Review: Desert of Exile

Comfortably one of the most refined AoS experiences in the genre, Desert of Exile (Rising_Dusk, 2007-2009) is 6v6 map built on the premise that simple, elegant systems can produce complex and engaging gameplay when their parts are allowed to interact. It is the first AoS to adopt this philosophy seriously, and it is executed with confidence and discipline. There are no half-baked systems, nor any peripheral mechanics that don’t quite tie in with the rest.

The game features a fresh approach to hero design, clear and satisfying objectives, and a unique flavour of teamwork and synergy that distinguishes it from its peers.

The central lake is walkable, aside from two deeper patches in the middle.

Map Features:

Desert of Exile is true to its name: featuring a barren wasteland with nothing to see but the bases and some stray merchants selling items. The terrain is largely flat and unobstructed, save for carefully moulded cliffs which enclose each team’s base and side-bases.

In the absence of any distractions, the main “map objective” is collecting Faction, the familiar resource from Advent of the Zenith which was awarded for hero kills, and here is also awarded for participating in destroying enemy structures.


Taking out an exposed outer tower with a large push awards Faction.

Like it’s predecessors, Desert of Exile has four lanes, with a Keep spawning the troops on each lane. If a Keep is destroyed, troop production will permanently cease on that lane, allowing enemy troops to advance onto the primary structure. Additionally, losing a Keep in one of the side-bases will give the enemy team control of its adjacent healing fountain, and cause one of the two healing fountains in the main base to deactivate! If both side-bases have been lost, a team will have no fountain healing inside their main base.

Losing fountains is a very “suitable” reward for taking out side-bases, as it makes attacks on the main base more effective: particularly the loss of the front-most fountain which makes defending the mid entrance sustainable. I really like the natural progression that this mechanic carves out for players.

Keeps are also the place to buy extra units, which will be queued up and march with the next troop waves that spawn. There are eight available unit types, including cheap backup, specialised attackers, and heavy pushing tools. Below are the Hallowed Order units; the opposing faction has functionally identical ones with different names.


The (×2) denotes that one unit of that type will be queued for each of the next two waves.

The most expensive units cost Faction (which is considerably more scarce than gold), but that’s fair considering that used effectively, they can take out buildings and help pay for themselves. The cheaper units have fixed pricing brackets, which I like a lot because it lets a player’s choice of units be strategically motivated, rather than picking the best thing you can afford.


Long-ranged ‘Vagabonds’ whittling a tower from a distance.

Anyone on the team can buy extra units at any time, which allows reinforcements to be queued remotely for distant lanes. Keeps can spawn at most four extra units per wave, limiting how much lane pressure can be applied at once.


Conditions:

Integral to hero design and combat are conditions: a selection of 8 negative buffs loosely modelled on those in Guild Wars, which are shared and re-used by the entire cast. With almost no exceptions, these are the only afflictions that may be placed on enemy units:

  • Bleed: 10 physical damage per second
  • Burn: 20 magical damage per second
  • Blind: 90% miss chance on attacks
  • Cripple: 50% movement speed slow
  • Knockdown: Stunned
  • Maim: 75% attack speed slow
  • Mute: Silenced
  • Ruin: -10 armour

One immediate advantage of this system is that it makes efficient use of information. New players don’t have to adjust to different damage-over-time values on graphically distinct buffs from all the different heroes; there’s just burn and bleed. Similarly, there is exactly one slow at 50%. This efficiency carries into the hero ability descriptions, which can be concise because they don’t have to explain the specifics of each debuff, only its name and the duration for which it will be applied.

Individual heroes range between having all eight of the conditions at their disposal, to having none of them (typically it’s 2-3 each). But even heroes and abilities which don’t inflict conditions directly still have a lot of interaction with them: perhaps a bonus effect if the target is blind, increased healing against a bleeding target, or more damage for each burning unit nearby. The limited number of conditions in the game makes them relatively frequent in the hero pool, hence many heroes can synergise particularly well with each-other by providing the right conditions at the right time.


Two burn-lovers working in harmony.

Other abilities interact with conditions in a more general way, such as transferring them to chosen unit, extending their durations, cleansing them, rewarding the presence or absence of any conditions, or reacting to new conditions being applied. Desert of Exile’s custom buff system is the most advanced of any Warcraft III map, and it’s utilised to the fullest to create lots of interesting dynamics, all with effects that are (as an added bonus) easily described in plain English.

