Dota Underlords Tier List: Best Alliances and Builds
Artist: Valve
Date: 11/2019

In this article, you'll learn about the best Alliances, Builds, and Strategies in the current Underlords Meta. All in-depth guides about specific builds can be found in our Dota Underlords Builds collection.

If you're interested in information about each individual hero, you can get it in our Underlords Heroes Tier List.

Best Alliances:

There are two main criteria when we order the Alliances.

  • Late Game Potential: at the end of the day, the best late-game builds determine the first positions in any Dota Underlords lobby. We tested 50+ builds in Freestyle mode (details below) to get a good grasp of how they perform in the late game and our results were a major influence on the order of the Alliances in the tier list here.
  • Reliability: how easy and reliable the strategies are to implement in real games is another big concern. If they require multiple things to go your way (specific items, multiple specific heroes, specific upgrades), then they are at a disadvantage against more flexible builds. This situation is made worse by the Jail system for some Alliances.

The best Alliances in the current meta are ordered below with info about each one. We share our thoughts about both the late game strength and reliability of each Alliance in Dota Underlords.

The icons are clickable and reveal the detailed info:

S Tier
A Tier
B Tier
C Tier:

Warriors: very good reliability, very good late game potential.

  • They are a versatile opening strategy. (3) Warriors can transition into Hunters, Mages, Warlocks, Trolls, Savage, but you can also simply commit to Warriors, etc.
  • They are strong in the early-mid game even after the (3) Warrior armor nerf. The individual heroes have good stats (especially survivability) and it's easy to utilize other secondary Alliances that are strong in the early/mid game (Druids, Savage, etc.).
  • They are strong in the late game: their six-hero bonus is great stat-wise, and Warriors are the Alliance with the most AoE control in the late game, which is extremely valuable. They have great 4 and 5 cost heroes and aren't too vulnerable to the Jail system.

In conclusion, Warriors are quite reliable, and at the same time they have some of the best late-game strats in Dota Underlords.

You can find our Dota Underlrods Warriors Guide here.


Primordials: the very best late-game potential, average reliability

Primordials are simply the best Alliance in the late game. After the big update, they got the third tier to the Alliance bonus (with (6) Primordial heroes) in addition to an Ace effect and two primordial 5-cost heroes.

This in itself makes them very strong in the late game, but another huge benefit is that they don't need upgrades to be strong. Enigma and Void can be a one-star hero in the build and it still performs extremely well. Moreover, the actual Alliance effect doesn't scale with the power of the actual hero (even a one-star one-cost Primordial will produce the powerful rank three Eidolon). This makes the build very cost-efficient and means you need much less gold to beat convincingly other strategies in the late game.

The downside of Primordials is reliability. They are vulnerable to the jail system because Arc Warden, Enigma, and Void are all necessary to reach the full potential of the Alliance. Moreover, the fact that you are searching for two (or even three) Ace heroes means that you need lvl9 as fast as possible, which isn't possible in every single game. That said, the low-tier Primordials are quite good (Morph and Tiny specifically), which means that Primordial drafts are actually not bad in the early/mid game by any means.


Hunters: very good reliability, average late game potential

We entered the 50+ Builds Test with the expectation that Hunters will perform extremely well. It turned out, however, that the best Hunter build in the late game is a Warrior build - (6) Warriors (3) Hunters.

The (6) Hunter builds performed well, but they weren't spectacular.

The real strength of Hunter builds, however, isn't their late game strength, but their reliability. They have a good gradual power curve (good at all stages of the game), and that they aren't vulnerable to the Jail system or the shop - there are quite a few interchangeable Hunter heroes you can use in the builds. This means they are reliable and easy to build.

So, even though they aren't the strongest late-game build, they are one of the Alliances that can certainly give you great results on average (you can even argue they are S-tier because of this, but we are more comfortable with putting them as the first A-tier Alliance).

For more details you can check out our Dota Underlords Hunters guide.


Heartless: very good late game potential, decent reliability

When the Big Update came out, we expected Heartless to be a very good secondary or tertiary Alliance bonus but something that's hard to build a strategy around. It turns out, however, that the Heartless bonus is good enough to justify using it as the primary bonus on top of which you build your strategy.

Armor is very important in Dota Underlords and armor reduction is just as good offensively as bonus armor - defensively (Warriors). The main benefit of Heartless is that it amplifies the damage of your whole army, not just the Heartless heroes.

The usual successful Heartless builds are (6/4) Heartless combined with Hunters. The great performance of (6) Heartless (4) Brutes in our late game test, however, suggests that other combinations are certainly possible and viable.

Last but not least, Heartless drafts are relatively reliable because most of the time you don't need specific heartless heroes and upgrades to reap the full benefits of the bonus. The important upgrades in Heartless builds are usually of the carries members of the secondary Alliance you're running in the build (Lefestealer being an exception).


