no towns, there is essentially no plot, and there are no mini-games or dialogue trees. There is only dungeon crawling. The goal of Rogue is to get a player into a dungeon as fast as possible, and keep him or her there indefinitely—always crawling more dungeon. Rogue doesn’t have any features that were not, in some form, originally invented in D&D. And yet, Rogue seems very different from its source material, because it focuses so intently on doing one thing. Rogue can kill the player character more often than the average D&D campaign, because the player is back in the dungeon after a few button presses. Moreover, Rogue can force the player to learn all the tricks of its dungeons in a short period of time, because there aren’t that many tricks. The player doesn’t have to worry about party composition and job classes. Rogue is also able to have a greater degree of volatility in its random operations because the game is too short for any single game event to have far-reaching effects. The player can never really lose more than a few hours of progress. These tradeoffs are the secret to the success of Rogue, and all games that use the same design strategy. Rogue gets rid of certain RPG design ideas to embellish others. This design strategy is what I call specialization. A specialization-type RPG is any RPG which focuses on just a tiny part of the D&D source material. In spawning the roguelike genre, Rogue made a template for a certain kind of specialization which the Diablo designers would eventually use.

There are two design ideas which are embellished in historically important ways in Rogue. The first idea is procedural content generation, and the second idea is permadeath. Procedural content generation means that some parts of a game’s content, like level layouts, monster stats, and magical item properties are created by an algorithm rather than by the direct work of a designer. Rogue relies very heavily on procedural content generation. Procedural content allows for an effectively infinite variety of dungeons to explore, which is necessary for Rogue since there’s nothing else to do in the game. Dungeons & Dragons actually pioneered procedural content generation, although it relied on dice and the human imagination to implement it. Room layouts, item drops, and enemy encounters in D&D are often augmented by (paper) procedures. Rogue simply goes further than D&D does by making nearly every aspect of the game design procedural. Permadeath, on the other hand, requires that with every death of the player character, the game starts over from scratch, with nothing preserved. This makes Rogue longer by forcing the player to play fresh dungeons every time he or she dies. For arcade games of the 1980s, this is a common enough strategy for extending the life of a game’s content. For RPGs, with their focus on long-term character progression, this is not especially common. Parties in D&D campaigns are, of course, wiped out all the time—but Rogue embellishes this idea by making it more common, and by eliminating most of the player’s defenses against it. Few dungeon masters kill their players’ parties multiple times an hour, but in Rogue, this would not be surprising at all.

As an important historical aside, I want to point out that while the key embellishments in Rogue were a product of an artistic reaction against D&D, the game was also shaped by technological concerns. No commercial computer in 1981 could have simulated all of the major systems in D&D. Rogue saves time by abandoning a large amount of the standard D&D systems. Rogue goes further, however, in that it also abandons content in favor of algorithms. The tabletop RPG lives on content. That is, a designer (such as your group’s dungeon master) creates and/or deploys a series of pre-constructed encounters which add up to a campaign. Rogue’s designers turned this content generation almost entirely over to a set of algorithms which construct dungeons, place monsters, and dispense items. This lack of content saves a lot of disc space, which, in 1981, was a precious resource. It also saves the design team a lot of time and money spent on making content. The next section looks at how, as technology and time constraints loosened, later roguelikes and derivative subgenres preserved the primacy of procedural content even when they didn’t have to. Eventually, a technological constraint became an artistic constraint, and the roguelike became an ingredient in other kinds of RPGs.

