Saturday, July 19, 2014

TI4 Group Stages vs TI3 pick + ban rates

I've collected some data from DOTABUFF on hero pick and ban rates for TI4 versus TI3.

I've laid them out in two infographics as follows

Hero Pick Rates in TI4 versus TI3 sorted by frequency

This shows you what the average game team composition looks like in TI4 versus TI3 - the popular carries, supports, etc.

Hero Pick + Ban rates sorted by TI4 vs TI3 variance

This shows you the relative changes between TI3 and TI4. In particular, you can see Lycan and Brewmaster were ignored in TI3 and have become staples in TI4.

Most hero changes are self explanatory - they're more picked in TI4 because they received buffs - but there's a couple of subtleties.

Lycan - while receiving no direct buffs, his role has changed substantially to how it was in TI2 when he was last dominant. The nerf to wolves hp made it almost impossible for him to jungle at a decent speed, however they got buffed with earlier invisibility and crazy hp regen which allows him to harass supports out of the lane. Also, pushing lycan got a big buff when Midas and Helm of the Dominator were changed to not eat Necrobook units.

Windrunner / Darkseer - while receiving no direct nerfs, these were previously the "premier" offlane heroes in TI3. Since the offlane was changed to be less unforgiving, it made it easier for other heroes to lane there.

Chen - while receiving no direct nerfs, the rise of Doom and the higher frequency of Midas pickups made him unviable to play as his creeps would keep dying. Also with the rise of Ancient Apparition his team heal was rendered useless.

Wednesday, July 16, 2014

TI4 Group Stages Round Up

Fastest Matches of TI4 Group Stages

The most notable change in meta has been the appearance of a surprising fast push strat from China. Vici Gaming, Newbee and even DK executed this several times, creating surprising wins in under 15 minutes.

DK takes Cloud 9's rax in 11 minutes 
Lumi during the draft: "DK right now have too much raw health especially with the regen that Dazzle and Witch Doctor bring, if Cloud 9 don't pick the highest damage hero in the entire game, they don't have what it takes to pierce through it."

Cloud 9 immediately picks Storm Spirit /facepalm/ leading to the fastest loss in the tournament, ending at 11 minutes.

The problem was that Cloud 9 allowed a Lycan pick through for DK and then spent their entire draft picking anti-Lycan heroes without actually thinking of how they were going to deal with the healers that came in the last 3 picks.

Link to game start

ps: if 11 minutes isn't short enough, here's an entertaining 3 minute summary.

Newbee takes Titan's rax in 12 minutes

Newbee take a level 1 Rosh which gives them an unstoppable advantage in the laning phase. Titan with their late game Naga pick simply never have time to recover: they lose all their rax by 18 minutes, the fastest in the tournament.

Link to game start

Vici Gaming take Alliance's rax in 14 minutes

Alliance go for a level 1 Roshan - they get the kill and and the Aegis, but get teamwiped by Vici Gaming who manage to scout it out. Both teams start even on gold and experience and remain on par for the the first 7-8 minutes.

Vici Gaming manage to dive the Tier 3 tower 10 minutes into the game to score 3 kills inside Alliance's base, even though the Tier 2 tower is still standing. They even make it out without taking any casualties or even much damage at all, they just heal up and keep pushing. Such an impressive team: not afraid to dive, very precise teamfights, and are careful not to overextend.

Link to game start

Vici Gaming takes DK's rax in 14 minutes
Tobi as Vici take the T2 tower: "Vici can't go high ground now."

Vici Gaming immediately go high ground and take rax.

Good example of why gold and experience graphs don't tell the full story - both teams had 0 kills, and were relatively even on gold and experience. Yet Vici Gaming were able to fight into the base even through Dk's Chronosphere, and come out of that with 5 kills to no deaths and taking the rax. Their timing push had the 3 critical items they needed - Radiance on the Lone Druid, Mek on the Viper, and Blink on the Tidehunter.

