This article was written out of necessity back in 2005… I was looking for a new gig and one studio asked for, as part of its application, a paper analyzing one of several possible games. Diablo 2 was on the list. Great timing! I had gone from X-Men Legends, where I learned the ups and downs of action RPG’s to working on Lord of the Rings Online, where discussions of scope and the feasibility of various online choices was the topic of the day. Both games set a lot of speculation stewing in my head about what could be done with the Diablo formula and why it was successful in the first place.

Analysis of Diablo II

by Patrick Lipo

Introduction

When the first Diablo was being previewed in 1995, most people (myself included) were blissfully unaware of its all-out potential. “I played that exact same game on mainframes 10 years ago,” we’d say, patting ourselves on the back. While we were congratulating ourselves, we had forgotten how those games had something that kept us playing and playing.

Blizzard could have simply created a polished copy of Rogue, Moria or Hack and done well, but they managed to refine the experience even further. Diablo was about giving gamers what the wanted, or perhaps what they needed, whether they knew it or not. Building off that success, Diablo II was able to add significant new features without spoiling what the original did right.

What Was Done Well

Simple World Presentation

2D may be “dead” to some, but the use of a 2D field was key to Diablo’s initial accessibility to millions. Everything the player needed to know about his surroundings was right in front of him. North was always up, just like a map. Yet, the isometric view and the 3D-rendered sprites kept the game from looking old. The Sims made a similar choice, and enjoyed similar ease-of-use.

Simple Controls

Click where you want to go. Click what you want to attack. What could be simpler? What Diablo I & II offers is a intuitive, rhythmic, and even mindless player experience at the lowest level. This allows the game to transcend the moment-to-moment battles and make people think about longer-term goals, such as completing the dungeon or gaining the next level. They could have added more moves to the player character (as Blade & Sword attempted), but would have clouded what worked so well, and pushed the emphasis to abilities and loot.

Frequent Rewards

From the very first Quill Rat slain, the coins spew forth, highlighting the strong cycle of rewards in Diablo II. While combat with a single opponent is simplistic, each enemy carries its own surprise contents. Who cares if a tiny Fetish unrealistically explodes like a piñata filled with gold, weapons and armor? Each and every kill feels different and rewarding because the player gets the pleasure of collecting new spoils, and rooting through a full inventory of randomly-generated items can be like a miniature Christmas morning.

The level progression curve is equally rewarding. While an MMO or pen-and-paper derived RPG such as Baldur’s Gate must space level advances with huge sessions of play, Diablo II manages to reward the player often, beginning at about five minutes and smoothly progressing towards around an hour. These frequent level-ups give the player yet another gift-unwrapping session of choosing which skills to acquire or advance. And while another game might provide finely-granular skill points to allocate, each Diablo II skill improvement is noticeable, with a beefy jump in damage, number of minions, or power duration.

Identifiable, Overlapping Goals

A major force in Diablo I and II’s long-lasting appeal is their presentation of goals. The player’s quest objectives are bold and easy to understand, such as “go here”, “find this” or “kill all of X in this area”. Beyond quests, the player can easily identify personal goals for his character, such as “level up”, “get this high-level spell”, or “become powerful enough to wield this weapon”. All these objectives are dangled in front of the player like carrots on a stick… You go into a highly-populated dungeon and you know what to do. You look at your skill tree and you see what prerequisites you need to summon an Iron Golem. Check your inventory and you see that sword that you just need three more points of strength to wield.

Coupled perfectly with this is the way that all of these goals overlap. In some games, the completion of a level gives the player an opportunity to catch their breath and consider quitting their session. In the Diablo series, the completion of a dungeon may bring you most of the way to earning another level, encouraging you to finish it off. However, once you earn that level, you might be halfway through another dungeon, drawing you to player just a bit longer to finish that up… And so it continues.

My first awareness of this dynamic came from playing the original Civilization, which had a similar loop of drawing the player from completing one more unit to finishing up that last attack before quitting for the night. Encouraging this sort of compulsive play behavior is not desirable in every type of game… Tetris’s strength comes from the ease of picking it up for a quick game, and massively-multiplayer games become more expensive to host if their players are active for 16 hours a day. However, for games such as Diablo II and Civilization, the goal structure had the effect of keeping people playing until the light of dawn began streaming through the window…

Randomness and Repeatability

The random generation of items and dungeons in Diablo II is something that outwardly sounds like a nice bullet-point for the sales flyer, but ultimately is integral to the series’ enduring presence. The dungeons have enough variation to make successive plays through (with the same advanced character or an entirely new class) different enough to keep the sense of discovery, but they are not so random as to make the dungeons appear “patchwork” (as seen in the PSP release of Untold Legends). The monsters have a sliding-scale difficulty that helps them remain challenging throughout your replay curve. The items have a fantastic, smart variability that provides statistics and powers that are interesting at the times you really want them. That last feature is something that Dungeon Siege had difficulty replicating (where you often saw Colossal Two-handed Mallets of Wisdom™ or Magic Wands of Excessive Strength™).

Integration of the Meta Experience

The effort that was put into making Diablo II replayable was exploited to the fullest in providing a metagame as well. Once the player completes the full story, it wraps almost seamlessly into the next play-through at a higher challenge level. The advancement curve is such that multiple completions are needed to fully experience everything a class has to offer (and even then there are other classes to explore). This embrace of the player’s experience above and beyond a single telling of the game narrative is something that more games should incorporate.