Roguelikeness as a Commodity

The oldest theme in the history of videogame design is that technological constraints become artistic strategies. Tomohiro Nishikado (whom we will discuss again later) turned a glitch into the foundation of all mainstream videogame design. In a similar fashion, the designers of Diablo 2 took procedural generation and permadeath (two ideas instituted partly because of technical constraints) as they appear in Rogue and roguelikes and turned them into artistic resources. One of the most important things to know about Diablo 2 is that it plays with its source material and implements it in an idiosyncratic way. In essence, the designers of Diablo 2 saw that roguelikeness was not a binary thing; games can be somewhat roguelike. The most obvious example of this is the fact that the Diablo games take place in real time, whereas most of its roguelike predecessors are turn-based. There are plenty of other examples, however. Player death in Diablo 2 is clearly based on its roguelike antecedents, but with some adaptations for a wider audience. When a player-character dies (on softcore setting), that character loses gold and experience points. Some of that can be retrieved if the player can reach his or her own corpse, but some of it is irretrievable. Thus, death is penalized—sometimes quite heavily—but in a way that’s more appropriate for a mainstream audience. Then again, Diablo 2 also has a hardcore setting which reinstitutes permadeath, if the player wants the extra challenge. Procedural generation is also something the designers play with rather than employ entirely. Traditional roguelike dungeons are almost entirely algorithmic, whereas many locations in Diablo 2 actually have a great deal of fixed terrain and fixed enemy locations—far more than most people imagine.[4]

Some of the ways in which Diablo 2 deviates from the roguelike formula are actually very informative about the degree to which the designers wanted to recapture the old roguelike spirit. In early roguelikes, item identification is an important bottleneck that prevents players from being able to use just any item they find. Not only are many items unidentified when the player picks them up, but the item identification scroll itself is an unidentified item which needs to be used blindly in order to determine what it is. This kind of volatility and blind chance was one of the things that David Brevik said he liked about early roguelikes, but which didn’t fit the voice of the Diablo games.[5] Indeed, not only was that kind of blind volatility diminished, it was actually reversed in some of the game’s mechanics. In Diablo 2, Deckard Cain can mass-identify an entire inventory’s worth of items.

It would be easy to see this change as modern developers dumbing down their product to reach casual players. This is an assumption game critics have made for as long as there have been identifiable videogame genres. But, like many RPGs of its time, Diablo 2 is actually dropping one idea to embellish another. They wanted to widen their audience, yes, but they also wanted to speed up and expand the game.

By eliminating the identification bottleneck, Diablo 2 is able to center the game’s focus on the item acquisition loop. There was nothing accidental about this. Brevik wanted to recapture the experience of playing Angband and Moria, two roguelikes he enjoyed in college. He remembered that he and several friends would play in the computer lab together (even though these were single-player games), and they would shout to one another to come and look when a rare and powerful item would drop. “You basically never actually killed the balrog,” he said. “You would play for those really rare items instead.”[6] Max Schaefer said something similar, saying that his focus was in creating powerful items that allowed players to have differing experiences of the game.[7] Although permadeath and other roguelike design ideas were important to the designers, the primary experience that they wanted to recreate was the ecstatic discovery of a very rare item that caused a sudden swing in difficulty.

The Legacy of the Tabletop: Character Classes

Another way that Diablo 2 diverges from its roguelike ancestors is in the use of character classes. Although the character class was one of the foundational ideas of the RPG, the roguelikes from which Diablo 2 borrows many of its core ideas did not use character classes in a significant way. This lack was probably a product of their idiosyncratic reduction in scope. Diablo 2, despite having a comparatively massive budget, mostly sticks to the reduced scope which it inherited from the roguelike. Real-time gameplay doesn’t really take Diablo 2 away from the volatility, danger and endless (but narrow) variation of the roguelike; it just recasts those things in real time. Character classes, however, are a major deviation from the roguelikes which directly inspired the game. Every character class feels distinct, and their differences are reflected in the underlying systems of the game. Below is a graph which visualizes the primary statistical difference (HP) between classes in Diablo 2.

[8]

Although the most obvious trend is the divide between fragile casters and hearty melee fighters, the classes exist on a spectrum rather than in two undifferentiated lumps. There are quite a few differences between classes, although we're only going to examine the statistical parts in this section. In the next section of this book we will also examine the action-game differences between classes, like differences in mobility and crowd control skills. For now, we’re going to focus on simulated skill systems (RPG stats) and the few class-based items the game employs.