This game is also notable because Vici Gaming's support SD was only level 5 at the end of the game.

Saturday, March 8, 2014

Adventures in Low MMR (Part 2)


This is a series of articles on my adventures on a low MMR account. My main account is at 4200 rating from playing mostly support. This is an experiment to see what the experience is like climbing the ladder while supporting from a low MMR account (2200). Here's the rules I am playing by - Declare I'm supporting at the start, control map with wards throughout the game, buy the courier, don't take farm from any other heroes, obtain detection as necessary, item focus will be healing and defensive in nature. I do not play a solo lane of any kind.

Series 2

I score 5 straight wins, bringing me from 2386 to 2509. Overall there was only 1 game with an abandon.

Game 6

I won the sixth game as Crystal Maiden (duo safelane with Necrophos) at 2386 rating

Warding: The enemy team did not buy a single observer ward.

Summary: The Necrophos plays very passively and is very poor at last hitting, so we don't accomplish very much in lane at all. Our Axe and BH offlane creep cut against an Ursa and Earthshaker, who are helpless to prevent it. They score first blood on top and an early tower, propelling our Axe to an unstoppable midgame.

Game 7

I won the seventh game as Crystal Maiden (duo safelane with Windrunner) at 2411 rating

Warding: The enemy Death Prophet kept the map warded, with even some sentries to catch out BH and Kunkka out. This is the second game with enemy wards.

Summary: The enemy sent the Death Prophet to solo offlane against us. Her item choices were extremely perplexing - she bought a Blades of Attack and nothing else to start with. She gets harassed out at level 2, walks home, heals up and... still doesn't buy any regen consumables, and instead buys a set of wards. She comes back to the lane and is instantly killed. She respawns, and comes back to the lane, and dies again pretty much instantly. Windrunner and I are both level 6 at this point and we take the tower and then go help the rest of the map, while Death Prophet is still level 3 with no items. The enemy has a split push lineup (NP, DP and PL) and at low MMR levels I suspect most players have no clue how to deal with it and wasted a lot of time running from lane to lane trying to defend against the split push only to have the enemy withdraw before they got there.

Game 8

I won the eighth game as Dazzle (duo offlane with Death Prophet) at 2436 rating

Warding: The enemy team did not buy a single observer ward.

Summary: There's a particular trend where at low MMR people are allergic to picking heroes with disables. I get a typical ally Death Prophet who isn't able to contribute anything in the laning phase. Things start out very badly. Our mid lane feeds first blood, and our top lane counterpicked themselves by getting two melee heroes against a Dark Seer who maxes ion shell first. Unsurprisingly they're unable to get any farm and end up feeding the Dark Seer a double kill. I decide to rotate while refreshing the wards and score a few kills mid and top, but it takes until well into the midgame until we manage to stabilize the GPM and XPM deficit and start to pull ahead. I've decided to buy a Bloodstone from now on because I might as well try to blend in.

Game 9

I won the ninth game as Crystal Maiden (duo offlane with Naga Siren) at  2460 rating.

Warding: The enemy team did not buy a single observer ward.

Summary: The enemy picks a very gank heavy lineup (Doom / QOP / Bloodseeker / Bounty Hunter) while our carry Alchemist starts the game by maxing his passive. Naturally we get crushed in the mid game, with multiple team wipes for our team. But eventually the Alchemist farms himself to victory. This is also the first game I have seen with a BKB.

Game 10

I won the tenth game as Crystal Maiden (duo offlane with Brewmaster) at  2485 rating.

Warding: The enemy team did not buy a single observer ward.

Summary: Finally a lane partner that can do something in the early game! Brewmaster is a deadly hero at low levels, due to his above average stats and powerful thunderclap. I grab the starting haste rune and immediately score first blood with it. We manage to score 8 kills on top lane in the first 9 minutes to set us up with an unstoppable advantage. At this point the demoralized enemy team starts disconnecting.

