25th Aug, 2018By Softmints11 minutes

Three years ago, Dota2 finally broke away from maintaining parity with DotA Allstars, and started to explore new horizons under the continued stewardship of its lead developer Icefrog. Moving to a new platform is an exciting thing, and being on the source engine has opened up "infinite" possibilities for what Dota2 could become.

In this article I want to explore how some of those possibilities have been realized through the last three years of content patches in Dota2, and discuss how DotA's roots may still play a role in its future.

The recurring theme here is that content patches have focused on taking underutilised mechanics and integrating them more fully with the game to increase the overall depth.

Lets get started by looking at the parity-breaking patch itself:

6.84 – Actives and Passives

Icefrog's commitment to maintain DotA Allstars for so many years after Dota2's release earned a lot of goodwill from the community. For some, it may have been a sensitive issue to finally part ways and lose parity. This patch needed to deliver a smooth transition.

6.84 introduced a number of items, including:

  • Lotus Orb: Grants a target a shield that reflects single-target spells.
  • Octarine Core: Cooldown reduction and spell lifesteal, in one item!
  • Silver Edge: Introduces the debuff Break, which disables hero passives.
  • Solar Crest: Introduces the debuff Accuracy, which counteracts evasion and is a less extreme form of True Strike.

Furthermore, on the heroes side:

  • Morphling becomes able to transform into an allied hero.

All of these new mechanics are goodies that users on the DotA forums had been requesting for years. More knowledgeable forum-goes would routinely dismiss these suggestions, knowing that they were infeasible or outright impossible due to technical limitations.

It's no accident that these mechanics arrived together in 6.84: they made a clear statement about new possibilities for the game and got players excited about what was to come. Seeing players discover interactions like Tiny's Toss target being bounced back at him by spell reflection was a lot of fun!

Activity, Passivity, Creativity

However, there is more going on in this patch than meets the eye.

To understand what's happening, we need to examine the two adjacent content patches: 6.82 and 6.86. Between them, nine heroes that previously had two passives, received a new active ability to replace one of their passives.

As we all know, DotA has its roots in Warcraft III, which was an RTS. Its heroes had four abilities (which DotA inherited), and as a general rule they never had two passives.

Reincarnation icon for Tauren Chieftain
The one exception: Tauren Chieftain's ultimate was also passive.

Early DotA versions broke that trend. Multiple passives became quite common, with heroes like Skeleton King and N'aix even having three passives. Why did this happen?

As someone who's done this kind of mapping, one reason is that it's simply easier to make passive abilities! Actives need more scripting and graphics and testing. Also, when you come up with a cool active ability it's natural to want to share it right away. Having to finish several more actives first tends to slow things down.

A second thought is that active abilities rarely benefited from items: Warcraft didn't have native support for mechanics like cooldown reduction or spell amplification. This meant players couldn't easily find overpowered builds that were based on their actives. Passives allowed for more creative expression and satisfying builds because they interact with attacks, health, and other stats that were appearing on items. Players had no problem with passives!

That said, there's no doubt that actives are more fun to use and create better gameplay moments. In my opinion, Warcraft's policy of "no more than one passive per hero" was the right balance. In the 6.82–6.84–6.86 patches, Dota2 took advantage of its new technology to move back towards that balance.

  • Firstly, Octarine Core was a step towards enabling active abilities to interact with items, ensuring the game wouldn't lose out on the build creativity and expression that passives were offering. Spell amplification and bonus cast range followed in 6.86, with the Intelligence stat granting spell amplification in 6.87.
  • Secondly, Break was introduced on both Silver Edge and as an Aghanim's upgrade to Shadow Demon's ultimate. This made passives interactable, at long last.
  • Finally, replacing passives with actives on several heroes has made the game more fun generally. It also ensures Break isn't strong against too many heroes, since less of them have 2 passives. I'm happy to see some heroes like Bristleback and Spectre keep 2 passives, as this keeps the game diverse.

The result is a much healthier environment where both actives and passives are fuller parts of the game.

6.86Attack and Cast Range

Icons for Dragon Lance and Aether LensTwo new toys catch our attention in this patch, once again on items:

  • Dragon Lance: Increases attack range.
  • Aether Lens: Increases cast range.