Historically, character classes in Western RPGs have been distinguished by the abilities they possess and the gear they equip. Diablo adheres to the former criterion, but not the latter. Each class features distinct abilities, especially in their passives. When it comes to gear, Diablo 2 experiments with a small amount of class-based gear, but most of it is usable by any class. In this section, we're going to primarily examine how character abilities distinguish each character class, and how that is a product of the traditional tabletop RPGs that influenced its designers. Most of this section will be about traditional RPG abilities and the RPG systems which support them. Although the character classes of Diablo 2 resemble their tabletop antecedents in many ways, there's a big difference beneath the surface. In tabletop RPGs, some characters hardly deal any damage to enemies at all. Instead, they heal, protect, enhance, obscure, redirect, negotiate, and many more things besides. In Diablo 2, all characters are constructed to slay hordes of enemies. Barbarians and paladins are much better at tanking than sorceresses and necromancers, but they have to be able to slaughter the demonic hordes by themselves, or else Diablo loses its unique voice. This is one of the primary reasons why this book contains so little about multiplayer mechanics. As more players join a Diablo 2 game, the HP, experience values and item drop rates increase in a way that's very favorable to the player, encouraging large parties. But there are no statistical changes to enemies that would require the tank/DPS/heal balance. A team of eight sorceresses will perform just as well as a more balanced team, and that's intentional.

Character Systems Overview

Before we get to each class, I want to unpack some of the systems that those classes share. There are three different dynamics which differentiate character classes: base stats, elemental availability, and mobility skills. Base stats are the numbers which make all of Diablo 2’s simulated skills work. Base stats are things like strength, dexterity, energy and vitality. These stats give characters physical damage and carrying capacity, hit rating and evasion, mana and HP—and a few other benefits besides. Every time a character gains a level, the player can invest five points into one of these base stats, but depending on character class, the benefits will be different. The player also invests in the base stats of skills, putting points into offensive spells, mobility skills, and passive abilities to raise their effectiveness. These too differ greatly between classes—and sometimes even differ greatly between different builds of the same class. Finally, each class has a unique elemental profile. That is, each class has access to a few types of damage, but no class has access to all the types.

BASE STATS

Like all RPGs, Diablo 2 is built on a system of simulated skills. A simulated skill is any skill which a character learns, rather than the player who controls that character. In Mortal Kombat, for example, the player has to learn and perform a button combination in order to use a fireball attack. In Diablo 2, the sorceress can learn the fireball attack at level twelve if the player chooses it. After that, the player only has to point and click. Because Diablo 2 is an action game, there’s some skill in aiming the fireball and other skills like it, but making the fireball more powerful is mostly a matter of stats rather than skill. (There are a small number of skills which really benefit from action mechanics, but we’ll get to them in the next section.) Fortunately, Diablo 2 offers quite a few different ways to augment the basic stats of a character or character’s skill, which is why these stats merit their own section.

VITALITY AND ENERGY

The most basic form of simulated skill in Diablo 2 is the stat point. Most RPGs have stat points that a player can invest in a character, either by random chance or out of a budget. Diablo 2 gives the player points to invest out of a budget earned after a character gains a level. Most of its stats closely resemble those that were originally invented in D&D with a few minor differences. For example, the vitality stat in Diablo 2 closely resembles the constitution stat from D&D. Investing points into either stat will allow the character to gain more HP. The exact implementation is different—points invested in vitality directly affect HP in Diablo 2 whereas in D&D, there is also some randomness in the process. The fundamental idea is, however, very similar. Energy does for spell resources (the mana pool) what vitality does for HP. The Diablo 2 energy stat doesn’t have a great analogue in D&D for this reason. In D&D, the character has to rest for a certain amount of time before replenishing his skills. In Diablo 2, the player has to use a potion or a skill which regenerates points. Also, energy stat has no effect on the power of spells. This is relatively rare—even in games which allow direct investment of points into skills, there is usually a separate character stat which also contributes to the power of some spells. In Diablo 2, this is only true of skills with a physical component. Spells in Diablo 2 only gain power from skill-specific points.

