Wednesday, July 31, 2013

Blog on hold

Combination of TI3 coming up, being sick, and having a bunch of other things going on (training, etc) means I'll be putting this blog on hold. Happy watching TI3 everyone!

Friday, July 26, 2013

Probability of a 7 game win streak in DOTA2

Everyone has experienced win or loss streaks in DOTA2 that seem to go on forever. Is this luck, or skill, or just business as usual?

If you're feeling lucky, this is the hero for you 
(Valve Ogre Magi release art)

Probability Theory

Modelling the probability of a single streak is trivial. Most people seem to have a win / loss streak of about 7, and if each game was "well balanced" they'd have a 50% chance of winning - the chance of 7 wins in a row is 0.5^7 = 0.8%. Wow, so nearly everyone has experienced a 0.8% probability event on their account, that seems pretty unlikely.

Or is it?

Modelling the probability of a streak of length K occurring within a sample of N attempts is a lot harder. The chance of, say, a 4 win streak (probability 0.5^4=6.3%) in 20 games isn't simply 6.3% x 5... it's not even 1-(1-6.3%)^4. It's complicated because while each event is independent, what's being evaluated is not - we're evaluating each flip and checking if it's a successful streak, and this evaluation is dependent on the previous attempts. We need to delve into Markov Chains for this.

Fortunately, there is a simple formula that covers this, which involves K-Step Fibonacci Sequences. The regular Fibonacci Sequence we all learned about in school has K=2 where each successive integer is the sum of the two before it. For higher values of K, say 3, each integer is the sum of the 3 values before it, and so on.

The formula for determining the probability of a streak of length K occurring with a sample of N attempts is 1- (K Step Fibonacci Sequence of N+2) / (2^N). We can demonstrate this works on some trivial examples which we can hand calculate, for example, the probability of a 2 win streak occurring in 4 games. (skip this portion if you're not interested in the math and head on to the next...)

The denominator (2^N) is the number of total outcomes. (say A is win, B is lose)

4 Wins - AAAA
0 Wins - BBBB

The 2 step Fibonacci Sequence is 1 1 2 3 5 8 13 21

I've bolded the ones above which have a 2 win streak (8 outcomes), total outcomes are 16, so the chance of a 2 win streak in 4 games is 50%. According the formula, the result of the Fibonacci Sequence for (N+2) is 8 as seen above, the value of 2^4 is 16, so the answer is 1-(8)/(16) = 50% which agrees with our result.

We can test this formula for a more complicated question, what's the probability of a 3 win streak in 5 games? 

5 wins - AAAAA
0 wins - BBBBB

The 3 step Fibonacci Sequence is 1 1 2 4 7 13 24 44

I've bolded the ones above which have a 3 win streak (8 outcomes), total outcomes are 32, so the chance of a 3 win streak in 5 games is 25%. According to the formula, the result of the Fibonacci Sequence for (N+2) is 24 as seen above, the value of 2^5 is 32, so the answer is 1-(8)/(32) = 25% which agrees with our result.

So about this 7 win streak...

Say the average player has 400 matches under his belt, and is wondering if his 7 win / loss streak is unusual. This means K=7 and N=400.

7 Step Fibonacci Sequence = 1 1 2 4 8 16 32 64 127 253 504 104 2000 3984 ...

Woah, we need to write this out until N+2 = 402 integers. Luckily computers can do it for us. Excel gives me a value of 5.2668E+119 for the 402th number in the sequence - that's a huge number, it is 119 digits long. I don't even know if we've invented a name for it. 12 digits long is "Trillion". 18 digits long is "Quintillion".

The denominator is 2^400. This is 2.5822E+120.

Finishing up the formula, 1-(5.2668E+119)/(2.5822E+120) = 79%.

This means that after 400 games, where the average win rate is 50%, there's 79% chance you will encounter at least one 7 game win streak. This means that out of a sample of 5 players, each of them with 400 games, 4 of them on average will have a 7 game win streak.

What this means is that a 7 game win (or loss) streak is not luck, or skill - it's just business as usual.

If you get up to 800 games played, there's a 79% chance you'll have an 8 game streak.

I couldn't calculate the equivalent 9 game streak chance for higher number of games because Excel apparently can't do numbers larger than 308 digits long (who knew).

If you have Excel or an equivalent (Google Spreadsheets, Open Office, etc) you can set up a K-Step Fibonacci sequence quite easily and calculate the various scenarios yourself.

No! It was not luck, but skill!
(The Celestial Starlight skin, created by Oni and Zaphk)

Wednesday, July 24, 2013

Design (Part 7) Screwing your teammates


This will be a series of articles about game design, using the popular games DOTA and League of Legends as illustrative examples. The specific words, "Design is an Optimization" was coined by the design director of Riot. It means that to him, design is an iterative process of optimization.