Mages: great late game potential, bad reliability

First, Primordials + Mages seems to be the best Alliance combination in the game right now. Second, Mages received a buff after Morphling became a Mage (our results in the late-game test for Mages improved drastically in spite of the KotL nerf).

The test results of Mages are quite impressive considering they are not ideally suited to beat the benchmark builds - both are tanky against magic damage - Knight bonus in the first case and the Scaled bonus in the second case. It's very likely that their score will rise if other benchmark builds are introduced. Moreover, they are at a disadvantage in the lvl9 test threshold because they don't benefit from three-star upgrades and prefer to reach lvl10 (i.e. their score will rise if we focus more on lvl10 rather than lvl9 builds).

Their good late game potential aside, Mages have some reliability issues:

The build is extremely reliant on the Humans bonus. Because of this, Mages are arguably one of the most vulnerable Alliance to the Jail system in the whole game. Losing any of their three Human Mages can be a huge hindrance.

Moreover, Mages perform disproportionately well with Humans and Primordials, and they underperform with other Alliances (even viable ones like Inventor Mages and Warrior Mages). Taking into account Primordials are also vulnerable to the Jail system, this contributes to the problem.

Our Dota Underlords Mages Guide.


Assassins: very good late game potential, unreliable

Assassins are, on average, the second-best late-game Alliance in our late-game test, which was a big surprise. All five tested Assassin builds performed quite well against both benchmarks.

(A disclaimer is that Assassins are quite decent versus both benchmark builds by definition (Knights & Hunters), and their late-game score might fall if they are tested against builds with more control and burst damage (Warriors,  Primordial/Mages, etc.).)

A major reason for their late-game strength is that a three-star Slark with Mask of Madness is hands down the strongest carry in Dota Underlords. The doubling of all hero HP makes rounds longer, which gives Slark a lot more time to build up his stacks and become unstoppable. Moreover, the fact that he has double the HP makes him harder to burst down and stop early in the round.

The fact that you can easily get the (3) Elusive bonus in any Assassin build is also a big benefit for Phantom Assassin - the other exceptional Assassin carry. It's worth mentioning that the best-performing Assassins build in the test was (6) Elusive (3) Assassins, which is a good proof of how much PA benefits from the Elusive bonus.

That said, Assassins are disproportionately reliant on PA and/or Slark (ideally on three stars) combined with specific items, most importantly by far - Mask of Madness. This makes Assassin builds quite unreliable and forcing Assassins blindly is unlikely to give you great results.

Our Dota Underlords Assassins Guide.


Warlocks: very good late game potential, below-average reliability

Initially, we thought Warlocks would only be a strong secondary synergy, but it turns out (6) Warlock drafts are entirely viable in the late game. (6) Warlocks (3) Healers and (6) Warlocks (3) Insects performed exceptionally well in our late-game strength test.

That said, Warlocks are unreliable because you need Disruptor to reach their full potential and because they are quite hard to build in the early-mid game (their low-tier units are low-impact and mostly utility-focused). Because of this, we believe the best way to utilize the late-game power of Warlocks is to use the (4) Warlocks bonus (with Disruptor) in combination with another Alliance that will carry you in the early/mid game - Warriors, Brute, Savage, Scrappy, etc.

Our Dota Underlords Warlocks Guide.


Scrappy: decent reliability, decent late-game potential

The important thing to remember about Scrappy is that there is a big difference in the performance of the different Scrappy builds.

(6) Scrappy (3) Assassins is extremely good in the late game, albeit very reliant on a three-star Slark (or at least PA) with Madness.

(4) Scrappy/Inventor (3) Mage, on the other hand, isn't as item-reliant, but it doesn't perform consistently in the late game depending on what build's it's facing. The Inventor explosions are very central to the build, which means it performs very well against melee strats, but very badly against ranged strategies (the Hunter benchmark build in our test was a big struggle for this strat). Moreover, it's not too reliable as well because of its reliance on finding Techies and ideally - three-star Inventor upgrades.

All in all, Scrappy builds give decent results on average, but it is a good idea to have a good look at the lobby before deciding what strat to go for and to lean more onto the Scrappy/Assassin combination rather than the other Scrappy variations.

You can find out more info in our Dota Underlords Scrappy Guide.


Brute: very good reliability, decent late game potential

Brutes are an interesting Alliance consisting of tanky heroes with a mix of offensive and defensive potential.

All in all, Brute builds are not the best in the late game, but they perform relatively well thanks to heroes like Lifestealer (who becomes a formidable carry on three stars with items like Mask of Madness, especially when combined with Magnus Empower) and Doom who has great stats and a powerful ultimate.

Their main strength, however, is their consistency. They provide you with a cost-efficient tanky frontline capable of dealing some damage, which makes it possible to go for various options for secondary Alliances (Brawny/Hunters, Heartless, Druids, etc.).