To give a sense of condition-based combat, I’ve listed some hero abilities below:

  • Abolish: Damages enemy units in the target area, and knocks down blind foes for 3 seconds.
  • Dragonbane Hide: Reduces all incoming damage by 10% for each condition this hero suffers from.
  • Upheaval: Channel to deal damage each second to foes in a target area, and cripple them for 4 seconds. Damage is doubled against knocked down units.
  • Martyr: Heals allies in the target area, and transfers all their conditions to the caster.
  • Epidemic: Spreads all conditions on the target unit to all nearby foes, with 150% duration.
  • The Prevailing Winds: Upon being inflicted with a condition, this hero deals 50 damage to nearby foes.
  • Twister: Sends out a slow-moving sandy cyclone which deals 20 damage per second, and applies 3 seconds of blind per second.
  • Absolution Aura: If a nearby ally is suffering from no conditions, they take 40% less damage from all sources.

There is a lot of diversity among the heroes. Some are proactive, trying to apply certain conditions to set themselves up for a good fight. Others are more reactive, moving conditions around or extending them. Meanwhile, there are heroes who are happy to be in the thick of things, tanking conditions and triggering bonus effects as they do.

Since the negative effects of conditions are fixed and shared by all heroes, durations and ability cooldowns end up being the key to balancing them. Durations are chosen to be whole numbers (2, 3, 4) for simplicity, and easy addition and counting. Some spells have particularly long durations, such as Scarlet Maiden‘s 45 second bleed, which emphasises the importance of having condition cleansing or transference available.


While condition-based combat has many perks, it has a few disadvantages as well. For example, while each individual ability might be simple and elegant, their high dependence on interaction with other abilities means that players need to know how all the abilities in a fight work before they’re playing the game properly.

Systemically high interactivity comes at the cost of a (mildly) steeper global learning curve. This works out as a very acceptable trade-off for Desert of Exile, but we must remind ourselves that there is a trade-off taking place: one which could be problematic for a more modern game. In particular, a steeper global learning curve might alienate spectators, or make life difficult for casters who are trying to explain what’s going on without regurgitating the entire game.

A related concern is readability of the game state, which is largely an issue due to Warcraft III limitations. As it stands, there is no way to see the remaining duration of a unit’s conditions, nor is there any visual representation of conditions moving from one unit to another. This makes combat more difficult to follow, particularly for players joining the fray mid-fight. This problem is solvable though, and its resolution might assist with the learning curve issue discussed above.


Trebuchets:

In earlier versions, each team had two Trebuchets isolated on a cliff behind their main base. These enormous wooden war machines were invulnerable, had global attack range, and their sole purpose was to fling rocks at the enemy non-stop! In practice, this meant that every four seconds, a pair of rocks would be fired at the nearest visible enemy units, dealing damage in modest area wherever they landed, possibly hurting enemy heroes in the process. Having some serious artillery on the back lines definitely feels cool, though the mechanic was a bit awkward in practice, since it exclusively hindered melee heroes, affected how lanes would push in a somewhat unintuitive way, and incoming projectiles were not indicated in advance.

Trebuchet rocks were later changed to land at a random point near their original target, opening the possibility of hitting ranged heroes. They were also updated to apply a random condition (cripple/ruin/maim) on-hit, making them potentially devastating to get caught out by. The thinking behind this might have been to “mix up” the conditions on any given lane and reward quick reactions, but in practice randomly getting hit, slowed, and dying to an enemy hero proved neither fun nor satisfying. Not long thereafter, Trebuchets were removed for good.


This is what global range looks like. Or, at least it used to.


Items:

Desert of Exile’s items can be divided into two categories; Regular items and Faction items. The former category consists of ‘cheap’ items which can be bought in each team’s base. These comprise:

  • Potions, which provide non-combat healing or mana regeneration over time, as well as an instant ‘cleanse conditions’ potion. Potions cannot be stacked, and are bought as single-dose or double-dose (the latter being more than twice as expensive, since it’s slot-efficient).
  • Permanent stat-padding items which each cost a modest 200 gold, and offer minor stat boosts such as +2 strength, +4 attack damage, +10% attack rate, etc. It is possible to buy several of the same item, but heroes are of course limited by having only 6 inventory slots.
    • One of the items grants +15 movement speed, but has the disappointingly fair downside of -2 armour. This item can be bought several times, but the negative armour will stack too, making the extra speed a risky choice.
  • While they’re not exactly items, it’s possible to buy a cheap flying courier to ferry items around, and workers which can repair buildings at the cost of gold. Workers are very important for maintaining bases, as tower life is relatively low, and towers will slowly burn down if left unattended on low life!