Thus if some aspect of DOTA or LOL seems inferior or different to the other, it is because a conscious choice has been made by its designers – the trade-offs have been examined, the benefits judged more than the costs.

With friends like these, who needs enemies?

One of the biggest causes of frustration in team games is getting screwed by your own allies, either unintentionally (they were inexperienced) or intentionally (trolls!).

Of course, the fact that this can happen at all has to be designed into the game. Riot believes such mechanics should be removed at all costs, as they can cause frustration and toxic interaction between players. Where possibly they have been modified - for example, Blitzcrank's Rocket Grab can only grab enemies - it can't displace allies, while Pudge's Meat Hook can. In general, Riot is careful to ensure that your abilities can't negatively impact your own allies, though there are a few exceptions, for example, Anivia's Crystalize can potentially block allied movement.

Icefrog doesn't care, and has left many such avenues open to players in DOTA2...

Abilities in DOTA2 that can actively harm allies

Bloodseeker's ability Bloodrage is a buff that will silence the target, increase their physical damage output, and do damage to them over time, and can be cast on both allies and enemies. It's meant to be cast on friendly physical damage dealers who don't rely on spells, or enemy spellcasters for the silence effect and who don't benefit from the increased damage. As you can imagine there is a lot of potential for this ability to be used wrongly - getting a long duration silence from your ally is infuriating as a spellcaster, or even being killed from the damage.

Naga Siren's ability "Song of the Siren" will put enemies to sleep, but makes them immune to damage for the duration. This skill is meant as an initiation and positioning skill (allowing your team to get into position before the fight) or as an escape skill. Used at the wrong time however this skill can cancel the damage and stuns your team was trying to inflict on the enemy team, causing your team to lose the fight in the end.

Nature Prophet's ability "Sprout" is erects a ring of trees around the targeted area. It's usually meant to trap enemies, but it is often used on himself as a defensive measure. However, improper use of this skill can trap allies within it by accident, causing their doom.

Tiny's ability "Toss" will toss the nearest unit towards a target of his choice. "Nearest unit" could be anything - a creep, an allied hero, or an enemy hero. Sometimes he uses it just for the damage potential, but it can also be used for positioning - for example, Tiny could blink into the enemy team, and throw their carry back into your team so you can kill him. Or Tiny could throw an initiator like Axe or Slardar into the enemy team to wreck havoc. There is a potential for the wrong target to be chosen, however, as Tiny may accidentally throw someone he didn't intend to - throwing the enemy initiator into your team, for example, or throwing your support into the enemy team. Also, a popular method of trolling allies at the end of the game is to throw them towards the enemy team so they die.

Further discussion

There really isn't much to say on this topic.

The trade-off is pretty clear here - you can design some really interesting abilities and make them very powerful, but only if you balance them with drawbacks and the potential for failure - which also increases the skill cap of the game. For example, if the Naga Siren's Song of the Siren didn't stop damage and stuns from applying to enemies, how much weaker would it need to be to be balanced? If Prophet's Sprout didn't stop allied movement through it, how much weaker would it need to be for it to be balanced? By removing the negative effects of those abilities, the primary effect has to be made much weaker, diluting the power and impact of the skill.

DOTA2 is designed around the philosophy that skills should be powerful and game changing, with massive impact. This is balanced by longer cooldowns, higher mana costs, and bigger potential for failure when compared with skill in LOL. The drawback is that the minimum skill requirement to play the game at a decent level gets a lot higher, as it's not only possible to fail at using the skill, but to fail so badly that your allies are directly harmed by it as well. LOL's philosophy is that they want to lower the barrier to entry yet be able to keep the skill cap the same - it's just as hard for a pro player to land a Rocket Grab or a Meat Hook. LOL's argument is that you can have your cake and eat it too - making it easier for new players and casuals to get into the game, while not degrading the play experience for the top players.

That is true as far as gameplay is concerned - and as I've mentioned, LOL is very concerned with gameplay - but this does hurt the game in terms of giving it interesting and impactful skills. High risk, high reward and high lethality moves in the game frequently give rise to infamous DOTA2 teamfights like "The Play" (also see - multicam edition) that simply can not happen in LOL, which overall makes LOL less interesting to spectate - in my humble opinion, anyway.

Tuesday, July 23, 2013

Competitive Roundup (Part 7)


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.