Dragons: secondary synergy

Dragons are quite versatile. You get a lot of value for the two hero slots that you need to invest. Dragon Knights + Viper or Puck is a great addition to most Knights or Mages drafts. Viper + Puck, on the other hand, is central to the Assassin + Mages strategy which utilizes the magic resistance debuff and Viper's Corrosive Skin passive.

For more details, you can check out our in-depth Underlords Dragons & Knights Guide.


Scaled: secondary synergy

Scaled a very strange Alliance because they have a very straightforward use. They are simply S-tier against Mage strats. They are, however, low-impact if your opponents don't have a lot of magic damage. The Scaled bonus is easiest to get in Warrior strategies (and very important because armor is useless against magic damage). 


Trolls: average reliability, average late game potential

Trolls got a huge boost in reliability after the Big Update with the introduction of Dazzle (you don't have to find the 5g Troll Warlord to get the full (4) Trolls bonus). That said, they are still vulnerable to the Jail system. Moreover, getting a two-star Troll Warlord makes a big difference in the late game, which means that they are by no means one of the more reliable strats.

The late-game potential of Trolls is an even more interesting topic. (2) Trolls are a good tertiary Alliance bonus in a lot of strats, but to use the full (4) Trolls bonus you usually want to run them either with Knights or with Warriors.

In our late game test, Warrior/Troll builds performed exceptionally well (the (6) Warrior (4) Troll lvl10 build was the strongest lvl10 build we tested), while Knight/Troll builds performed exceptionally badly. This was a surprise, but it suggests (4) Trolls by themselves are not game-winning and they need other factors to be successful. Judging by our test, it seems the control coming from Warriors is exceptionally valuable for them, while the lack of control in Knights is a problem in the late game.


Elusive: good late-game potential, bad reliability

The (6) Elusive (3) Assassins build was one of our best-performing builds in our late-game test. The other Elusive builds weren't as good, but they were not terrible as well. The rework of the Elusive bonus simply synergizes extremely well with the Elusive Assassins, which turns Phantom Assassin into a super-carry.

Elusives, however, have two major reliability problems:

  • First, they lost their only tank (Treant), which means a full (6) Elusive draft is actually very hard to build in the early/mid game.
  • Second, they are overly-reliant on a single hero - Phantom Assassin. This is a problem considering the jail system. And even if PA isn't banned, if other players in the lobby are contesting her (e.g. Assassin players), this could become a big problem.
Out Elusive Builds Guide.


Druids: very good reliability, bad late game potential

After the Big Update, Druids became quite a bit more flexible, which means they are an interesting addition in a lot of strats. That said, their role in the game didn't change much - the star-level advantage they provide is extremely strong in the early and mid game, but its impact falls off drastically in the late game.

In fact, (4) Druid strategies have the second-worst score in our late game strength test.

This means that having (4) Druids in your final draft is a good idea only when you're snowballing from their early and mid-game strength and have a better economy than your opponents. If you don't have such a big lead, it might be a better idea to stick to two Druids or to transition out of your Druids entirely in favor of stronger late-game heroes and Alliances.


Knights: the anomaly in our test

Knights were by far the biggest surprise in the late game test - they had the worst average score by a large margin.

It seems quite unlikely that Knights are actually the worst late game Alliance in Underlords - there might be another reason for the bad results:

There are just a few (realistically two) good Knight builds and a whole bunch of underwhelming ones.

The good builds are the benchmark build (Knights/Healers/Dragons) and (6) Knights + (4) Trolls. All other possible builds (Knights + Mages, Hunters, or (4) Knights (4) Troll variations) underperform a great deal and drag the average of Knights down a lot. If we simply delete these builds (forget they exist) and use only the good ones, the average of Knights will rise drastically.

So, if you stick mostly to the two good Knight variations, you should be able to have good results with the Alliance. Knights are quite reliable because of a good power-curve and good mid-game strength.

That said, even the good Knight builds aren't amazing in the late game. They lack any kind of meaningful control and are hard to split apart to avoid AoE, which means winning with Knights usually happens with a significant economic advantage.

Since upgrades are also important for Knights, it's a great idea to run them when they are uncontested (which happens more often after the Big update).

For more details, you can check out our in-depth Underlords Knights Guide.


Savage: good reliability, bad late game potential

Savage strategies have always been momentum-based - i.e. you can use them in the early-mid game to secure a winning streak, but they fall off hard in the late game: Savage builds have the third-worst average late-game score in our test, and this wasn't a big surprise.

Unsurprisingly, they perform way better versus slower strats (our Knights benchmark) and much worse against fast, nuke-damage straits (our Hunters benchmark) because they need time to accumulate bonus damage.

The best performing Savage build we tested was the Savage/Trolls/Warriors variation.