Thursday, March 6, 2014

6.80 Hero Pick Rates and Win Rates

Working from the data collected by Phantasmal at Dotametrics in this post I've put together some more infographics on the pick rates and win rates for 6.80 by skill bracket. The estimation of Very High corresponding to >4000 rating and Normal corresponding to <3200 rating is a bit iffy: I remember seeing some DOTA2 stats site make that correlation, by cross referencing players (added by a bot to survey ratings) against their matches and how it's flagged in the system. The flags for Normal / High / Very High are still present in the system but no longer accessible to players via the in game GUI. That being said the estimation I'm using "feels" right so far, and I'll continue using it until I see evidence to the contrary.


Hero Pick Rates

Hero Win Rates

Note: Terrorblade and Phoenix are going to be highly picked this month because they've only just been introduced to the game, so they're anomalies.

Tuesday, March 4, 2014

Adventures in Low MMR (Part 1)


This is a series of articles on my adventures on a low MMR account. My main account is at 4200 rating from playing mostly support. This is an experiment to see what the experience is like climbing the ladder while supporting from a low MMR account (2200). Here's the rules I am playing by - Declare I'm supporting at the start, control map with wards throughout the game, buy the courier, don't take farm from any other heroes, obtain detection as necessary, item focus will be healing and defensive in nature. I do not play a solo lane of any kind.

I will attempt to point out some mistakes made by both teams in the post-game analysis, which may be helpful for players at that particular skill bracket, and also because it will be interesting to see the evolution of play skill from low to high skill.

Series 1

I score 5 straight wins, bringing me from 2265 to 2363. Overall there was only 1 game with an abandon.

Game 1

I won the first game as Crystal Maiden (duo safelane with Lycan) at 2265 rating

Warding: The enemy team did not buy a single observer ward.

Early Game: We lose our lane, as Lycan is unable to farm well against an Axe and Ursa. Eventually we decide it's more profitable for Lycan to go jungle and for me to leave and go roam other lanes.

Mistake 1: The enemy team had 4 heroes which don't come online until they get their Blink Daggers. None of them did, which is why their kill potential was very low. No more than 2 heroes should require Blink Dagger, so you can prioritize farm to them.

Mistake 2: They try to counter the Shadowblade on our Sniper with sentries. This is the wrong use-case. Sentries are used to prevent invisible heroes moving towards you, while dust is used to prevent invisible heroes from escaping. You would buy sentries to protect yourself against a Nyx or Shadow Fiend. You would buy dust to help you secure a kill on a Sniper.

Game 2

I won the second game as Venomancer (duo safelane with Spectre) at 2290 rating

Warding: The enemy CM kept the map well warded, but neglected detection for our Riki.

Early Game: Our bottom lane feeds the enemy Juggernaut 5 kills in the first 9 minutes, giving them a sizable gold and exp advantage. My lane was a loss, as the Spectre I was babysitting had constant lag issues for the first 10 minutes and was unable to farm. I was unable to even leave the lane to stack and pull. The enemy ended up with a 3000 gold and exp advantage going into the mid-game.

Mistake 1: My team picked 4 melee heroes. This is a bad idea. It is however an extraordinarily bad idea when the top lane has an Axe, mid lane has a Sniper and bot lane has a Juggernaut + CM combo. Naturally we lost all 3 lanes.

Mistake 2: Bloodseeker should not build a late Radiance, as the burn damage is a non-factor late game. If his farm won't allow him to get an early Radiance then he should go for cheaper items like a Basher + Orb of Venom rather than walking around with brown boots 30 minutes into the game.

Mistake 3: The enemy Juggernaut had a fantastic start, scoring 5 kills in the first 9 minutes and getting the double the GPM of any of our heroes. However, Juggernaut's power falls off dramatically in the late game. Teams that utilize Juggernaut need to be push oriented to turn the early kill advantage into a tower advantage, or run him as a secondary carry so they still have a primary carry that will win them the late game.