Mouseports vs Na'Vi


Tournament:  D2L Finals (Game 1)
Teams: Mousesports versus Na'vi
Casters: Ayasee and Maelk

I really like Na'Vi's drafting strategy in this game - it's a strong mid-game team that can push and teamfight at all stages of the game. It's a flexible draft that could adapt to counter whatever Mousesports threw at them. Mous picked a Spectre as their hard carry, so Na'Vi went for early push and managed to destroy two lanes of barracks by 20 minutes. The Spectre managed to farm a fast 20 minute Radiance just as their second lane melee rax was going down - it was a fast Radiance, but simply not fast enough. One final push on the last lane at the 23 minute mark sealed the game.

There was no good quality Youtube upload of this that I could find, so you will have to watch the VOD from Twitch at this location. It's one single VOD of the entire finals lasting 4 hours and 16 minutes and there's a lot of filler and waiting in between - skip to the times I indicate below.

Draft Stage starts 0:51:30
Game start 1:13:48

Na'Vi vs Mouseports 


Tournament:  D2L Finals (Game 2)
Teams: Na'vi versus Mousesports
Casters: Ayasee and Maelk

Na'Vi allow Mous to pick up Puck again, Fata's signature hero, but they have a nasty surprise for them - they last pick Outworld Devourer.

Last week I highlighted a match where iG counterpicked Na'vi's mid Storm Spirit with Outworld Destroyer. Because of how badly OD beats other Int heroes mid, Na'vi had to move their Storm Spirit to the off-lane. This forced a very unfavourable laning situation for Na'vi and they lost the game badly.

This time Na'vi does the same thing to Mous and counterpicks Puck with the OD. Mous however, stands their ground, sending Puck mid anyway, and the result speak for themselves - by about 6 minutes and 15 seconds into the game, OD has 35 CS and 31 denies, while Puck has 10 CS and 0 denies. By 11 minutes in OD is nearly level 10 while Puck is still level 6.

Another interesting thing to note is how Funnik abandons top lane as Windrunner and just jungles the ancient camp with a Quelling Blade for the first few levels, it's reasonable common for him to do this. Mousesports do pick a KOTL as their last-pick in an effort to contain Na'vi's aggressive pushing but the game still ends in 27 minutes, with Na'vi racking up nearly 1000gpm of gold advantage per minute due to their superior hero picks and laning.

The VOD is just a continuation after Game 1 linked above, the times are listed below. 

Draft Stage starts 2:04:53
Game start 2:16:15

Mouseports vs Na'Vi


Tournament:  D2L Finals (Game 3)
Teams: Mousesports versus Na'vi
Casters: Ayasee and Maelk

Game three was a long one, with Mous playing very well throughout the game, their Anti-Mage carry having 21k net worth versus Na'vi's Gyro having 16k net worth. However, all it took was one perfect teamfight by Na'vi to turn the tide of battle completely - and with the momentum from that they assault Mous' base relentlessly and end the game right there.

This VOD captures that perfect teamfight till the end of the game, a sequence lasting only 3 minutes, completely reversing whatever happened in the first 40 minutes of the game. You could watch the whole thing as well (it's part of that long Twitch VOD linked above), but it's just a lot of back and forth until the end.

Friday, July 19, 2013

Let's Play (Part 6) Bounty Hunter


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 release artwork for the Bounty Hunter

Bounty Hunter (Gondar)

Gondar, the Bounty Hunter, is based off Akama, a broken Draenei who was once the Guardian of the Black Temple in the Outlands before it was taken by the Horde. In the Frozen Throne campaign, he and his clan assist Illidan Stormrage in defeating the Pit Lord Magtheridon. The Draenei are an ancient and enigmatic race, and Akama himself possesses fearsome stealth and combat abilities.

In DOTA2, the Bounty Hunter is a support-type assassin, stalking his enemies while in stealth in order to track enemy movement and set up ambushes for his allies. His kit includes stealth, a crippling first strike attack, and a shuriken that interrupts enemy actions. His ultimate, Track, marks the enemy, giving allies vision, enhanced movement speed and most importantly, additional gold if they manage to kill the target. Above all, the main reason teams run a Bounty Hunter is for the gold advantage he brings to the team.

The Bounty Hunter is a popular competitive pick, as he has the rare ability to solo the off-lane even against multiple opponents. This is because he is a hero that primarily needs levels (to get his ultimate as soon as possible) yet does not need gold (as he is relatively item independent). It would be rare for teams to run a Bounty Hunter in any other lane as they do not play to his strengths.


The gold advantage provided by Track is substantial - see further down for more information.

Extremely high "first strike" damage from Jinada allows him to solo weak supports with ease.


Poor scaling on abilities means he isn't able to fight straight up against enemy carries.