Faction items are so-named because they cost a whopping 8 Faction each! Bought from the shops situated on the battlefield, they are substantially more powerful than basic stat-padding items, though they do little stat-padding themselves. Their effects are situational, clever, and let heroes significantly change how they interact with the game’s other systems. Here are some examples:

  • Power Inverter: All damage this hero deals with spells is made physical if magical, and magical if physical.
  • Crown of the Ages: Increases durations of conditions the hero causes by 25%, and decreases durations of conditions the hero suffers by 33%.
  • Heart of Rendabi: Causes any magical damage you deal to ruin foes for 4 seconds.
  • Phylactery: Adds mana regeneration, and restores 15 mana to the hero each time it is inflicted with a condition.
  • Spectral Shard: Increase the damage of any conditions you inflict by 50%.
  • Devil’s Thorn: Returns 50% of the physical damage you receive to its source.

While it’s not directly emphasised by most hero abilities, the distinction between physical and magical damage becomes important once items are involved, as half of the Faction items react exclusively to physical damage, or to magical damage. Hence, whether a hero’s abilities naturally deal physical or magical has a major impact on how they can itemize. The other half of the Faction items affect or interact with conditions in some way.

In practice, it was rare for a hero to buy more than two Faction items, due to their expense and specialised applications. Players tended to pick the 1-2 items which worked well with their abilities, and filled out the rest of their slots with stat-padding items like +2 armour, which were favoured over having a Faction item with conditional or inconsistent benefits. This implementation was effective at rewarding players who earned Faction, but didn’t achieve quite as much build diversity as I expect was planned.

Faction items seem like they would have been better off as upgrades. Warcraft III’s limited UI would have made that impractical, but theoretically it solves the problem of Faction items competing for inventory space with regular items. I also would have liked if upgrades had multiple ranks: allowing players continue to specialise their hero without making them uniformly “more powerful”. Even if the higher ranked upgrades weren’t cost efficient, simply having them available relieves players of feeling “maxed out” before a match is close to ending.


Hero Design:

Like Dusk’s previous work, Desert of Exile involved collaboration with a number of artists to create unique, desert-themed characters which fit into the game’s distinctive setting. Lore was less of an emphasis compared to Advent of the Zenith, though it was far from neglected: with historical accounts available for each character in the hero-select screen. There are 30 heroes in the game, and the standard game mode is asymmetric, with 15 heroes in each team’s hero pool.


Reskinning the naga as sand-lizards was a personal favourite.

The heroes themselves have only four abilities: three regular and one ultimate. Each ability is intentionally kept simple: typically dealing damage of a certain type, and having some degree of condition interaction, just like the examples cited previously. Rather than being powerful and fight-deciding on their own, ultimates are often channelled or utility spells, and/or have some dependence on conditions.


That’s not to say ultimates can’t spill any blood…

By design, each hero is limited in what it can individually accomplish. Its own abilities probably have 1-2 workable synergies, but beyond that, teamwork is necessary to unlock each hero’s potential, and that depends greatly on which heroes are in play. Similarly, opposing heroes will have different ways to deal with conditions or apply their own, making for a complex web of possible interactions in any given combat scenario. Add all the possible item effects to the mix, and the potential depth that’s the crux of the game should become apparent.


Overview:

Desert of Exile is one of the few AoS maps that has a timeless appeal: it has a clear central idea, and explores it rigorously and in great depth. Not only this, but conditions are given a suitable stage on which to shine: clear and reinforced map objectives direct players’ attention directly to combat, and there’s no clutter like a jungle or shops with recipe items for players to get lost in. The combination of both its clever idea and disciplined execution is what made Desert of Exile such an excellent map, and a fine example for the rest of the genre to follow.

Download: Here
Further reading: Interview with Rising_Dusk

Advertisements

5 thoughts on “Review: Desert of Exile

  1. This map was brilliantly polished, and had a lot of great ideas. It was also completely devoid of many common mistakes made in other AoS maps, and not to mention relatively bug-free.