Game 3

I won the third game as Dazzle (duo safelane with Centaur) at 2315 rating

Warding: The enemy team did not buy a single observer ward.

Early Game: I secure first blood for my lane with the starting Double Damage rune, and get another kill a minute later, winning us the lane easily.

Game 4

I won the fourth game as Crystal Maiden (duo offlane with Tidehunter) at  2339 rating.

Warding: The enemy team did not buy a single observer ward.

Early Game: I don't know what to say, we got 6 kills in our lane in 5 minutes, we were facing a Wraith King, Bounty Hunter and a jungling Lycan who ended up dying when he dived our tower.

Game 5

I won the fifth game as Crystal Maiden (triple safelane with Bloodseeker and Earthshaker) at  2363 rating.

Warding: The enemy team did not buy a single observer ward.

Early Game: I help secure the first few kills of the game, first blood on bottom and then ganking their mid Invoker as he comes for the rune. We did not intend a trilane but the Slardar and Storm Spirit demanded solo lanes, it's ironic that those two heroes did not end up doing very much.

Wednesday, February 26, 2014

Let's Play (Part 9) Ember Spirit (Xin)


Every week or two I generally try to play a new hero to discover what it’s like playing it, and in this series of blogs I will make a post on my experiences with it.

Official Ember Spirit Artwork

Ember Spirit (Xin)

Ember Spirit (Xin) was based off the Fire Panda aspect of the Pandaren Brewmaster, Chen Stormstout. In the World of Warcraft, the Pandaren are a secretive race who long ago left Kalimdor when they saw the Night Elves' obsession with magic and the troubles it would bring. They only formally resumed contact with the Alliance and the Horde after the Sundering. In the Warcraft 3 storyline, a wandering Brewmaster - Chen Stormstout - meets Rexxar in the Barrens, and they both play a key role in the narrative of the Orcs landing in Kalimdor.

In DOTA2, the Ember Spirit is part of a trio of elemental spirits - Storm Spirit (Int), Earth Spirit (Str) and Ember Spirit (Agi). What they all have in common is very high mobility, which makes them hard to kill, and also gives them great reach, allowing them to catch up to enemy heroes even from tremendous distances. The Ember Spirit's playstyle revolves around using Sleight of Fist and Fire Remnant to zip around the battlefield, allowing him to completely dodge enemy spells and land Searing Chains (a 3 second immobilize) on fleeing enemies. In the early game, his Flame Guard makes him a dangerous opponent, as it burns nearby enemies and also protects him from magical damage: support heroes getting dived by an Ember Spirit can often do nothing but burn slowly to death.

At present time he has a 43% public win rate (from DOTABUFF) after receiving a series of nerfs since his introduction.


Xin's skill set is almost item independent - allowing him to contribute to teamfights with a bare minimum of items. This allows him to play almost any position in any lane. Unlike most jack of all trade heroes, his Sleight of Fist has incredible scaling with damage items in the late game.

Xin is also very difficult to pin down and kill. He contributes to teamfights from extreme range without putting himself in danger, and he can escape danger instantly by teleporting to a Fire Remnant.


He has poor stat gain and thus can be killed easily if initiated upon.

Landing his primary disable - the Sleight of Fist and Searing Chains combo - can be unreliable - more on this later.

Sleight of Fist + Searing Chains

This is going to be a much longer post than usual, because this hero has a lot of issues worth mentioning.

Searing Chains ensnares two random enemies within a 400 radius, so with careful positioning you can try to avoid wasting this skill on creeps.

Sleight of Fist has a 700 cast range, and will allow you to hit anything within 550 radius of the cast point, potentially allowing you to strike at targets 1250 away from you. Each target will be attacked in sequence, one target every 0.2 seconds, and finally you end up back where you started. You are invulnerable during the attack animation.

