The Devil Is in the Details of Action RPGs – Part One: The Logistics of Loot


Role-playing survival game is willing to take risks

While the title may suggest otherwise, I am not in the Diablo 3 beta. As I’ve been counting the minutes for either Diablo 3 or Torchlight 2 to be released, I ran through Torchlight 1. Playing it, I noticed several things that didn’t seem right with the mechanics that I wanted to take a closer look at.

When it comes to the action RPG genre, any fan knows about the cycle: you fight enemies to get loot to help you level up and repeat. In other words, the magic phrase is: Fight, Loot, and Level. If any of those three are not represented correctly, it can bring the experience down. We’re going to ignore “Fight” for this post, as everyone should know what is good or bad about it.

Loot is the big one, and is one of the main draws of any action RPG. With loot, there are two schools of design: set or random. Set loot, means that the designers hard coded every item, piece of equipment and weapon in the entire game. Demon’s Souls and Dark Souls are currently the best examples of this practice.

The advantage of set design is that by knowing every piece of gear in the game, it gives the designers freedom to get creative. In Dark Souls, each weapon type is unique in its feel and utility. This also allowed the designers to easily set up a general pace of getting equipment and balancing it out with enemy encounters.

There are two disadvantages to set loot design. First is that it has a cap, there is such a thing as the “best sword in the game” or “best piece of armor”. Meaning, that eventually the drive for better loot disappears, which is one third of the pull of playing action RPGs. Playing Demon’s Souls; I lose a lot of the motivation to continue playing new game +s as there is no new equipment to find.

The other has to do with PvP; set loot largely turns PvP into a race to get the best loot before anyone else. When I played PvP in Dark Souls, no matter how great I was at avoiding damage, all it took was one hit from someone’s high level weapon to kill me instantly. This forced me out of PvP until I could grab better weapons which would take awhile.

Randomized loot design which is used in most action RPGs, is that instead of defining set pieces of gear in the game. The designers set up algorithms for loot generations. If you look at Diablo 2, every item that has unique stats or bonuses comes with a prefix/suffix or prefixes, such as “burning” or “spiked”. These adjectives defined what kinds of bonuses are attached to the gear and from there the weapon is given the amount of that type. That means that my “freezing, burning axe” could be different from your “freezing burning axe”. Items are also graded in terms of rarity. This allowed the player to quickly see what equipment is more powerful and affects the bonuses from the adjectives. Diablo 2’s loot table is still one of the best of the genre with all the variables that go into generating loot.

Obviously the big advantage of randomized loot is replay ability. You never know if that chest or enemy will drop some super piece of gear. New gear provides both a visual boost (better gear = shiner avatar) and of course the stat boost. With Diablo 2, the harder the difficulty level, the chance of finding rarer gear is increased further encouraging play.

The problems with random loot and where Torchlight fits into this post, is that there is more to it than just creating random gear. In order for loot to motivate people, there must be an ascending trend of power over time. Meaning the further the player gets, the better the loot they find.

In Torchlight the loot table is not as refined as Diablo 2 was. For example while playing on hardcore mode; I used a chest armor I found within the first 5 floors of the game, as my only piece of chest armor for the entire game. While the idea of being able to find any equipment anywhere in the game sounds good on paper, it does cause two problems.

First is that it breaks the flow of the game. Enemies are designed around the generalized loot in the area. Meaning, if the best armor in the area can only block 3 points of damage, then enemies shouldn’t be set at dealing 30 damage per hit. If the loot table isn’t balanced with the enemies it can lead to the player either demolishing everything, or barely able to survive. Not properly balancing loot and enemies also makes it difficult to determine where to introduce new enemies or strengthen existing ones.

That leads to problem two, having the randomized element of the game work against the player. In Torchlight, my first character on very hard difficulty did not get lucky finding new pistols and armor to use. I went 5 floors using the same gun and armor. When I arrived in a new area, I could barely kill anything and enemies were nearly killing me with each hit.

The problem with Torchlight is that the loot table is not ascending as much as Diablo 2. If I find a rare item on floor 3 in Torchlight and another on floor 5, there is a good chance the former is as powerful or stronger then the later. However in Diablo 2, finding a rare sword at the beginning of an act and at the end, you are practically guaranteed that the latter is stronger than the former.

Looking deeper at Torchlight one of the problem areas I saw has to do with the types of rarity. Ignoring normal or white weapons Torchlight has the following categories: green for magical, blue for rare, gold for unique, and purple for set items (items that go together.) The problem with this is that with only a few categories, it makes it harder to find better gear.

If you get lucky and get gold equipment early on, chances are you won’t find anything to replace it for a long time (such as 4 or 5 floors or more). Likewise if you are stuck with a blue or green item, you’re going to find plenty of them which may or may not be better then what you have. Due to the rate of finding blue items which most unique monsters drop, it lowers the value of green items outside of the very beginning of the game.

Another issue with Torchlight is that there is more quantity then quality with loot, some unique enemies and chests drop multiple pieces of the same equipment type all within the same level range. This makes it a crap shoot when it comes to getting new gear. Sometimes you’ll find something that is miles above what you have, and other times you’ll find 2 or more pieces of equipment equal to or worse then what you have. As an example while fighting level 11 enemies, I saw loot as low as level 8 dropping. If the quality of loot increased at a faster rate, that would elevate some of the issues.

Going back to Diablo 2 it had the following categories (not counting normal or low quality): high quality, magical, rare, set and unique. That’s 5 to Torchlight’s 4, meaning there is a greater spread of items to find. In Torchlight my chance of getting a unique item to replace a rare is low. However in Diablo 2, I have a much greater chance of replacing my high quality item with something better. Combine that with the quality of loot rising at a fast pace, makes the hunt for loot an enjoyable one and not an act of necessity.