His abilities tend to be more useful in situations where your team is even with the enemy or has the early game advantage - as Track only awards bonus gold if your team is getting kills. If your team is on the back foot and not able to score kills on the enemy, the Bounty Hunter can be a dead weight.

Shuriken Toss vs Jinada

At Level 9 you will have 76 base damage, and ideally you would have Phase Boots (24 damage) and a Ring of Basilus (6 damage) for a total of 106 damage.

Typically at this point you will have 1 level in Track and 2 levels in Shadow Walk (so the duration exceeds the cooldown). Players typically then have 4 levels Shuriken and 1 level Jinada, or the other way around - 1 level Shuriken and 4 levels Jinada.

Jinada crit is 150% at level 1 (12 seconds cooldown), going up to 225% at level 4 (6 seconds cooldown). This means your initial crit does 159 damage at level 1, or 239 damage at level 4 - improvement of 80 damage.

Shuriken goes from 100 damage at level 1 to 325 damage at level 4 (10 seconds cooldown in both cases). This is an improvement of 225 damage.

At first glance it would seem that Shuriken is better. However, some points to consider:

You may get to use Jinada twice (with its 6 seconds cooldown). If you add the additional damage it provides over a regular autoattack (125%) you gain an addtional 133 damage. This brings the total damage improvement up to 212 damage, on par with Shuriken. Not only that, but you get the slow applied a second time.

Shuriken is useful as an interrupt (it mini stuns) and you lose the utility of this if you're using it as damage - with a level 4 Shuriken build, you would be counting on using Shuriken once, waiting 10 seconds, and then using it again to finish the enemy off - they could use a TP scroll to escape.

Also leaving Shuriken at level 1 and relying primarily on Jinada damage is not as taxing on the Bounty Hunter's mana pool, which is quite limited in the early game. The level 4 Shuriken costs substantially more.

Overall, I would say levelling Jinada first is the superior choice, especially as it also allows stronger lane harass against the enemy.

There are situations when levelling Shuriken first is better, in games where you do not want to be in melee range of the enemy - for example, when the enemies have deadly melee combatants like Tiny or Slardar, or if they have dangerous burst damage heroes in the early game like Lion. Then you would have to play more of a caster-type Bounty Hunter attacking from range with Shuriken and Track.

Levelling Shadow Walk first is also not a bad choice if you anticipate spending a lot of time in stealth as it not only improves the duration, but also improves the fade time (allowing you to more easily disjoint projectiles) and the damage increase is also substantial (going from 30 damage at level 1 to 120 damage at level 4 is a 90 damage improvement).


This amazing skill is the whole reason for the Bounty Hunter's existence. Not only does it give you vision of the enemy - allowing your team to initiate on them much easier, and also preventing them from initiating them on you - the +20% movement speed buff to all your allies in the fight makes such a big difference in attacking and retreating. It also prevents heroes from using stealth abilities to escape fights or initiate them. Last but not least, the gold advantage it provides is massive throughout the game.

Present at Fight Track Level 1 Track Level 2 Track Level 3
Just you 150 200 250
You + 1 Ally 200 300 400
You + 2 Allies 250 400 550
You + 3 Allies 300 500 700
You + 4 Allies 350 600 850

Team gold gained per track kill

This means that if your team somehow kills all 5 enemy heroes while remaining alive, your team would be awarded an extra 4250 gold. Over the course of the game, as long as you are doing your job with Track, your team could be trading "even" with the enemy (2 kills for 2 kills) yet have the gold advantage slowly tip in favour of your team due to the extra gold provided by Track. The total team score could be 20-20 yet your team could be up 8000 gold by late game. Ensuring Track is applied to as many heroes as possible should be your main priority in fights.


Track counters heroes with stealth like Rikimaru or Invoker, and also counters heroes with Shadow Blade as their core like Drow Ranger or Sniper.


Teams with hard disables or heroes with PBAOE (point blank AOE skills) like Centaur, Slardar or Tiny can make life very difficult for the Bounty Hunter. If he comes in close to use his Jinada he can be instantly killed if the enemy has detection.

Item build

As the Bounter Hunter is relatively item independent he buys movespeed and sustain items that will allow him to be more active in the map. You should always have a scroll of teleport available to use - if any fight breaks out you should immediately teleport to the location and join in. Even if all you accomplish is tracking the opponent before he dies, the bonus gold would be worth it in most cases. The items listed above is my preference for the Bounty Hunter core - Phase Boots, Urn of Shadows, Drums of Endurance, Vladmir's Offering. It has an easy start that will allow you to remain in the lane despite harass (ring of regeneration and sobi masks) that transitions into a relatively cheap core that gives you good regeneration and survivability.