The length of the green line is the 700 cast range of Sleight of Fist. The blue circle shows the 550 radius AOE covered by it. The picture above was taken with the range finder active - it's disabled by default. The range finder shows a solid green line (like above) if it's within range, and a faded green line with a red portion showing how far you have to walk before it activates, but I personally don't use it - it feels like using intuition is superior to trying to pay attention to a rangefinder. All you have to know is that the blue circle can be a bit further away from your character, but not by much.

"Luke, you switched off your targeting computer! What's wrong?"

To enable the range finder, you can manually enter the console command dota_disable_range_finder 0 but you'll have to do that every game. To set it permanently, you need to edit the DOTA2 launch options to look like above. Setting -novid gets rid of the silly intro Valve animation (with the bald guy) and -console enables the console, they're not necessary for this but it's nice to have. Anything after the + sign gets entered into the console.

The obvious synergy is to use Searing Chains during the Sleight of Fist animation on targets that are far away from you. Say an enemy is 1000 distance away: you can cast Sleight of Fist, and then immediately - during the 0.2 second slashing animation - cast Searing Chains. The ability to reliably land a 3 second immobilize and at least 500 damage on an enemy 1000 distance away - for only 160 mana - is major part of what makes Xin so powerful, but also requires impeccable timing. This ability can even hit and immobilize enemies hidden in fog - as a bonus, the immobilize effect will cancel teleports, making escaping from Xin an almost fruitless exercise.

All of Xin's abilities have no casting animation. This is an important fact because this allows you to easily dodge spells and attacks using a well timed Sleight of Fist, or even Fire Remnant.

Dodging a spell with Sleight of Fist to score First Blood.

Dodging a spell with Fire Remnant to avoid death.

When Searing Chains and Sleight of Fist are maxed, Xin's playstyle revolves around hanging back at the edge of the battlefield, using the Sleight of Fist and Searing Chains combo every 8 seconds to deal heavy damage and proc the immobilize effect on his enemies. Only when the enemies are close to death or fleeing does Xin charge into the fight with Fire Remnant to finish them off.

Flame Guard

This is an extremely situational skill. Flame Guard burns enemies in a 400 radius and also absorbs magic damage directed at Xin - if the shield absorb gets used up, the burn effect disappears too.

At level 3, Xin can have two points of Flame Guard, which deals 40 damage of burn damage per second in a 400 radius, and absorbs 200 points of magic damage. Used correctly in conjunction with 1 point in Searing Chains will make you a strong contender for first blood, especially with allies who can synergize with your 2 second disable - but it all depends on the type of opponents you are facing. A pair of magic nukers can instantly dispel your shield, leaving you vulnerable and stunned / slowed in melee range, while pure physical damage users can ignore your shield and kill you faster than you kill them.

However, taking two points early is be detrimental to your mid game, as a lot of Xin's effectiveness only comes online once he maxes Sleight of Fist and Searing Chains, which he usually does at level 9 (4-4-0-1). Flame Guard is quite inefficient at 2 points - consumes a lot of mana for not much benefit - but maxing it delays your core skills even more. You either end up with 2 useless skill points at level 11, or you delay your full effectiveness until level 13. Generally if you can avoid it, don't take Flame Guard, unless you have a very favourable lane matchup for it.

As the burn range of Flame Guard is identical to Searing Chains, it can serve as a useful indicator as to whether the enemy is in range of Searing Chains - you can see the burning animation on them if they're taking damage.

Fire Remnant

Most of Xin's strategic playstyle revolves around careful pre-placement of Fire Remnants.

You store 1 Fire Remnant every 35 seconds, and placing it on the ground leaves it there for 45 seconds. You can click to dash to any Remnant - if there are multiple Remnants on the map, you will dash through each of the others before arriving at your destination. Each Remnant then explodes, dealing AOE damage.

Usage of this skill requires deliberate forethought.