In the past, it was considered impractical to have an item increase attack range. Sniper has had an "increased attack range" passive since the Warcraft days, but its implementation used a unit upgrade; similar to "improve flying unit weapons" as might be seen in Starcraft/Warcraft.

The catch was: unit upgrades can't be downgraded, so if this trick was tried with an item, players would be able to pick up the item, drop it, and still enjoy the increased range! There are a couple of solutions to this, like making the item undroppable, but none are elegant and I didn't see this attempted in any Warcraft map that I can think of. Dota2 was breaking new ground here.

The two new items cost roughly 2000 gold each, and didn't have further upgrades. I really liked this decision, because it created a niche for medium-cost items offering a genuine utility (for carries in particular) that late-game items didn't.

Players now had a compromise to consider: they would lose gold when selling these items late-game item slots grew scarce. An underdog who still had their Dragon Lance instead of a Daedalus would have a tangible attack range advantage! That struck me as very interesting. (Sadly this compromise was not to last, as Dragon Lance got an upgrade item a few patches later.)

6.87 Stock Start Delay

Warcraft III had item shops that heroes could purchase from (that's where we get the secret shop). It was commonplace that shops couldn't sell certain items until a fixed time several minutes into the game, which included classics like Boots of Speed. More powerful items like Scroll of Protection took even longer to become available.

A human hero waits at a shop, where Boots of Speed among other items are still unavailable.
Can't an archmage get some better service around here?

The reasoning was sound: shops only carry a limited stock of items (so there is competition to buy them), and making players race for them immediately from minute 0 would have been degenerate.

Eul's DotA and its successor Thirst for Gamma made heavy use of this "stock start delay" mechanic on both base shops and secret shops: with some items taking up to 10 minutes to appear. It seems Dota2 has surprisingly picked this old mechanic back up! 6.87 saw two new items making use of it:

  • Icons for Infused Raindrop and Tome of ExperienceInfused Raindrop: which appears after 3 minutes.
  • Tome of Experience: which appears after 10 minutes, and restocks every 10 minutes.

I don't see any suggestion that we'll see more of this mechanic. It seems to be a balance necessity for Infused Raindrop, while Tome of Experience is an anomaly of an item. Yet, we do see the complexity of Dota2 deepen with their inclusion.

7.00 Attribute Bonus

One of the campaigns that came out with The Frozen Throne was a miniature RPG featuring a half-orc hero: the beastmaster Rexxar.

Warcraft heroes normally max out at level 10, but this campaign was styled as an RPG and needed a steadier level progression. To accommodate, Rexxar and his friends ended up having a max level of 18.

Rexxar stands upon a desert cliff, accompanied by his trusty bear Misha
The opening scene of the Founding of Durotar campaign.

In Warcraft, the rule is that you get a new skill point every level. To keep consistency with this, the design team needed to give Rexxar more abilities to put skill points into. His basic abilities were adjusted to have 4 levels each (up from 3), but this wasn't enough.

Introducing a fresh new ability wouldn't worked, because one purpose of the campaign is to teach players about some of the new heroes that would appear in RTS gameplay. It was important to keep parity between the RPG beastmaster and the ladder one, so they had to have the same abilities.

The yellow plus icon of Attribute BonusThe solution was Attribute Bonus: a simple stat-padding passive to give players something to click on, and keep the rules about skill points consistent. It was intended for a single-player RPG, and it's a fine solution. You can't lose the Founding of Durotar campaign anyway.

When The Frozen Throne came out, map developers gained the ability to change the hero level cap. Most games did so: raising the cap to better match the rhythm and progression of their gameplay. This tended to be paired with Attribute Bonus to pad out skill points for the extra levels. DotA Allstars was among the maps that did this.

If Attribute Bonus had never existed for the Founding of Durotar campaign, I think we would have seen a wider variety of solutions to "what to do with extra skill points". Maybe DotA would have taken a different direction. Some games, such as Dota Outland, Ninjas on Battle.net, and Advent of the Zenith certainly did. However, most games (across all genres) saw Attribute Bonus as an easy solution, used it, and didn't think about extra skill points any further.

The talent tree icon, with two Lina heroes doing different things
One of the promotional images for the 7.00 patch.