    But if I had to pick one tag for this map, it would be “anti-patterns”. Anti-patterns everywhere. If you think that this:

    Dragonbane Hide: Reduces all incoming damage by 10% for each condition this hero suffers from.

    …is interesting, fun, user-friendly, or otherwise good design, you need to think about it again.

    This isn’t my only grief. Condition interaction is over-used – it feels forced and boring.

    Simplified items and resources are only simplified in appearance, but infinitely more complicated and frustrating than having no items at all.

    Queued spawns make unpredictable, hesitant, reactive gameplay.

    And one other thing I remember pissing me off about this map: periodic effects use a curiously low frequency. Yes, I can see things occurring at low frequency, yes in spite of argument, it does affect gameplay. Yes, it makes an otherwise polished map look like garbage.

    This map had a lot going for it, and I reckon if it was still in development, many things would be different. But this game was certainly not good design.

  2. @cokemonkey: I feel you are being overly dismissive.

    If you think that (…Dragonbane Hide) is interesting, fun, user-friendly, or otherwise good design, you need to think about it again.

    It would be great if you could elaborate on this. It’s solidly designed, but you can’t really do many things in Wc3 engine, that would make this feel more user friendly (like highlight best targets, show damage reduction, etc.)

    And one other thing I remember pissing me off about this map: periodic effects use a curiously low frequency.

    What periodic effect? Khasmin’s AoE on condition? Condition duration? I’m not sure what you’re talking about.

  3. It would be great if you could elaborate on this. It’s solidly designed, but you can’t really do many things in Wc3 engine, that would make this feel more user friendly (like highlight best targets, show damage reduction, etc.)

    I don’t think simplifying the user interaction is enough to save this item – the design is inherently bad because it rewards the player for being effected by debuffs.

    What periodic effect? Khasmin’s AoE on condition? Condition duration? I’m not sure what you’re talking about.

    Periodic effect means effects that occur periodically – every spell in this map seems to move at a low frequency, and this is visually unappealing.

  4. I don’t think simplifying the user interaction is enough to save this item – the design is inherently bad because it rewards the player for being effected by debuffs.

    The design isn’t bad in itself. It’s the old and tried concept of playing from behind. Basically, the more you are damaged the more damage you deal or more evasion you have.

    every spell in this map seems to move at a low frequency, and this is visually unappealing.

    Still no idea, what this refers to. Spell animation seem on par with Wc3. Not sure what it means for a spell to move at low frequency?

  5. Dragonbane Hide is not a direct reward for you, it is a means to encourage a certain change in tactics from the enemy. You don’t seek to proc it as Dagurnott, the enemy seeks to avoid it, and you exploit their desire to avoid it. As an aside, the bonus was larger when it was first added, and we quickly found out in the testing games we played that day that Dagurnott could manage to gain health when the entire enemy team was focusing him. Good times.

    “Periodic effects using a low frequency” was an intentional design choice, if I recall correctly. He means abilities like (I no longer recall the ability names, to my disappointment) Khamsin Mistresses’ secondary, all of Maltheron’s area damage abilities, Arro Kree’s ultimate, and so forth, which mostly dealt chunks of damage on half-second or full-second ticks rather than “smooth” continuous damage. Exceptions include Skittel’s acid spit and Burn, which dealt smoothly continuous damage. Most condition-inflicting abilities with a duration also inflict conditions on one second ticks or longer (e.g. Atrius cloud extends blind by Xs every second, Genobee’s ult knocks down for Xs every Ys, etc.), mostly to keep Kassar from gibbing everything nearby or Last Rites users from becoming invincible.

    As for why this was a design choice – you might be interested to know that at least one recent game (Company of Heroes 2) in fact decreased the damage tick frequency of a DoT effect (Molotovs) while keeping the damage per second the same. The intent was to allow players to more intuitively gauge how much damage they could expect to take when making the decision to not move out of a DoT. DoE’s reasoning is largely the same, and I think it’s a more than fair tradeoff for a slight loss in visual appeal. I believe at one point Skittel’s acid spit used a slower tick, which made it more effective than intended, because in practice you could dodge sufficiently to only spend fractions of a second standing in the cloud, but slower ticks instantly inflicted a larger chunk of damage than you would have taken from smooth damage over time; thus the ability was changed to not use discrete ticks.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s