1. Escape Mechanism - you always want to leave a Remnant in a safe spot you can instantly retreat to if you get ganked. However, it only lasts for 45 seconds, and it's easy to get caught out without one.

2. In Combat Mobility - you can use it during fights to reposition yourself, either to move away or towards the battle. The Fire Remnant itself can be targeted at 1500 distance, and it runs there at 2.5x of your current movespeed (typically, you have 360 speed with brown boots, which equates to 900 speed for your Fire Remnant) . You on the other hand, travel to it within 0.4 seconds, or at 1300 speed if that's faster. This means that if you cast a Fire Remnant and immediately dash towards it, you'll catch up to it and detonate it before it reaches its original destination, which can prove disastrous if you were trying to use it to get up / down a cliff. Also bear in mind that the Fire Remnant is affected by slows - if you are crippled by slow effects, escaping with a Fire Remnant will be difficult as it also travels slowly.

As an example, if you had 360 move speed, and cast the Fire Remnant at the max 1500 distance, it would take 1.7 seconds for it to get there. In order for you to actually arrive at the Fire Remnant at its correct range, you would need to give the Fire Remnant a 1.3 seconds head start before dashing towards it. (in other terms, the Fire Remnant needs to travel 1140 distance out of 1500 distance, or take 1.3 seconds out of 1.7 seconds move time). If you're under attack, you can use Sleight of Fist to make yourself invulnerable while waiting for the Fire Remnant to travel the required distance.

This is very similar to Puck's waiting time when using Illusory Orb to escape. Puck uses Phase Shift to avoid damage while waiting for the Orb to travel a sufficient distance.

3. Rune Control and Bottle Recharging - If played mid, Xin can make good use of the bottle. When he's on low HP / MP and his bottle is empty, he can place a Fire Remnant, teleport home, quickly regenerate to full and refill his bottle before dashing back to his Fire Remnant. This can allow him to rejoin an ongoing fight fully recharged - the downtime isn't much of an issue, considering the 8 second cycle window of his Sleight of Fist and Searing Chains combo. In this way he's a bit like Tinker, able to leave and rejoin an ongoing battle fully healed. In any case, ANY time you want to go back to the fountain to heal you should place a Fire Remnant in the field so you can quickly return after healing if you want. He can also control the runes by placing a Fire Remnant on the rune spot 30 seconds before it spawns. Fire Remnants are also useful for taking shortcuts across the map, especially so when jungling.

4. Burst Damage - if there's a particularly good opportunity you can level up Fire Remnants to use them as burst damage - as there's no casting animation, if your team manages to catch a few enemies in a group you can unleash all three remnants and detonate them for 600 points of AOE damage.

5. Defend against split push - Let's say your team is pushing their tower and wants to force a teamfight, the enemy sends one or two heroes to counter-push another tower in response - you can leave a Fire Remnant behind, teleport back to your tower to defend, then dash back if a teamfight starts. This gives Xin some pseudo global presence like Furion / Spectre / Morphling.

6. Provide Vision - the Fire Remnants have 400 vision range, so you can use them to scout or provide vision of incoming ganks.

Here's a quick clip showing the usage of Fire Remnant and the Sleight of Fist + Searing Chains combo, where the Ember Spirit easily catches up to a fleeing Naga Siren even though she uses her Song of the Siren to escape.


Battlefury is a standard item on Xin, allowing Sleight of Fist to cleave, multiplying the damage on targets close together. It has great synergy with teammates abilities that cause the enemy to clump up, for example Magnus and Dark Seer, or enemy abilities that generate lots of summon units or illusions.


Xin is easily picked off and killed by heroes with long range initiation, such as Storm Spirit.

Enemy team composition that is mainly ranged will be able to avoid grouping up and render his Sleight of Fist and Searing Chains less effective.