Some players like to play the Bounty Hunter as a farming carry and build Phase Boots + Battlefury on him, for about the same cost as the above. The Battlefury is a highly efficient item in terms of the damage and regeneration value you get from it, and it will make your Jinada opener crit for a very large amount as well as enable very quick neutral farming. This isn't generally recommended though - Battlefury is normally bought on farming type heroes who spend the early and mid-game farming to procure a set of end-game items - this is counter to the entire concept of the Bounty Hunter, who is designed to be extremely active from level 6 onwards in order to secure his team a large gold advantage. Not to say this isn't possible, but it's certainly not the most optimal way to play him unless the team was somehow depending on you to be the #1 position carry.


As the off-lane solo you will be expected to buy the first set of wards to stop the enemy safelane neutral pulls. You should also pick up a stout shield, a tango, and maybe two branches. The single tango can be enough to sustain yourself until you buy the ring of regeneration (if you're following my build) - the Bounty Hunter has better than normal base health regeneration (0.75 per second, versus most heroes who have 0.25 per second).

I played a series of Very High Skill games on Bounty Hunter on the weekend, and as you can see in the map above, in one game as the Radiant off-laner I managed to ward their neutral pull with a somewhat unusual and useful ward spot I mentioned in a previous post about dual purpose wards. The enemy support was unable to deward my pull blocking ward, and so I was technically able to "win" my lane - reach level 6 in a reasonable amount of time without dying so I can begin ganking other lanes, while keeping their support under-levelled - he's unable to find a source of XP  through neutral pulls as he is trying to give the lane XP to his carry. Watch the screenshot I took above of the minimap where I ward the dire pull (my green ward) and he unsuccessfully places two sentries (hollow red circles) in an attempt to find it, wasting time and money doing so.

Wednesday, July 17, 2013

Design (Part 6) False Choices


This will be a series of articles about game design, using the popular games DOTA and League of Legends as illustrative examples. The specific words, "Design is an Optimization" was coined by the design director of Riot. It means that to him, design is an iterative process of optimization.

Thus if some aspect of DOTA or LOL seems inferior or different to the other, it is because a conscious choice has been made by its designers – the trade-offs have been examined, the benefits judged more than the costs.


Games are fundamentally about allowing the player to make choices. We present the player with a scenario, and present them a variety of options - and the idea is, the more skilled you are, the likelier you are to pick the better option and thus have an increased chance of winning. This can be seen at every level of the game - from what heroes you pick, to what lane you go to, what items you buy, how you level up your character, and most importantly, how you participate in battle.

Choices are interesting if all of them are potentially viable at some point. For example, in Scissors-Paper-Stone, you have a choice of 3 actions, one resulting in victory, one in draw, and one in defeat, but you don't know which one - it's up to you to predict your opponent's move. A "False Choice" is choice that results in an optimal (victory) or sub-optimal (defeat) outcome in all possible scenarios - that's not a choice at all, it's just an illusion of choice. You have to choose it / avoid it all the time, so you might as well not have been given that choice. Identifying a false choice is actually harder than it looks, as we'll see later!

Riot have talked several times about the issue of false choices in their game, and they've done a remarkably good job of removing them. DOTA2 however...

False Choices in DOTA2

Clockwerk's skill "Power Cogs" is an example of a false choice in DOTA2.

Clockwerk's Power Cogs will damage and knock-back enemies that touch it

If an enemy walks into the outside edge of the Power Cogs, he gets damaged, loses mana, and gets knocked backwards, back to where he started. So why would anyone choose to walk into the Power Cogs in the first place? It's a 100% lose situation: there is no reason why anyone would do this, ever.

Let's see how some similar environmental hazard skills work, and how they avoid presenting players with false choices.

Jakiro's Macropyre will damage enemies walking over it

Jakiro's Macropyre - creates a damaging field of fire on the ground. It's another environment effect that you generally want to avoid, but you could have a valid reason for walking through it. Maybe you really need to get to the other side and are willing to suffer the damage to do so. A choice is given to the player, walking through the fire is probably a bad choice but it can situationally be the right thing to do.

Disruptor's Kinetic Field prevents enemy movement through it

Disruptor's Kinetic Barrier - creates a wall and prevents enemies from walking through it. Enemies touching the barrier are just halted - they don't receive damage or knockback. No choice is given to player in this situation, they are simply unable to walk through the wall.

To be fair, Power Cog's isn't designed specifically only to create the false choice that it does. It's primarily used to trap people inside it with Clockwerk (so he can kill them with Battery Assault). You can tactically force the knockback and life/mana drain effect on enemies by using it just outside melee range - not so close that they'll end up trapped inside with you, but close enough that when the Power Cogs spawn it pushes them away. In some far fetched scenarios, you could have your teammates use skills to push enemies into your cogs and cause them to take damage.