You can get more details in our Savage: good reliability, bad late game potential

Savage strategies have always been momentum-based - i.e. you can use them in the early-mid game to secure a winning streak, but they fall off hard in the late game: Savage builds have the third-worst average late-game score in our test, and this wasn't a big surprise.

Unsurprisingly, they perform way better versus slower strats (our Knights benchmark) and much worse against fast, nuke-damage straits (our Hunters benchmark) because they need time to accumulate bonus damage.

The best performing Savage build we tested was the Savage/Trolls/Warriors variation.

You can get more details in our In-Depth Savages Guide.


Brawny: decent late game potential, bad reliability

Despite not being able to test them in our late-game-strength experiment, Brawny builds are with a good late game potential while being unreliable by design because of the way the bonus scales.

If you get early two-star brownies that are able to accumulate kills, they will snowball into the late game. If you don't, you'll struggle to generate a meaningful amount of kills.

This makes Brownies an opportunistic Alliance - if the shop allows for it, it's a great idea to make use of them (in combination with Brutes, Savages, Hunters, etc.). If it doesn't - you're better of ignoring the bonus.


Humans: secondary synergy

After the rework Humans simply became an extension of the Mages bonus. They are extremely important in Mage strategies and quite unimportant in all other builds.


Demons: secondary synergy

Before the Big Update, running a Demon Hunters + Demons strategy was not great, but it was at least viable.

Now, however, with the Demon bonus change, a full Demons strat is almost impossible to make work.

If you have too many Demons they will be silenced for most of the fight, which means not only that they won't use their spells often enough, but also that they won't get the damage increase in time.

After the Big Update, Demons are simply a secondary synergy viable in many other strategies. The big difference is that instead of running a single Demon, you can often run two, or in very rare occasions - three.


Deadeye: secondary synergy

You don't care about the Deadeye synergy outside of Hunter strategies - it's simply not that impactful, and it's not game-breaking even with the Ace bonus. That said, the Deadeye units are actually decent, so you're not actively trying to avoid them.


Shamans: secondary synergy

(2) Shamans are actually a very decent tertiary bonus for some strategies. The (4) Shamans bonus, however, doesn't justify the additional two hero investment at all.

We even tested a (4) Shamans build and then we tested the same build but with one of the Shamans replaced by a better standalone hero, and the second variation performed significantly better.

This simply means that building a strategy around the full (4) Shamans bonus is not a great idea in this meta.


Blood Bound: unreliable, below-average late game potential

Blood Bound is the most unreliable strategy in the game because of the simple fact that it depends entirely on a single item (Big Time Contract). Without it (ideally multiple copies), you cannot even build a decent Blood Bound build.

To top it off, the late-game potential fo Blood Bound builds is not great - with the increase of all hero HP in the big update, the high damage coming from BTC carries is no longer absurd and you will rarely see a hero getting one-shot.

This means that even if you have the item, you usually want to use Blood Bound for a mid-game win streak and transition out of it in the late game.

60+ Builds Tested

We tested more than sixty builds in the new freestyle mode to check their late game strength and efficiency. We were testing for the lowest possible build cost to reach the same power-level (to be able to beat two benchmark builds) on lvl9 or lvl10.

Here’s the summary of the results:

If you click on the image, you’ll be taken to the spreadsheet with the detailed table of the builds including their individual results, builder links, and freestyle mode code for each encounter so that you can check them out yourselves.

You can read our thoughts about the results of every individual Alliance above in the Best Alliance section.

We also say a few words about the best performing indicidual builds in the test below in the Best Builds section.


Cost Efficiency:

We are testing the builds for late-game cost efficiency. We are trying to take all builds to the same power level (until they are able to beat the benchmark builds) and to do it for the least possible cost (while trying to keep things realistic – we keep special repositioning to beat the specific benchmark build to a minimum). This means the best builds in this test are the builds that become the most powerful for the least gold. We also record the average damage the builds deal once they are able to beat the benchmarks, but we optimize for cost, not for the damage, so the numbers in the damage column are less important and reliable.

The Benchmark Builds:

In the ideal situation, each build would be tested against each other build in a round-robin fashion. This, however, would take an unrealistic amount of time for one person to set up and test, so we settled on a more realistic method: to test each of the builds against two standard/common benchmark builds – one common tanky build (6 Knights 3 Healers), and one common DPS-oriented (6 Hunters 3 Warriors). The idea is that some builds would naturally do better against the first while struggling against the second and vice versa, and the average results would be a good representation of the late-game performance of the builds in general.

Knights, the Tanky Benchmark Build:

One of the most standard and popular builds in the game that focuses mostly on universal survivability (good both against magic and physical damage). Before the big update it ran (2) Warlocks. Now, it runs (3) Healers, but the idea is the same.