Item build

Bottle is a standout item for the early game, as Xin's regeneration is poor. A full bottle will restore 405 health and 210 mana. Compared to a Perseverance, which grants roughly 2 mana per second (if built as first item) - it will take 105 seconds to restore as much mana as a full bottle does, at 1/3 the cost. Xin makes good use of the bottle because of his ability to effectively scout and get runes with Fire Remnant (although dashing to the remnant uses up 150 mana itself, barely paying for itself) but more importantly because he can make more liberal use of teleport scrolls or walking home - he can leave a remnant in the lane before walking / teleporting home and immediately dash back out into the field. This gives far more opportunities to refill his bottle compared to other heroes. This also means that Boots of Travels is a viable first item for him - as Xin is relatively item independent, he can afford to buy this as a first item so he can heal and refill his bottle at will.

Battlefury is Xin's core DPS item, due to the way it interacts with Sleight of Fist - every target in the area gets hit by a single attack, which means that any one target might be in range of multiple cleaves, increasing damage taken exponentially. The item is also ideal because it provides raw damage (Xin doesn't need attack speed) and also provides much needed mana regeneration.

Note on cleave mechanics: notionally it's 35% of your raw attack damage, but because it's unreduced by armor or evasion it typically ends up as roughly 50% of your normal attack. The radius of the cleave is 250, so it ends up a rather large circle 500 distance in diameter - this is larger than the typical AOE stun radius used in the game (225 radius, for the AOE stuns that Lina, Leshrac and Sven have). So, larger than you might have imagined - try envisioning one circle of AOE damage like that for every target hit by your Sleight of Fist. Non hero units take half damage from Sleight of Fist and so the cleave damage is reduced accordingly.

There's several standard builds on Xin, depending on which orb effect you want to utilize.

Brown Boots into Drums of Endurance, then Battlefury - this is the high DPS build that is the most common. If you're feeling confident you can skip the Drums of Endurance, as it's just there to provide a bit of a HP and MP buffer for safety. If there's a ganker targeting you, you need to take countermeasures: if it's an invisible hero you need to carry a gem, if it's a long range initiator (blink LC or SS) then you need Linkens late game.

Brown Boots into Eye of Skadi, then Battlefury - this build is more early game ganking oriented, with a Point booster and very useful Orb of Venom early on. Applying the Skadi debuff (35% move and attack speed slow, goes through BKB, 5 seconds duration) on the entire enemy team will yield a large advantage in the mid-game fights, as well as giving you a very large health and mana pool. The Skadi debuff duration (5 seconds) nearly matches Sleight of Fist's cooldown (6 seconds) allowing you to kite enemies to death - this is a key feature of Xin's playstyle, as he can not stand up to and trade blows with the enemy carry. This is my preferred build by far.

Arcane Boots into Desolator - This is an alternative non-Battlefury build, where you build Arcane Boots to compensate for the lack of mana regen. This could be a good choice where the normal Battlefury cleave is unlikely to be useful and you prefer to apply the -7 armor effect to the whole enemy team. For example, an enemy team that has no summons or illusions, has mostly ranged heroes, your own team is mostly physical DPS and lacks a way to force the enemy to clump up.

Either stacking multiple Battlefuries and / or getting a Divine Rapier is the common "end game" plan for him.

Maelstrom is currently bugged (?) on him in that a second proc will not occur while the first is still chaining, limiting its viability on him. In any case, there is a lot of item-budget wasted on attack speed which makes it sub-optimal.

Shadowblade is possibly bugged (?) on him right now where if he activates Shadowblade's active, then casts Sleight of Fist within the 0.3 seconds fade time, all targets hit during Sleight of Fist receive the 150 damage bonus from the Shadowblade active. In addition, once Sleight of Fist is finished, you will remain stealthed, and upon breaking stealth you still get that 150 bonus damage attack. It may be viable as a second damage item after the first Battlefury, due to the mana costs involved, but I don't really recommend it.