This problem of unintended false choices can occur very easily with environmental type effects and it's a surprisingly difficult problem to solve - for example, Void's Chronosphere is intended to trap enemy players within it, however, anyone else wandering into the field while it's up will also get frozen. This is a false choice, as there is no reason why anyone would choose to walk into it, but it's there nonetheless. There's no real way to fix this problem in this case, as that's thematically how the skill works.

Further discussion - why it's bad, how to avoid

Why do we even care if there are false choices in games?

One is the minimalistic school of thought - having choices increases complexity. But false choices don't increase gameplay, so there's no reason to have them. Either take them out (like the Kinetic Field as described) or modify them so that some choice / gameplay is preserved (like Jakiro's Macropyre as described).

The other is that false choices are an unnecessarily negative form of feedback. Competitive multiplayer games need to give players feedback on how they're doing, so they can get better at playing them. Most of the time we want to use positive feedback - rewarding the player for using a skill correctly. For example, getting Mirana's Sacred Arrow to connect will reward the player with special dialogue and a powerful stun on the enemy. A false choice like walking into Power Cogs or Chronosphere is a very negative form of feedback - it's basically the game telling the player "Why did you do that, you idiot?" At least if it was a potentially valid choice - like being chased by enemies and trying to run through the Macropyre instead of around it - and the player died - the feedback the player would receive would be more like "Maybe this wasn't such a good idea in this specific situation, maybe try running around it and see if it works better." You're basically telling the player he made an error of judgement rather than that he was just being stupid.

Karthus' Wall of Pain slows enemies that walk through it

A League of Legends redesign of those skills will probably involve deactivating the active portions of those skills and turning them into an inert barrier after the initial cast - essentially still preventing people from entering the Power Cog or Chronosphere area but not punishing people who do so by accident. League of Legends has a couple of champions who have persistent barrier type skills, like Veigar or Karthus, which punish the player for walking through them, but never to the point where it becomes a false choice. Karthus has the skill called Wall of Pain that will slow you when you walk through it, but at least you've still moved some distance in that direction so you might want to do that anyway!

Removing false choices makes the game more forgiving, as there are less ways to punish the players for misplays - this is especially important for newer players. DOTA2 may simply not have this concern, as it's designed primarily with veteran players in mind - they might actually like the fact that having mechanics like this amplify the skill gap between better and worse players, giving the advantage to the more proficient player.

Monday, July 15, 2013

Competitive Roundup (Part 6)


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.

iG outpick Na'Vi


Tournament:  Alienware Cup (Lower Bracket Finals Game 3)
Teams: iG versus Na'vi
Casters: LD, 2GD, Luminous, Merlini, & Bruno

I really like the drafting strategy in this game - one of those games that are won in the drafting stage alone.

Na'vi get second pick and thus are able to pick up two heroes in a row, and they go for Naix + Storm Spirit, an aggressive combination that has served them very well in the past - the infamous Naix bomb, where Naix infests into the Storm Spirit and they go ganking. From the point Storm Spirit gets his Orchid no hero can solo farm safely - even traditionally safe farmers like Anti-Mage - due to the instant activation on the Orchid silence.

However, iG immediately counter-pick with Outworld Devourer. This creates a laning problem for Na'vi, where their Storm Spirit is unable to solo mid (Outworld Devourer is not only a strong mid but is especially strong against Intelligence heroes). Yet Storm Spirit needs a solo lane for a fast level 6 - without getting his ultimate he is very easily killed and can't contribute much to the game. iG further seal the deal by picking Treant, where his global defensive potential counteracts the Naix + Storm Spirit global ganking potential.

In the end, Na'vi have terrible lanes, with Storm trying to solo their safe-lane and getting repeatedly killed by the iG aggressive tri-lane, their Naix trying to solo mid against the Outworld Destroyer and losing, and they just get crushed overall and lose the game badly.

Funnik plays Bone Clinkz

Tournament: Alienware Cup (Finals Game 2)
Teams: versus Na'Vi
Casters: LD, 2GD, Luminous, Merlini, & Bruno

This game is particularly notable because there are no players in any well known competitive team that use Bone Clinkz except Funnik, and even Funnik has not used Bone Clinkz for months.

Na'vi run a free farming Clinkz (Funnik) while LGD have a free farming Spectre (Sylar). Obviously Spectre is a stronger lategame carry, however, once Clinkz finishes his Orchid (at which point Sylar is still farming his Sacred Relic), he starts terrorizing everyone on LGD, to the point where Sylar only dares to farm when his entire team is standing behind him to save him from being ganked by Clinkz.  Even that doesn't save him once Clinkz gets his BKB and Daedalus.