The result is that until 7.00 arrived at the end of 2016, DotA was using and balanced around the archaic solution of Attribute Bonus, which was never intended to do anything for balance or work in a multiplayer setting. Certainly, there were cases where early points in Attribute Bonus were an innovative move (Juggernaut used to do this sometimes), but overall the +20 to all stats was wasting clicks and making everything more homogenous.

The move to talents was monumental, and befitting the leap to a new version number. Over 800 new talents had to be written, and gradually balanced over the versions to come.

Talents were a positive move for the game's design. The flexibility they offer to balance not only individual heroes, but specifically being able to do so at later stages in the game where items would previously have been dominant is a big benefit. They force some rare irrevocable decisions on players, which is nice to see. 

Two of the key mechanics that were introduced to the game through talents were respawn time reduction (formerly appearing only on the item Bloodstone), and passive gold-per-minute talents. The former was eventually removed due to disrupting the flow of the game, but gold-per-minute talents have made a big difference in giving certain heroes (particularly supports) the option of economic independence.

Other features were added in this patch, such as Shrines, the Backpack, and Monkey King, but I think the story is really centred on talents.

7.07 Dispels

The emphasis in this patch was on dispels. These were a fundamental mechanic in Warcraft III, where caster units played an important role in armies due to their powerful buffs and debuffs. Examples of their spells include:

  • Slow: Decrease a unit's movement speed by 60% and attack rate by 25% for 60 seconds.
  • Spirit Link: Chains up to 4 units together, causing 50% of damage taken by any of them to be distributed among the others. (This makes it very difficult to focus key targets down.)
  • Unholy Frenzy: Increase a unit's attack rate by 75% for 45 seconds. The unit loses 4 health per second while this is active.
  • Faerie Fire: The target loses 4 armour, and shares its vision with you for 90 seconds.

The dispel magic icon from Dota2's storm brewlingThese powerful magics that could turn battles were counter-balanced by dispel abilities which every race has access to. They always affect an area: a single dispel can eliminate many buffs if the enemy is clumped together.

The result is that it's generally possible to keep some of your army buffed, but having too many buffs out at once will make the enemy's dispels more efficient, and frees up their casters' mana to do other things. Note that running out of dispels is terrible: look at the duration on those abilities! It was a good system, and worked well in the context of Warcraft.

Lane-pushing games have a different dynamic because there's no army to think about: everyone controls one hero. There are few targets worth dispelling, and scarce few abilities that apply buffs in the first place.

In most lane-pushing games, dispels have been rare and strong, while debuffs have short durations (under 10 seconds). I think this equilibrium was arrived at because short durations are less oppressive against heroes that don't have access to a dispel, and if something is too oppressive players will tend to complain.

The exceptions to this tend to be older abilities that were adapted directly from Warcraft's RTS gameplay: such as Queen of Pain's Shadow Strike or Doom's Doom. Meanwhile, undispellable stat reductions like those of Undying or Slark seem to be the modern solution for long-term debuffs, perhaps because they are relatively unnoticeable to players and don't feel as oppressive.

Anyway, back to the patch. The new content includes:

  • Nullifier: An item whose active dispels a target, and prevents them from using their items for a short time. Can be dispelled.
  • Spirit Vessel: An item whose active temporarily reduces healing received on an enemy, or increases it on an ally. Can be dispelled.
  • Meteor Hammer: An item whose active stuns and leaves a damage-over-time debuff on targets. The damage over time can be dispelled.
  • Aeon Disk: After being brought below 70% health, this item activates to dispel and briefly protect its wearer from further damage.
  • Necronomicon Level 3: Now has a single-target purge (dispel).

There's a strong trend here of making more buffs and debuffs available to dispel. We also see this in the new heroes:

  • Pangolier's Heartpiercer: This passive applies a debuff to enemies which activates after a 2 second delay. It can be dispelled.
  • Dark Willow's Cursed Crown: Places a debuff on an enemy which will stun the area around them after a 4 second delay.

The innovation we're seeing here is hero abilities with delayed debuffs to make dispels more attractive. If you dispel before they activate: the ability is fully countered, just like dodging a projectile! These two heroes were cleverly designed with this patch in mind.

We also see a number of hero talent and ability changes:

  • Morphling: New ultimate dispels self on cast.
  • Razor: New ability periodically dispels nearby enemies.
  • Slardar L25: Corrosive Haze undispellable.
  • Sven L15: Storm Hammer dispels enemies.
  • Templar Assassin L20: Refraction dispels.
  • Troll Warlord L25: Battle Trance dispels your team.