Minimizing Dead Playtime

One final element that helped give the Diablo series appeal was its conscientious reduction of dead time at any cost. Most RPG’s have some measure of uneventful busywork or travel, but elements such as the Town Portals virtually eliminate any dead travel time in the game. RPG purists doubtless were infuriated at this break with “reality” and “world sense”, but this addition had a far, far, far more positive effect on the player experience than a negative one. Diablo II added sprinting and item highlighting that identified and alleviated tedious bits that existed in the first game, showing that the developers considered this issue important to track down and solve.

What Could Be Improved

More Random Side-Quests

The randomized content of Diablo II is inspiring, as is the simplicity of their quests. One thing that I would do to maximize the value of such a powerful and versatile system is create far more simple side-quests than the game originally provided. The content structure and world layout of Diablo II makes a natural potential for creating hundreds of quests with variable properties that an industrious (and thorough) character can embark on. The component-based map structure allows the game to sprinkle quests into almost any map, each with a named monster and a rare or unique drop, so that adding new dungeons to a previously featureless play zone can provide an entirely new feel. There could be only a limited number of quests available for each play-through, so that it might take the player dozens of characters to see all of the possibilities.

These side-quests could also work with Diablo II’s replayability. By tracking the player’s completion history with different characters, the game could open up specialized quests on subsequent run-throughs. Complete the paladin on the hardest difficulty and your next character might get some holy artifact. This could bring more long-term goals than Easter eggs like the cow quest already provide.

Feedback for Hit-or-Miss

A difficult issue with real-time games that use to-hit rolls is what to do when the character misses his attack. Typically, a miss is shown as a normal hit with no effect or sound. Diablo is this way, allowing the player to click frantically at an enemy, but with only some percentage of the attacks resulting in damage, the rest passing through uneventfully. The player feedback on this is weak, resulting in a little bit of mystery around what is a “good” or “bad” attack total.

Having taken on this issue in the action-RPG X-Men Legends, there are a few things that can be done to help better represent it to the player. The first is to play a “dodge” or “parry” animation on the opponent that shows that player why damage was not done. This can be exciting, adding in new motion to the interaction, but it must be done carefully to avoid confusing the player (for example, big dodging motions might make the player think that the AI is doing something to keep away from him, as though he is doing something wrong). It can also put new pressures on the character animators, particularly if you wish to synchronize the animation with the incoming attack (although this is less necessary with Diablo’s smaller characters). Finally, in a game with many attacks coming into a single target, deciding which ones to respond to can become almost arbitrary.

In X-Men the above was impractical due to memory and manpower limitations, so an alternate approach was taken. A failed attack roll is deemed a “weak hit”, with almost no effect and an unsatisfying “thup” sound. For successful hits, an effect is selected from a set of increasingly intense impacts, depending on how much the player’s attack totals exceed the enemy’s defense. In addition to the hit sound, a secondary “rumble” sound is mixed in to give extra “oomph” to powerful hits. The result of this tactic is that when the player first meets a new creature that is fairly tough, he does weak hits, but as he begins to gain experience and outclass it, he is rewarded by much more powerful effects to go with his increased damage-dealing.

More Dungeon Interactivity

Diablo dungeons are very good at providing exactly what they need as far as functionality. Their interactivity needs are very simple… a key may unlock a door or trigger an animation, flipping a bit in the dungeon and little else. This makes “what you do” in the dungeons fairly limited. By adding a few moving elements such as sliding walls and mobile platforms, certain situations could gain more of a time element, such as protecting a caravan or moving through an area before a wall crushes the player.

Also, if destructible structures and walls were added, player spells could have much more tangible impact on the world, and monsters would be able to smash their way through obstructions for dramatic effect.

Encounter Generation

At this point I’ve given suggestions on content, presentation and technology, but played it fairly safe (any schlep can say “more quests!” or “break stuff!”), so I’ll add something that might have more impact on Diablo II’s gameplay. The standard play structure of Diablo II involves creatures that sit and wait for you to clear them out, after which an area is empty until reset. For the sake of contrast, this could be enhanced by creating encounters that come to the player instead. These could be used to occasionally liven up travel through a cleared-out area, or add tension to certain objectives with ambushes, retaliations or pursuits.

These encounters would need to be generated with the same care as the rest of Diablo II’s randomized content, fitting with the appropriate biome, challenge level and terrain features. They would also have to be provided at carefully timed moments, so as to not betray the feeling of accomplishment that a player feels when walking through an area he devastated. A useful technique would be to let the player in on exactly what is happening by announcing the attack with a battle cry or even a special title (“Raptor Vengeance!”) when one is triggered.

In the case of random quests, these encounters could greatly enhance the sorts of events that can occur. Finally, generated encounters could potentially provide the game with the feel of hand-crafted content without the manpower and testing challenges typically experienced by heavily scripted games.

Conclusion

I hope that this analysis of Diablo II was not so drawn out that I lost you two pages ago. The game at its core is so simple, yet it did so many things right. It is amazing that more games haven’t benefited from the lessons it brought to the industry.