Hunters, the DPS Benchmark Build:

(6) Hunters with a tanky (3) Warriors frontline is another classic build that is on the other end of the scale – it relies on a very high damage output to win fights. Even though (6) Hunters (3) Insects or (6) Hunters (4) Heartless might be trendier now after the Big Update, (6) Hunters (3) Warriors is THE classic (6) Hunters build and is unlikely to go anywhere even after further balance changes.

Same level, no Underlords:

Since we are testing builds for hero cost efficiency, we needed to test them on the same level in order to equalize as much as possible the experience and reroll gold cost. We settled on lvl9 whenever possible. The reason is that lvl9 is a good compromise between upgrade-value strats and level-value strats.

Builds that stay on lvl7 and 8 to search for 3-star upgrades often push to lvl9 towards the end of the game to sneak in some extra Alliance bonuses and to upgrade any 4/5 cost heroes. At the same time, builds who want to push towards lvl10 as fast as possible often don’t have the luxury to do it right away and are forced to stay on lvl9 to search for some key heroes and upgrades and stabilize before they push towards lvl10.

This means lvl9 is the “fairest” late game meeting point for most builds in the game. That said, some builds simply need lvl10 (e.g. (6) Warriors (4) Warlocks). To accommodate for them, we tested lvl10 builds separately against lvl10 benchmarks (the same as the lvl9 benchmarks but plus 1 hero and minus some upgrades).

To eliminate another level of complexity and make the results more reliable, we also decided to test all builds without an Underlord.

Reading the results:
Underlords Builds Test Freestyle Mode

  • Average:

The lower the cost the better. E.g., on average, the 4 Primordial strats we tested needed 78.75 gold invested in heroes to beat the 102g Knight and 104g Hunter benchmark strats, which needless to say is good.

  • Median:

We give the median for a different reason. Let’s say we test 3 good Primordial strats and only one bad one. At the same time, we test 5 bad Troll strats and only 2 good ones. This might be fair because those are the viable strats of said Alliance, but it could also just be our own lack of creativity to think of more good Troll strats.

Obviously, the Primordial strats will have a better average in this case. The median, however, remedies this situation a little bit. If we tested more bad strats rather than good strats for the specific Alliance, the median will be higher than the average. In this case, the best thing to do would be to see which builds for the Alliance are performing well and focus only on them rather than to disregard all builds in the alliance as a whole.

  • Standard Deviation:

The standard deviation, in general, represents how volatile the builds within an Alliance are. For example, Assassins have a low standard deviation, which means the different Assassin builds perform similarly to each other. Mages, on the other hand, have a huge Standard Deviation, which means that there are some great Mage strategies and some terrible ones. Primordials + Mages is simply miles better than Knights + Mages.

  • Disclaimer:

First, the fact that a build has great/bad results in the experiment is only half of the picture and doesn’t necessarily mean that it’s the best/worst build possible. Late game potential is just one (very important) side of what makes a build good. Reliability is the other, so we talk about that in more details in the Alliance and build explanations above and below.

Second, it’s certainly possible to improve some of the individual builds and scenarios in the experiment with further tweaks, which would give the specific build a slightly better score. Of course, we couldn’t afford to spend an indefinite amount of time on each scenario to make it perfect, so such “mistakes” are inevitable. Nonetheless, we are fairly confident that further tweaks wouldn’t lead to massive changes to the final average results.

Third, we were unable to test strategies with Brawny heroes in them because for some reason Freestyle mode generated an unrealistically huge HP number for the Brawny heroes (400+ kills). If we’re able to fix this problem we’ll revisit the Brawny strats. 

Here are a couple of non-build-specific general lessons from the experiment that you can take to your own games:

  • Not all upgrades are equal:

Utility heroes give much less value when upgraded to a higher star level than tanks and damage-dealers. For example, a one-star Disruptor is usually more than enough to make a (4) Warlock build very strong. You use him for his Ace effect and the AoE silence. His stats and the damage of the silence don’t matter that much, so a two-star upgrade on Disruptor doesn’t make a huge difference. For contrast - a two-star Troll Warlord is much more impactful than a one-star Troll Warlord and can make a big difference.

Even more interestingly, there is a big difference between a two-star and a three-star Axe despite the fact that he is a much lower tier unit. Axe is all about the stats – his EHP is quite a lot higher on three stars, which means he can live for longer and buy your backline damage dealers more time to win the fight. In that sense, getting your Axe to lvl3 might actually be more important for your final placement than getting your Disruptor to lvl2.

Understanding which units are your upgrade priorities can help you prioritize bench space easier and improve your results as a whole. The rule of thumbs is that you need upgrades on the heroes that need high stats (tanks and damage-dealers), while you don’t on heroes who you keep mainly for the Alliance and Ace bonuses.

  • Positioning is game-breaking in the late game:

This sounds like an obvious statement. However, I don’t mean just spreading your heroes to avoid AoE, etc. Absolutely minor changes on the board can turn the outcome of the round when the two armies are close in power.