Wow, that's a lot of buff interaction!

I think a hallmark of this patch is that Venomancer's Venomous Gale ability became dispellable, after many years to the contrary. Venomancer deals all his damage over time, so the enemy having dispels (a tiny proportion of matchups) would have been a hard counter to him.

7.07 gave him some talent changes that make the hero less dependent on his Gale, and let him fit better into a game where dispel is now a fully-integrated mechanic.

Ongoing Trend – Evasion

In Warcraft there were very few sources of evasion, so few mechanics to counter it existed (if any). I don't believe True Strike ever actually makes an appearance in Warcraft itself: it seems to have been a beta mechanic. Thus, the two heroes with evasion as a skill, owners of an evasion item, or simply units standing uphill got to enjoy an advantage uncontested.

Dota2 inherited this trend. For many years in evasion appeared on a couple of heroes, one item, and it had exactly one counter: the expensive item Monkey King Bar whose True Strike passive causes attacks to never miss.

Dota2's patches have been steadily integrating evasion better as a game mechanic. The history goes something like this:

  • 6.82: The evasion-granting item Butterfly gained an active which allowed heroes to temporarily trade their evasion for increased movement speed.
  • 6.84: Solar Crest was introduced as an item which grants evasion, and its active can gift temporary evasion to an ally, or apply temporary Accuracy (evasion reduction) against a target enemy.
  • 6.87: Bloodthorn arrives: another soft-counter to evasion which grants temporary True Strike against a single target.
  • 7.00: Twelve heroes now gain access to evasion through their talents.
  • 7.06: The heroes Earthshaker and Witch Doctor have True Strike made available on suitable abilities.
  • 7.07: The Javelin and Monkey King Bar items cannot miss if they proc. Monkey King Bar's chance to not miss is now 70%.
  • 7.14: the Maelstrom and Mjollnir items cannot miss if they proc.

At this point, evasion is a well-integrated mechanic and part of gameplay which has no hard counters, and a wide variety of soft counters. The result is much healthier than the previous binary relationship between "I have evasion" and "I don't have evasion".

What Next?

Each of the patches we've discussed above has added depth to the game by taking a previously niche mechanic and making it better integrated and something that more heroes can interact with.

Most of these mechanics have their roots in Warcraft III. With a probable content patch looming after TI8, it is an open question whether the team will dig once again into Warcraft's mechanics, or take the opportunity to introduce something new.

If it's the former, then here's the Warcraft mechanics I think are most likely to be chosen:

  • Magic immunity is a fundamental part of the game, yet the only item that grants it is Black King Bar. Several other items have had their effects decomposed into smaller items like Shadow Amulet and Talisman of Evasion: that pattern could continue. One form of immunity that appeared in Warcraft was Anti-Magic Shell: a buff that can be dispelled (Repel is based on it). Could something similar appear on an item?
  • Powerup items have seen recent popularity, with four bounty runes being added, and the river now spawning two after 40 minutes. Powerups were a significant feature in Warcraft; perhaps we'll see heroes with abilities that can spawn or manipulate them?
  • Buildables, essentially the ability to construct small towers/buildings around the battlefield. Warcraft supplied the item Ivory Tower which let heroes place a tower without workers, and several old maps adapted that into their gameplay. Dota2 already has moved towards more interactive buildings (shrines), and has a mature territory-control game through wards that could be built upon (pardon the pun).

If the team decides on new mechanics instead, my guess would be:

  • Terrain interaction has appeared on the three newest heroes (Tree Dance, Rolling Thunder, Terrorize), suggesting we could see more of this going forward: perhaps the ability to create temporary cliffs or walls. This has come up in forum suggestions before: with abilities like Pit of Malice being imagined as actual craters that de-elevate the terrain.

Any of the above could serve to further deepen the game. Whatever the mechanic, I expect the 7.07 patch to be used as a model: with new heroes and some talent changes playing a role in its introduction.

As a final note, I want to emphasise that a hell of a lot of thought went into Dota2's patches. It's one thing for me to sit back and analyse, but quite another to plan these moves in advance, even with the backing of a design team. Icefrog is highly acclaimed for his work on balance, but what really steers the ship forward is content patches: and credit is well-deserved there too.