STRENGTH AND DEXTERITY

Strength and dexterity in Diablo 2 are also close to their D&D/roguelike ancestors, with one small change that has some huge consequences. Strength allows characters to wield heavier armor and heavier weapons, and to deal more damage with melee attacks. All of that is totally orthodox RPG design. Dexterity plays a huge role in hit rate for physical attacks—for both melee and ranged attacks. This is also orthodox to the tabletop games that existed at the time of Diablo 2’s release. The only real deviation from contemporary tabletop stats is that dexterity also governs ranged damage. We’ll cover this in more depth in the section on hit rates, but the basic idea is that dexterity is twice as useful for ranged attackers as strength is for melee fighters. Not only does dexterity help their damage, but it also makes their attacks more accurate at the same time. Because ranged attackers get to “double dip” in dexterity, many ranged attacks are balanced (one might say penalized) in a way that melee attacks are not. In theory, this is a perfectly acceptable tradeoff, but in execution there are some oversights. We’ll examine those abilities in the character sections, but it does seem like the genesis of this problem is the fact that ranged attackers get two bonuses.

The only really unorthodox dynamic in the implementation of stats in Diablo 2 is the way that stat points confer different amounts of benefit for different characters. In most RPGs, a single point in any given stat confers the same amount of general benefit, regardless of class. There may be class-based abilities which apply higher or lower multipliers to those stats, but basic rolls depending on a stat like wisdom or charisma will be the same for any two character classes who have the same value in that stat and no other modifiers (like feats). In Diablo 2, the returns on HP and mana investment differ significantly across the characters. These returns are intuitive, however; the Barbarian gets twice as much HP from vitality as the Sorceress, and the reverse is true for mana and energy. The other characters are somewhere in the middle.

DEXTERITY, LEVEL AND HIT RATES

In Diablo 2, magical attacks have a 100% hit rate if the spell animation comes in contact with the targeted monster’s sprite. Physical attacks only hit if the player has a sufficiently high level, a sufficiently high hit rating, and a little bit of luck. The single most important determining factor in hit rates is the attacking character’s hit rating, measured against the defending monster’s defense rating. Below is a graph showing how increases in hit rating increase the overall chance of hitting a monster of an equal level.

[9]

Investing five stat points (one level’s worth) gets a character 25 points of AR, which is the subject of the x-axis. Character level is the other factor under the player’s control that affects a character’s chances to hit with a physical attack. Below is a graph that shows how gaining levels affects hit chances if attack rating remains static.

[10]

The slope of the line in the second graph is shallower, although it can be hard to tell just by looking. Even after gaining 30 levels relative to the character’s target monster, the chances of hitting only go up by about 12 percentage points. Still, player characters will have a hard time hitting enemies that are a higher level than they are, and an easier time hitting enemies that are a lower level. Although it is never spelled out anywhere in the UI, one of the ways that character classes differ is in the way that level is applied to their hit ratings. Certain classes get bonuses to their “effective level” when making a physical attack. For example, a level 15 barbarian making a melee attack is treated (for the purposes of hit rating) as though he were actually level 35. That’s usually worth about nine percentage points in the chance-to-hit rating.

[11]

These bonuses mostly make sense. The three characters that tend to fight up close the most often, the barbarian, paladin and assassin, all get significant bonuses to their effective level. The casters, meanwhile, are penalized when attacking, since they really should not be doing so very often. It’s the middle two classes, amazon and druid, that have problems—but we’ll see about those in the respective class situations.

DEXTERITY, LEVEL, DEFENSE AND BLOCK RATE

Dexterity also contributes to defense and blocking, two simulated skills which serve the same ostensible purpose, but which operate independently of one another. Both defense and block rates are responsible for deflecting enemy physical attacks so that they do no damage. Neither has any effect on spells, which are instead governed by elemental resistances. Defense works exactly the same way for the players that it does for the enemies, except that players get to factor dexterity into the equation. The contribution of dexterity isn’t much, however.