The challenge of using loot as a motivator is that the player shouldn’t be surviving from one piece to another, and at the same time, going hours using the same gear also doesn’t work. That does it for part one, in part two we’ll take a look at leveling and see if Diablo 2 still stands as the best in this area.

Josh Bycer.


  • A lot of rogue likes do really well with fewer categories. Regular, magic, “better magic”, artifacts, and random artifacts. Magic is, of course, the +4 sword. Better magic is the +4 sword of westernesse, which in addition to the +4 bonus gets other abilities (it has been awhile, but I think it's double damage to orcs, goblinses and such and also some nice passives like dark vision).

    Not every roguelike employs random artifacts and some never liked the mechanic. But they world well in various *bands. You just never know what you are going to get, even factoring in how item spawns are somewhat tied to depth/dungeon level. Further, since you have a certain goal in terms of character building, the value of a specific randart is even more subjective. That randart ring with +2 str, +2 end, and 3 other buffs looks great on paper but you had those buffs covered and you needed Nether and Drain resistance. Rogue Likes frequently find you using gear for ages, but I think they do a great job with the “thrill of discovery” aspect. I think it has less to do with the categories and more to do with there just being better balance in the actual loot generation *and* mechanics.

    Torchlight can throw a zillion mods at you on an item. It's not as easy to decide what you do and don't need, and frequently it's a MMO-like case of “well, this one has more, bigger numbers so, uh, I guess that's for me” (it looks Diablo-like on the surface but I think it's more MMORPG like). Classical Roguelikes have those sorts of numbers but *many* buffs are just flags. You get this resistance, or a level of that resistance (and there are only a few levels, or the numbers cap at a relatively low and easily digested amount).

    At the end of the day, I think you need to do the following to have the loot aspect work:

    1. Have a decent system in terms of mechanics/stats. It needs to be clear why this or that item is desierable in most cases. That won't always be true – it can take experimentation when chosing between a faster weapon or one with other buffs, e.g. Just piling on numbers is not that useful. While there are people praising the D3 stat overhaul, I'm not particularly fond of the JRPG/MMORPG/Diablo 2/TQ systems wherein the “ability scoreS” just get added to and added to ad nauseam.

    2. There is a delicate art when it comes to how you generate the random stuff and dole it out. There must be some possibility of finding something awesome “out of depth” (I think Ringil could spawn at floor 30/1500, if very rarely. You would be using Ringil for the next 70 levels, or most of them). But a key thing to remember with Rogue Likes – *that* experience exists inside a certain mechanics continuum.

    Because it's *memorable* when you find Ringil “early”/out of depth. But as RL players are fond of saying “the RNG giveth, and the RNG taketh. Mostly it taketh, though”. Ringil can spawn out of depth but really mean and nasty things can as well.

    So yeah, I want to be more careful about that in a “longer”/non-permadeth game. Anyway, this is easy to say needs to happen but I suspect it's tricky to do in practice. But not impossible by ay stretch – plenty of games *have* done it.

    3. I tend to agree that “the yellow that spawns at the beginning of an act will/should be better than the one that spawns at the end”. That's likely not 100% true but then you're in a comfortable mechanics situation. Oh, the beginning one has 15% lightning resist. We'll keep using it until we secure a better source, at which point there is a gear shuffle and re-prioritizing lets us ditch the beginning one for the end.

    That said, I don't consider it absolute but the best example I have of this is a RL that didn't follow traditional mechanics in this respect: I.V.A.N. A game needs to understand what it's going for to answer these questions.

  • Regarding point 2, that is a big deal and something that Diablo 2 got right where Torchlight didn't. In D2, there is a chance for rare items to drop off of any enemy, with of course a greater chance off of unique mobs.

    In Torchlight, my best gear only came from fighting unique enemies. While in D2 just now, I found a unique (highest rarity) shield off of a random enemy. The shield is probably 5 times better then anything I'm going to find for a while and that's ok.

    It's fine to have a chance at super all powerful items dropping at any time. However, the game should not be balanced on that chance.

  • Diablo 2 actually did (mini) bosses better than it did loot, IMO. While there wasn't, strictly speaking, “out of depth” stuff nightmare and hell really broke out exotic boss combos. But even in normal, you could run into sticky tactical situations (well, that was barely true in normal).

    I can't think of a single ARPG in this mold that did bosses in a satisfactory manner outside of Diablo 2 (I can't remember enough about Diablo to judge it). Bosses in Torchlight mostly feel like extremely high HP pinatas that sometimes hit like a battleship. where as in D2 you had to deal with things like LE/Spectral/etc. The MSLE boss in act 5 Nightmare is objectively tougher than the one in act 1 (barring a few really weird combos maybe, like rolling the lesser versions of those dude that cast meteor strike in a1). At least it has better stats. You may or may not be better equipped to deal with it, of course.

    But they were both tactical issues that required you full attention. Torchlight certainly required attention but it wasn't the same. The path to get the loot wasn't as satisfying, and neither was the loot.

    And I think that too many games took the wrong lesson from Diablo 2 by mostly focusing the better drops on the bosses. Though TQ did chests better than most games in the genre I think.

    Things are more interesting when random evil enemy #4566 can drop something awesome. I honestly can't remember if most of my gear in torchlight was from bosses or not. I haven't played it in ages. But I did remember an overall dissatisfaction with the loot/looting in the game.

  • Ah LE, the gift that keeps on giving nightmares.

    Dealing with LE or any of the enchanted enemies is what kept me from trying hardcore mode.