Clinkz is a hero with a very niche role in DOTA2 and is often misunderstood, leading to his poor win rate. He's not a carry (despite appearances), and does very poorly in teamfights. His skill set is deceptively simple but in reality requires a lot of skill to play properly. Clinkz is a hero that has a really bad early game and really bad late game - he only shines for a small window in the mid-game, and only if he managed to farm a fast Orchid. When he does shine though, he's an absolute terror, as this match shows.

Link to draft stage

Other thoughts

The Alienware Cup was really well run and had very enjoyable commentary from a variety of people with good insight into the game. I might review more games from it next week. Also there was the StarLadder Finals going on which I missed most of (there's only so much DOTA2 you can watch in a week) which I guess I will review next week. And right now the Defense is starting up again. Also I nearly didn't get to post this because I've been playing too much Endless Space lately but that's a topic for a different blog, methinks...

Thursday, July 11, 2013

Voice Acting in DOTA2 and LOL

Voice Acting in LOL

There's really not too much voice acting in LOL - a typical champion has about 17 different lines of dialogue, mostly related to moving and attacking. Some examples - Ahri, Nidalee, or Garen. There's no context related dialogue, or conditional dialogue (more on that later).

Voice Acting in DOTA2

There's a lot more voice acting in DOTA2. A typical hero has over 300 lines of voice acting, related to a variety of actions such as spawning, killing an opponent, getting first blood, using spells, dying, reviving, etc. There's even a whole host of lines related to actual chat (thanking someone, laughing), calling missing in your lane, expressing their disappointment at failing to connect with a skill shot or failing to kill a hero, and of course, special rival dialogue where your hero will taunt the enemy in a specific way related to their character after the kill.. Some examples - Juggernaut, Crystal Maiden, Bounty Hunter.

The main attraction, however, is the context specific conditional dialogue. The game is smart enough to detect certain unique events (whiffing with a skill, killing a specific rival, performing a special combo, buying a certain item) that is notable enough to warrant special dialogue that is quite often humorous in nature. Discovering these special lines of dialogue is always a fun and amusing exercise. I'm just going through the heroes that I know below, I'm listing down 24 special dialogue triggers that amuse me but there's obviously way more than this.

1. Juggernaut triggers town portal while using Blade Fury
(Bladefury renders Juggernaut magic immune for the duration, allowing him to escape with a Town Portal scroll safely as he cannot be stunned or interrupted)
  • Play I'm out of here.
  • Play Hate to hit and run.
  • Play I'll be back to finish this later.

2. Juggernaut fails to kill his enemy with Omnislash
(Omnislash is often sure kill, unless you mess it up and it hits too many targets or the enemy manages to dodge it)
  • Play What a waste!
  • Play Argh, I had such plans...

3. Juggernaut kills Lich
(Rival kill dialogue has a 25% of occuring, normally it just picks a random kill dialogue)
  • Play I'm the Juggernaut, Lich!

4. Axe executing opponent with Culling Blade
(His Culling Blade ultimate is an instant kill on opponents with low hp)
  • Play Culled!
  • Play Wheat before a scythe!
  • Play I said good day, sir!

5. Axe failing to kill with Culling Blade
(If the enemy HP is not low enough the instant kill effect will not happen)
  • Play Axe misjudged?
  • Play Axe misjudged?
  • Play You're tougher than Axe thought!

6. Axe using a blink dagger to catch up to the enemy, then killing him with Culling Blade
(... why not)
  • Play Not so fast!
  • Play You been blinked!
  • Play No escaping Axe!

7. Sven using Storm Hammer to stun a teleporting hero
(Sven can use his stun to prevent heroes from escaping by interrupting their teleport channel)
  • Play You look stunned.
  • Play Don't move.
  • Play Not so fast.
  • Play Stand and fight.
  • Play Stand your ground.

8. Windrunner missing her Shackleshot / Powershot
(Windrunner has two skill shots that many players find difficult to land)
  • Play Lucky you, slipped off the kebab.
  • Play Dang!
  • Play What the? I missed!
  • Play Unbelievable, I never miss!
  • Play What?

9. Witch Doctor about to kill with Maledict 
(The Witch Doctor has a strong damage over time spell called Maledict, if the game engine knows the enemy is about to die from it, he will say one of the following lines)
  • Play Wait for it…wait for it…
  • Play He's about to pop.
  • Play Dat make a big mess.

10. Ursa purchasing Vladmir's Offering
(With this item Ursa can easily solo Roshan, it's commonly made as soon as possible on him)
  • Play Behold Vladmir. In my claws, Roshan is undone.
  • Play Behold the offering. In my claws, Roshan is undone.
  • Play Roshan shall bow to me!
  • Play Roshan! I come to reclaim what you stole!