For example, we had two corner box formations fighting. Something as simple as swapping the position of Dazzle and Witch Doctor within the formation (both utility heroes, not the main carries or tanks in the draft) changed the outcome of the fight. The reason in this case was that Dazzle was a more valuable unit in that particular fight because the round was long (the fight was against the Knight benchmark build) and putting him on the backline meant he can cast more heals over the duration of the fight, bringing a lot of value. Multiple Witchdoctor Casks, in comparison, weren’t as impactful.

In other fights, however, the opposite might be true. For example, against Assassins a couple of stuns would certainly be more valuable than a couple of heals because Assassins nuke heroes quickly anyway and reduce healing done.

So, if you’re fighting against Assassins the Witch Doctor should be the hero you’re protecting more rather than the Dazzle, while against Knights the opposite is true.

  • Damage is not linear:

A very small difference in power could lead to a big difference in the damage inflicted at the end of the round. What we mean is that a strategy would be losing by ~15 damage against a benchmark build, and then improving it the slightest (e.g. one more hero upgrade) would cause it to win by ~15 damage instead.

The main culprits for this effect are usually AoE damage and heals. Some strategies win or lose gradually, but AoE damage and healing strategies usually have drastic damage swings.

This is easy to imagine with Mages – if your board is slightly weaker than the enemy board, a lot of enemy heroes will survive on a sliver of HP and will inflict a lot of damage to you at the end of the round. Hunters are the opposite – you bring down enemy heroes one by one, so it’s entirely possible that the last standing enemy hero is the one that beats you, inflicting minor damage at the end of the round.

Another great example is Warlocks. They utilize a lot of healing. This means that if they win a round, a lot of the heroes in the Warlock build will survive and will deal a lot of damage to the enemy player at the end of the round. It’s very unlikely that a Warlock build wins the round with just one or two heroes left.

Again, you can check how each build performed in the test in this spreadsheet.

Best Underlords Builds:

The example lineups are tabbed - you need to click the title to reveal the information about the alternative lineups.

Best Knight Builds in the current Meta:

Knights are one of the most popular and reliable strategies historically.

The key things to keep in mind for Knights:

  • There are a lot of underwhelming Knigut builds that you should avoid unless you're a very experienced player: Knights + Hunters, Knights + Mages, (4) Knights + Trolls variations.
  • Use the (6) Knights variations, either with Healers/Dragons or with Trolls.
  • Keep in mind Knights are mostly an upgrade-value strategy: it's a good idea to stay on lvl7 and 8 for a while until you find some important upgrades

This is the classic, arguably most popular and reliable variation. It focuses almost entirely on survivability and sustain. To compensate for this, it's extremely important to find a three-star upgrade on your Luna (and ideally on multiple other Knights) and decent items for her and Dragon Knight (they are your damage dealers).

This is the second viable option - the (6) Knights (4) Trolls variation. It focuses much more on DPS because of the (4) Trolls attack speed bonus. While an upgraded Luna is also important in this variation, getting to lvl9 relatively fast is also vital because a two-star Troll Warlord is a big power-spike.

You can check out other builds in our in-depth Dota Underlrods Knights guide.

Best Hunter Builds in the current Meta:

Hunters are quite straightforward - they are single-target backline damage dealers. Some key things to keep in mind to make Hunters work:

  • Hunters will give you great damage output, but you usually need to think about a good frontline because they are very squishy.
  • Medusa makes a big difference. Even though Hunters benefit from upgrades, don't delay going to lvl9 for too long or you might find her too late.
  • Drow is Heartless, so try to include at least one other Heartless hero for the (2) Heartless armor reduction bonus (your whole lineup deals physical damage).

Out of the possible (6) Hunter builds, the (6) Hunter (4) Heartless variation is one of the best options - it deals extremely high physical damage and it has a good gradual power curve (you can do very well in the mid game with (4) Heartless (3) Hunters).

(6) Hunters (3) Insects is the other trendy and well-performing variation.

The classic (6) Hunter (3) Warriors option, however, shouldn't be underestimated. It's a great combination of high damage and a tanky frontline with great control.

You could argue Hunters and Warriors are the two best Alliance to combine in the whole game. The (3) Warrior (3) Hunters opening is one of the most flexible in Dota Underlords and there are tons of viable late-game variations coming out of it you can use to win games.

You can check out other builds in our in-depth Underlords Hunters Guide.

Best Warrior Builds in the current Meta:

Arguably, the best Alliance in the game. Extremely flexible, but with very high late game potential. Very tanky with amazing crowd-control.