Sunday, January 26, 2014

Competitive Roundup (Part 10)


There are so many tournaments going on at the same time that it can get overwhelming trying to keep up with them. In this series of blog posts, I will share matches which I have found notable and worth watching in the past week or two.

TA Off-lane


Tournament:  Starladder VIII (game 2)
Teams: Na'Vi versus Alliance
Caster: Beyond The Summit

I've been kind of slack with these, but I'm getting around to it! Today's focus is on the off-lane role played by Funnik. There was some speculation that the solo-offlane versus defensive tri-lane meta would fall out of favour after changes to lane creep equilibrium, but that hasn't happened. What has happened though, as a sum of all the changes, is that the roster of possible off-lane hero has expanded, which makes things a lot more interesting.

Rather unusual picks from Na'Vi - nice to see the casters confused and get the lane matchups wrong. We see the return of Invoker, which is always exciting, and a really interesting off-lane by Funnik on TA. Alliance had Wisp in their lineup, which means they will dominate the mid-game because of their strong gank potential, but lose out in terms of teamfight - their two supports (Tusk and Wisp) cannot influence a 5v5 teamfight as much as Na'Vi's Treant and Dazzle. Both critical teamfights were won by the Treant being able to blink in and Overgrowth.

TA being played off-lane is a threat not so much because of what she does - which isn't much, since TA as a ganker requires the gold and exp of a solo mid - but is a threat because it forces the opponent to divert resources into controlling her, lest she gets too much farm and snowballs out of control. Ultimately Na'Vi had 3 strong carry heroes (TA, Luna, Invoker) while Alliance only had one - the Alchemist. Alliance were able to score repeated ganks using the Wisp Relocate but that was not enough to suppress the farm of all of them at once. Knowing they were on a clock Alliance tried several times to break into the Na'Vi base but the Treant's Living Armor was able to repeatedly heal the T3 tower back up. At the end, even Funnik's off-lane TA was able to solo the Alchemist in the final fight and still have nearly 100% hp remaining.

Link to game start

Nyx Offlane

Tournament:  Starladder VIII (game 3)
Teams: Alliance versus Na'Vi
Caster: Beyond The Summit

Very typical picks from both teams. This game exemplifies the differences between Alliance and Na'Vi - Alliance tend to prioritize winning every lane, making sure all their heroes get good farm and experience. It's usually more typical for teams to sacrifice some lane compositions in order to have better teamfight and midgame control.

Funnik running Nyx offlane is an example of this. Nyx gets absolutely crushed in the off-lane - he has to face up against Naga, Jakiro and Gyro. The three of them are able to keep him away from the creeps and keep him severely underlevelled and underfarmed. On the other hand, Admiral Bulldog gets Timbersaw, which is able to get a much better start as he can farm the lane without fear even against 3 opponents.

However, the Nyx pick does pay off in the mid game, as the fear of ganks from Nyx and Storm Spirit inhibit Alliance's movement and farming to the point the Luna can roam the map uncontested and pick up a huge amount of farm. Funnik plays a very safe off-lane, rarely giving up kills, and then coming back strongly in the mid-game to change the tides of the game.

Link to game start

Bonus game: Singsing's Mirana

This isn't a competitive game but I had a great time watching it. Singsing is probably the most popular DOTA2 streamer right now, he plays for Speed Gaming. He plays a lot of pub games, and is well known for playing Mirana.

In this pub match he faces off against S4, the solo mid player for Alliance. S4 counter-picks his Mirana with Shadowfiend, but due to lack of support he gets killed over and over by Singsing.

It's actually shocking how easily Mirana scores kills when played aggressively. Singsing makes it look so easy. In particular the presence of mind he displays at 3.45 is impressive: just after executing a gank and retreating from diving the enemy tower, the enemy support CM rotates in (too late) to help: instead of running out, Singsing skills up a point in Arrow and has barely enough mana to fire it to score a kill on the CM.

Link to draft stage