11. Troll Warlord meeting an ally
(The Troll Warlord's dialogue lines are parodies of bad mannered players in the game. (He is a troll, after all). Many of his lines involve insulting his allies. When he meets allies in the game, he might say one of the following.)
  • Play I'd be better off fighting alone.
  • Play Don't get in my way.
  • Play You've got to be kidding.
  • Play Why are my allies so weak and pathetic.
  • Play I can see I'll have to carry us to victory.
  • Play Try not to feed.
  • Play Just try not to feed.

12. Treant Protector using Tangoes
(Well, he is a Treant, and the tangoes are healing items that require you to destroy a tree to gain health... so naturally he feels quite bad about it)
  • Play Sorry, friend.
  • Play That was rude of me.
  • Play Apologies.
  • Play I'm sure you understand.

13. Timbersaw using Tangoes
(Timbersaw hates trees, so naturally when he uses tangoes he has something to say too)
  • Play Mm god I hate trees.
  • Play Another tree down.
  • Play Mmm splinters!
  • Play Delicious.
  • Play De-li-cious.
  • Play Won't need a toothpick.

14. Sniper killing a hero that used a bow
(Sniper is very proud of his gun, and naturally looks down upon people who use other weapons)
  • Play Oh, how cute!
  • Play Oh, feather bullets!
  • Play Heh, crummiest gun I've ever seen.
  • Play What's wrong, your string break?
  • Play Aw, isn't that quaint?
  • Play Get a gun, you'll live longer!
  • Play Guns are the future, my friend!
  • Play My gun is quick!

15. Sniper killing a hero that used a blade
(Sniper is very proud of his gun, and naturally looks down upon people who use other weapons)
  • Play Someone brought a knife to a gunfight!
  • Play You pull a knife, I pull a gun.
  • Play Someone brought a knife to a gunfight!
  • Play Inconceivable!

16. Slardar killing a hero after a long chase
(Slardar can sometimes score kills after an extended chase, due to his increased movement speed and ability to track his opponent's movements with Amplify Damage)
  • Play I cross the world to take you down.
  • Play I came a long way to see you die.
  • Play Tsee heeh heh he he heh heh.

17. Silencer stealing intelligence
(Enemy heroes dying near the Silencer have some of their Intelligence permanently stolen by him. That's bad enough, but it's insult to injury that Silencer taunts them as well)
  • Play Once a warrior, now a fool.
  • Play Your intelligence flows to me.
  • Play My intelligence grows.
  • Play Have you forgotten something?
  • Play Can you feel your grasp slipping?
  • Play Ha! How about a little brain damage?
  • Play Your mind is mine! Not like you were using it anyway.

18. Shadow Shaman trapping opponents with Serpent Wards
(Using Serpent Wards directly on an enemy hero can sometimes box them in completely, rendering them unable to move - it's something you always try to do, but don't always succeed at)
  • Play My serpents paralyze!
  • Play Entrapped.
  • Play You're not going anywhere.
  • Play Keep an eye on that one.
  • Play Hold that one fast!

19. Faceless Void using Chronosphere correctly
(The Faceless Void will say some bad-ass sounding lines if he manages to land Chronosphere on enemy heroes, trapping them in a bubble of stopped time)
  • Play Entombed by time.
  • Play Sealed your fate.
  • Play Like insects in amber.
  • Play Time's pendulum stops.

20. Faceless Void missing with Chronosphere
(Sometimes though.. he might misclick and cast Chronosphere on nothing...O GOD THESE LINES ARE TOO FUNNY especially the voice acting itself)
  • Play Nothing to see here.
  • Play How embarrassing.
  • Play Dear Gods, my mistake!
  • Play Forgive me my friends!
  • Play An eternity of embarrassment!
  • Play I can't bear to look.

21. Mirana hitting hero with Sacred Arrow
(This is a skillshot that is somewhat difficult to land)

22. Mirana hitting hero with Sacred Arrow for maximum stun
(If it travels at least 1500 range it deals the maximum 5 second stun)

23. Mirana accidentally hitting a creep with Sacred Arrow
(Creeps tend to get in the way a lot...)
  • Play Out of my way!
  • Play Who spoiled that shot?
  • Play Son of a...
  • Play Gaaah!

24. Mirana hitting Roshan with Sacred Arrow
(She often uses it against Roshan while the team is attacking, being an easy to land 5 second stun)
  • Play I could have hit Roshan with my eyes closed.
  • Play Poor Roshan.
  • Play An easy shot.