Keys to success with Warriors:

  • Arguably the biggest strength of Warriors are their flexibility. Open up with a (3) Warrior frontline and see what the shop gives you to decide what strategy you'll be going for. Knowing multiple viable builds will give you a big advantage.
  • Prioritize the crowd-control Warriors. Disables + Tankiness is what wins Warrior builds the late game. Tide, Kunkka, and even Tiny are great heroes to include.
  • Make use of the (2) Scaled bonus. You are vulnerable to magic damage, but there are two Scaled Warriors (Slardar and Tide). Problem solved. Slardar is also a good hero to prioritize for a three-star upgrade - great stats and the ultimate helps for the DPS output of your army.

The (6) Warriors (3) Healers variation performed the best against the two benchmark builds in our late-game-strength test. Combining a lot of survivability with a lot of sustain is never a bad idea, and the build allows for a lot of minor Allliance bonuses.

That said, we cannot move on without mentioning the (6) Warriors (3) Hunters variation. You could argue it's the best build in Dota Underlords if you take both reliability and late game potential in account.

Check our more Warrior Builds in our Dota Underlords Warriors guide.

Best Mage Builds in the current Meta:

Mages are vulnerable to the Jail system, but when their key units are in the game, Mage builds have very high potential.

Some important things to keep in mind:

  • Mages are mostly a level-value strategy. You need a two-star Keeper and ideally even a two-star Lich, and at the same time, three-star upgrades of lower-tier heroes don't make much of a difference. Get to lvl9 ASAP and if you can afford to - try to push for lvl10.
  • Mages are very reliant on the Human Alliance mana bonus. You should aim for (4) Humans in almost every Mage build.
  • Mages over-perform when combined with Primordials, so abuse this fact.

After the Morphling change, it's hard to argue that (6) Mages (4) Primordials (4) Humans isn't the best lvl9 Mage build in the current Dota Underlords Meta. If you have the luxury to push for lvl10, you can start replacing some of your weaker Primordials with high-tier heroes like Void and Arc Warden (if you have items for him) and you can include Lich. It's also possible to go for (6) Mages (6) Primordials (the best lvl10 build in our test), but by doing that you'll have to lose the (4) Humans bonus.

You can check out more Mage builds in our Dota Underlords Mage Guide.

Best Elusive Builds in the current Meta:

Elusive strategies were changed fundamentally in the Big Update. Frankly speaking, currently, they are very closely connected with Assassins, in almost a similar fashion to Mages and Humans. Some keys to keep in mind:

  • Your core of the draft is (3) Elusives and (3) Assassins. Don't try to force the (6) Elusive bonus right away. It's an option for the late game. Early on you'll have to use other (possibly temporary) heroes. With the loss of Treant, Elusives simply don't have good frontline heroes.
  • Your number one most important hero by far is Phantom Assassin. Getting her to three stars is vital. A good item is also game-breaking (the ideal option is Mask of Madness).

This variation, unsurprisingly, performs extremely well in the late game. The (6) Elusive bonus, the Void Ace bonus, and the (2) Scaled bonus make your Elusive Assassins extremely hard to kill. They already deal high damage and the draft has plenty of control, so you have all the tools you need for success. The problem is getting to that point. You'd usually need to use temporary cost-efficient frontline heroes (e.g. two-star Tiny, etc.) until you're able to complete the (6) Elusive bonus and get to lvl9 for the Ace heroes.

Best Savage Builds in the current Meta:

Some keys for success with Savage builds:

  • Savages are a snowballing early/mid game Alliance. Use your early strength to secure a good economy.
  • Savages fall off in the late game. Keeping savages in your final draft is certainly possible, but more often than not replacing most of them with better late game heroes will yield better results. A Savage/Warrior/Warlock opening that transitions into a Warriors/Warlocks late-game draft is the standard option.

The (4) Troll (4) Savage variation, unsurprisingly, performed best in our late-game test. Savages and Trolls have a very obvious synergy - the increased attack speed helps the Savages accumulate bonus damage faster.

The (3) Warriors and (2) Warlocks, on the other hand, help the build survive for longer.

Best Assassin Builds in the current Meta:

Assassins are in a good spot since the Big Update. Some keys for success with them:

  • (6) Assassin builds are certainly viable, but (3) Assassins with another major Alliance (e.g. (6) Scrappy) are more stable and should be your first consideration.
  • Assassin builds are upgrade-value builds - three-star upgrades on Slark and Phantom Assassin in addition to good items (most importantly Mask of Madness) are absolutely crucial. Stay on lvl7-8 until you find your important upgrades. Higher levels are a luxury.

The Insect Assassins are certainly one of the most interesting (6) Assassin variations after the Big Update. This build provides good control in addition to decent survivability thanks to the miss chance coming from the Insect bites in combination with the (3) Elusive evasion.

That said, most (6) Assassin builds perform at a similar level, so you defenitely don't have to stick strictly to Insect Assassins.

Best Scrappy Builds in the current Meta:

Scrappy lineups are quite reliable. They can work well in combination with Warlocks, Mages, or Assassins. The Scrappy + Assassins variation tends to give the most consistent results, but the others are certainly also viable.