Base Defense Rating = [Dexterity / 4] [12]

Even amazons and throw-barbarians (who put lots of extra points into dexterity) only get the same amount of defense from their dexterity stat that a paladin gets from a decent helm or boots. All that dexterity isn’t really helping the defense stat much. Moreover, defense itself is not that useful. Defense rating only works while the player character is walking or standing (not running), and it’S very difficult to get a high enough defense rating for it to help much after level 60. With a skill like the barbarian’s Iron Skin or paladin’s Defiance aura, defense can be augmented to the point that it offers real protection. Getting a significant amount of defense from items and dexterity alone requires some serious item farming—but we’ll cover that in the third section of this book.

The block ability is much more likely to provide significant protection to a player, although it has its drawbacks. Any character that equips a shield (or an assassin with the Weapon Block skill) can block incoming physical attacks. The primary contributor to the block rating is the shield’s own block stat, which is separate from that shield’s defense rating.

The other component in block rating over which the player has control is dexterity. Dexterity improves block rates a lot more than it improves defense, although the required investment is high. Block rate maxes out at 75%.

[13]

[14]

There are two graphs because there are two important dynamics to examine. First, it’s a lot easier to reach 75% block by investing in dexterity than it is to reach 75% evasion (armor) rating via defense via any means. But in the first graph, the player-character’s level is static. The second graph shows what happens to block rate as the player character gains levels. Block rates actually go down as the player gains levels unless the player is investing heavily in dexterity. For some dexterity-heavy characters this isn’t a huge problem. But for classes that aren’t stacking tons of dexterity, this dynamic forces players to find higher-level shields whose higher block rates can mitigate the lack of dexterity. (Really, all classes benefit from higher-level shields, but the need is greater for classes that don’t invest much in dexterity.) Indeed, the class-based differences in shield usage run deeper than dexterity investment. The classes themselves have block rate modifiers that operate similar to the hit rating modifiers. Any given shield has a 10 percentage-point spread for its block rate. For example, if the player equips their character with a Small Shield, the caster classes (sorceress, necromancer, druid) have a base block rate of 20%, the paladin has a 30% base block rate, and the assassin, amazon and barbarian will fall in between at 25%.

SKILLS

In Diablo 2, skills themselves have levels which the player can improve. Every time a character gains a level, they gain one skill point which can be invested into any unlocked skill. This is not unique in the history of RPGs; even contemporary games like World of Warcraft, The Witcher, Skyrim and the Deus Ex series use this mechanic. What’s unusual about Diablo 2 is the number of times a player can invest in a single skill. Every skill in the game can be leveled up 20 times, although many skills make for very poor investments.

The results of each point are fairly significant as well. Below I have visualized the growth in damage of some of the major skills from a few classes.

[15]

[16]

The growth is mostly linear, although spells have a little bit of increased slope at the end. The most natural response to this graph is to wonder if the linear growth of monster HP will simply neutralize any gains the player character makes in skill power. The answer is complicated. For the first difficulty setting, and a large part of the second difficulty setting, focused investment in the best skills will give the player increased killing power. In the second half of the game, deficits start to eat away at the player’s marginal killing power. At that point, the player has to synthesize gear, skills and good strategy, but that is the domain of part three of this book.

ELEMENTAL AVAILABILITY

The last major, systemic difference between the character classes is the types of elemental abilities available to each. Including physical-type damage, there are six elemental attributes an attack can have in Diablo 2. Each class has access to a few elements; none has access to all of them.

(Note that this chart tracks significant sources of elemental damage. For example, the assassin's Shadow Arts include some magic damage skills, but those are crowd control skills not meant for killing. I also did not count skills and builds which are widely deprecated, like assassin and amazon poison damage builds.)

You can see how the classes exhibit a great deal of variety in terms of elemental availability, and how there's not a lot of overlap. Six elements are more than enough to make seven classes (and more than a dozen possible builds) distinct in their elemental strengths and weaknesses. Yes, certain classes have a greater ability to split their skill points up between multiple trees than others. The “meteorb” sorceress, Tornado/Hurricane druids, Berserk barbarians, Traps assassins—all of these can mix two elements. Most builds cannot mix three elements without having a very specific gear set, which maintains the game’s balance.

In Diablo 2, the elemental system governs only which creatures or characters take greater or lesser damage from enemy attacks. The balance is mostly orthodox to the RPG tradition, although immunities are occasionally applied in counterintuitive ways. In most RPGs, a creature which is vulnerable to one element is strong against its opposite—icy creatures are vulnerable to fire, and strong against cold, for example. Sometimes that is true in Diablo 2 (certain families of enemy, like the Fallen, share fire resistance), but often elemental strengths and weaknesses don't make any intuitive sense. Thanks to the procedural addition of monster affixes, an enemy can spawn completely immune to fire and ice damage alike. But while any unique enemy can spawn with almost any elemental modifier, most of the elemental resistances in the game are fixed. Below are graphs which visualize the prevalence of resistances on the highest difficulty setting—the point at which elemental resistances become common.

[17]

There are a couple of important trends visible here. First, immunities to non-elemental magic and physical damage are much less common than resistances to fire, cold and lightning. This dynamic is in proportion with the prevalence of extremely powerful fire, cold and lightning attacks in the game, however. These three elements make up 49% of all types of spells or attacks the player can use.

But about resistances that don't reach the level of immunity, and yet are still widespread? The data are clear: physical resistance is extremely common on the highest difficulty setting. To a certain degree, this makes sense. Physical damage is the most common type of damage; only the sorceress class lacks a viable build which depends on physical damage. But there seems to be a huge oversight with regards to this dynamic. Physical damage is already subject to another form of mitigation. Skills which deal physical damage are almost always subject to attack rating. This is a disadvantage already; fire magic might be resisted by monsters, but those spells don’t miss unless the player aims poorly. Although there are builds which make physical damage viable in late-game Diablo 2, those builds are a lot harder to make and don’t offer much in the way of compensation for that fact. This is a flaw which we’ll see come to particular prominence in the bow-using amazon.

SYNERGIES

At high levels, most classes rely on a small number of skills to deal the bulk of their damage. But if those skills can only receive 20 skill points worth of investment, what does the player do with surplus skill points? The skill synergy allows one skill to enhance another, allowing players to invest skill points in a way that still enhances their primary skills.

When it comes to synergies, the big difference between classes comes down to the linearity, number degree of their synergies. A linear synergy is any synergy that takes place linearly across a skill tree. For example, the sorceresses Blizzard skill has synergies from the Frost Bolt, Cold Bolt and Glacial Spike; those three skills are prerequisites for learning Blizzard, so they are “in line” with the skill. When the player finally does invest in Blizzard, the sorceress already has a minimum of 15% bonus to its damage. The second important aspect of synergies is their number and degree. Blizzard benefits from three direct synergies, with an average amount of 5% damage per point invested. (I.e., Blizzard has three linear synergies, and their degree is 5% per point.) This means that a player investing all available points into Blizzard synergies might gain as much as a 300% bonus to the damage of the spell. Thus, certain classes have huge advantages in damage because of the linearity and degree of their synergies.

The impact of synergies on character classes is not always strictly quantitative. For example, several of the barbarian's combat abilities rely on synergies which are non-linear. Both Concentrate and Berserk receive significant amounts of magic damage from skills in the Warcries tree, which is completely independent of the Combat Abilities tree. The barbarian really does need those Warcries for buffing and crowd control purposes, however. So while the barbarian can't stack points into linear synergies the way that a sorceress can, his synergies encourage the player to invest in skills with great utility value. This is part of the barbarian's design philosophy, which requires the use of a movement and crowd control skill in conjunction with a main attack skill. Players tend to build their sorceresses around one big skill, and then focus on inflating the damage of that skill. For the barbarian, that strategy doesn’t work quite as well, and the synergies reflect it. Similarly, the necromancer's Curses skill tree lacks synergies entirely. Even the strongest curses won't kill most enemies, so the player needs to pair them with another offensive skill. Although the Poison Nova skill gets synergies from other poison skills, its best complement (what you might call a “soft synergy”) comes from the Lower Resist curse, which allows it to do a lot more damage than it normally would. Accordingly, all analysis of a class's synergies will try to take into account what those synergies (or lack of them) mean for the character.

Introduction - Back | Next - A Look at the Character Classes