Some keys to success:

  • Scrappy builds are usually upgrade-value builds. If you are using Assassins, three-stars on your Assassins is the priority. If you are mostly winning thanks to the Inventor explosions, however, you need to prioritize upgrades on your high-HP inventors (Timber, Techies, Tinker if you're running Mages).
  • If you rely on Inventors, however, you need to get to lvl9 sooner rather than later - the Techies Ace bonus is very impactful. This is one of the reasons why the Assassins variation gives the most consistent results - it doesn't need Techies to work well.

This is a good example of a lvl9 Scrappy Assassin draft. The most important thing in this build is a three-star Slark with Mask of Madness. Most other things are optional: 

  • You can run it on lvl8 without the Warlocks and Aces if you need to.
  • You can replace the Shadow Fiend with another high star-level Assassin (usually Phantom Assassin) if you wish.

Best Primordial Builds in the current Meta:

According to our late-game strength test (and out in-game experience), Primordials are the biggest contenders for the strongest late game build in the game. Some keys to success:

  • Primordials are a level-value build. They don't need upgrades, but they need multiple high-tier units (up to three Aces), which means pushing to lvl9 and even 10 as fast as possible is very important.
  • Primordials are very vulnerable to the Jail system. If Arc Warden, Enigma, and possibly Void are banned, avoid going for a (6) Primordial draft. Moreover, Primordials work best with Mages, but if Keeper and other Human Mages are in the jail go for other variations (usually Primordial Assassins).

This is a relatively standard (6) Primordial (3) Mages strat. If your Tiny is only on two stars, you can swap him out for a two-star IO placed next to the Arc Warden to increase his DPS further.

An this is the best-performing lvl10 build in our test - (6) Primordials (6) Mages.

Best Warlock Builds in the current Meta:

Warlocks are usually used as a secondary synergy for other durable drafts - e.g. Warriors + (4) Warlocks. That said, (6) Warlock build are also viable and quite strong in the late game. Some keys to success:

  • Don't force (6) Warlocks from early on. The cheaper Warlocks have terrible stats and cannot carry you in the early/mid game. Play up to (4) Warlocks early on, and push to (6) Warlocks only in the late game.
  • Disruptor is absolutely crucial if you're playing more than (2) Warlocks and pretty much doubles the effectiveness of the Alliance bonus. This means Warlocks are level-value builds. Don't focus on hero upgrades too much, try to reach lvl9 ASAP.
  • If you are running (6) Warlocks, your main problem is having enough damage. A three-star Shadow Fiend with a good DPS item is your easiest option.

This is arguably the most interesting (6) Warlock build after the Big Update (and the best performing Warlock build in our late-game test). The (3) Healer bonus actually amplifies the heals coming from the Warlock bonus, which needless to say is quite powerful.

In this build, getting the Lifestealer to three stars and giving him a powerful item is a great idea - he is one of few frontline heroes and your main carry. More stats on him help a lot.

You can find more Warlock builds in our Dota Underlords Warlocks Guide.

Best Troll Builds in the current Meta:

Key to success:

  • Don't rush the (4) Trolls before you have other heroes to benefit from the attack speed bonus. The early game Trolls are mostly utility heroes and they don't use the attack speed too well themselves. The exception is Batrider, who will be your main carry in the early/mid game.
  • Reaching lvl9 is important not only because you need to find Troll Warlord, but because upgrading him to three stars makes a big difference.

This is the best-performing lvl9 Troll build in our late-game strength test. The tankiness and control from the Warrior frontline and the sustain from the (3) Healers and (2) Warlocks buys the Trolls a lot of time to deal damage. Your most important upgrades are a two-star Troll Warlord and a three-star Slardar.

Beast Heartless and Brute Builds

Surprisingly, the best-performing Heartless and Brute build in our late-game test is the (4) Brute (6) Heartless build shown below.

That said, this is likely not the most reliable and consistent build for those Alliances. Heartless are likely easier to make work as (4) Heartless in Hunter builds, while Brutes are arguably most consistent when combined with Druids or Hunters/Brawny (if you get a good Brawny start).

Thanks for reading! If you enjoyed the article, stay tuned for more Underlords content (you can give us a follow on our social media below!).

You can see the rest of our Dota Underlords guides in our Library and our articles about the game in our blog.

A big thank you to everyone in the Dota Underlords community who voted in the poll - your involvement makes a big difference in producing high-quality content!

Special thanks go also to our sponsors, EGB, who make publishing free content like this possible. If you are into esports betting you should visit them - they offer great live odds and many bonuses on professional Dota 2 matches as well as all other popular esports titles!


Kyril "MrNiceGuy" Kotashev

Esports and gaming enthusiast since forever and founder of Dotahaven.