The numbers of ESWC
ESWC gave us a chance to witness some of the top DotA teams performing in a Dota2 tournament for the first time since The International. On that occasion, the game was brand new, and most teams were still experimenting with strategies and the limited hero pool. This time, they had more time to prepare, and thanks to that, it was an amazingly entertaining tournament. Today, we're going to talk about numbers. This feature is for all of you statistics fans: thanks to all the information we were able to gather from ESWC, we can bring you a full statistical analysis of that weekend. So brace yourself and enjoy the following paragraphs full of juicy information!
Statistics are a tricky thing. In the words of Aaron Levenstein: "Statistics are like bikinis. What they reveal is suggestive, but what they conceal is vital" and as such, they tend to deceive a lot. An insufficient amount of data may display conclusions that are completely inconsistent with reality, while on the other hand, an incomplete analysis, even with the right amount of information, can result in faulty interpretations. Take the case of the dreaded kill/death ratio, which lots of players love: by now, most people have realised that by itself, the statisic is an insufficient reflection on a player's skill level. This is for several reasons, such as the fact that support heroes generally tend to get lower K/D ratios, and that the (lack of?) skill of your opponent is also a key factor.
There's also another issue, the greatest sin of statistical analysis: Extrapolation. When you have information about a certain trend and you're missing a few non-vital parts, statistics are useful to help fill those gaps. That's called interpolating, and it's an invaluable tool. On the other hand, extrapolating takes the values of past events and tries to infer future instances of those events. The problem is that the perception of what happened alter the way we will react to next thing we encounter, and that human factor alters the specific things we are analyzing (the metagame, if you may). If a hero or team is entirely dominating the game, instead of believing it will keep on doing so, expect to see new strategies that aim towards stopping that specific hero or team composition.
Statistics, however, are an invaluable tool if you know what they actually tell. What information can we gather from them? For starters, we can detect tendencies, team playstyles and player form. With that information, you can get a heads up on what the metagame considers the most valuable heroes, detect weaknesses in team strategies, and force opposing players into a role or hero with which their performance is sub-par. On the other hand, for a team, it can help see what they're doing right or wrong, find the deficits in their game and adjust their playstyle to improve their game.
To wrap up this introduction, we have to understand that statistics are not the ultimate answer, but as you will see, it's a very useful parameter which we'll use to understand what happened in ESWC. It's worth noting that in order to make this feature, we took the following information from the ESWC matches: Match results and length, picks and bans (in order and which player used which hero), and finally, KDA for all heroes in all matches. Note that from the grand total of 25 matches played, there are three games (Next vs BX3, Virus vs M5 and SGC vs M5) where there's no ban information nor pick order information. That's because those matches weren't casted and the available replays don't show picks and goes into a All Pick screen instead. That's not fatal in itself, but take that into account when reading the following analysis.
|Top Tier Heroes |
|There are heroes that are so powerful that they're either banned in the first round of bans, or picked in the first pick. During ESWC, we had:|
Banned in the first round in at least 18 of the 25 matches, first picked in two, Furion is simply a very powerful hero with huge mobility, ganking and pushing power. On top of that, he can actively participate in every single teamfight. Most teams didn't want to see him in play at all, even banning him when having the first pick.
Banned in the first round in at least 11 matches, first picked in at least 5. A nightmare in the hands of Artstyle during The International, a fast, very tanky ganker and a huge asset to the team with its 8 second silence, he's arguably the strongest ganker in the game that can stay relevant late game.
Banned in the first round in at least 12 matches, first picked in at least 5. Considerably one of the best late game carries, while not as powerful as spectre, he can dish out lots of damage early thanks to mana burn and insane base attack speed and get away from most tough spots thanks to his blink.
Banned in the first round in at least 13 matches, first picked in at least 4. Puppey showed us in the International how powerful this hero can be. Huge jungler and invaluable in gank-into-push strategies, its creeps and their auras mean that a successful gank, will also mean a free tower.
Banned in the first round in at least 12 matches and first picked in at least 5. Two words: Dark Ritual. A skill that, by itself, can set back a hero's leveling by 2 levels by level 6 and give him infinite mana to harass makes him an insane solo hero. Add total item independence and a skill that can win teamfights by itself and you have a hero you don't want to face.
|Most loved heroes|
|Some heroes are so useful and versatile that they find a place in every single game, however they are not individually powerful so that they deserve a ban. Some will argue that heroes that fall into this category are the best balanced ones. Whether you agree with this or not, one thing can be said: Their skill set makes them suitable in every team composition and that's why they're universally picked.|
The most picked hero with 19 appearances. Can either solo or participate in a trilane. A long stun, long range farming skill and a solid escape mechanism more than makes up for a lackluster ultimate. Never a game winner, but useful at every aspect of the match.
Picked in 17 matches and the same recipe: Strong solo and trilaner, long stun, farming skill and 2 escape mechanisms that can also help initiate. Additionally, it can shift into a carry pretty decently providing there's no hard carry on the other side.
Picked in 16 matches. A tanky roamer/trilaner with a decent slow, but let's face it: the only reason Tidehunter is picked is because it has the best initiation tool in the game which doesn't actually require a dagger to be effective.
Picked in 15 matches, and with good reason. Great solo and trilaner, global presence and huge spell and attack range for a very squishy hero. Teams that picked Ancient Apparition found great success during this tournament.
Picked in 14 matches: The epitome of support heroes had a "curse" during the international, where whomever would pick her would surely lose. The curse became more of a trend in the ESWC with a hideous 20% win rate. That didn't stop teams from picking her though as her utility is great.
|Most effective heroes|
|The reason why heroes are picked and banned is because players believe those heroes will enhance their chances at winning the game. If that was the case, the most picked heroes would also have the best win rates. However, that is not always the case. There are always a few odd picks that have great success, perhaps because they were the favorite pick of the champion and of no one else, sometimes because it embodies a strategy that greatly counters the main playstyles of the tournament. |
Pudge (4-0 100%)
A favorite of 2 of the most succesful teams in the game: 4GL and Na`vi. Pudge was used by grizine and Dendi with unparalleled success. For a long time considered a "joke pick", Pudge's damage output in the early game, paired with its ability to deal enormous amount of damage early game has morphed him from just a fan favorite to a force to be reckoned with. A unique initiation skill, and amazing tankyness mid and late game makes Pudge a scary opponent if played well, and Dendi and grizine seem to have found the recipe for success.
Faceless Void (5-0 100%)
Another favorite of 4GL, also used by Na'vi in the amazing monkey vs. Na`vi game, and also in one game by Moscow 5, Faceless Void won the battle of the carries. While on paper he's not as effective as Antimage or Spectre, but he showed up in teamfight oriented teams with great results. Maybe therein lies his success: While not as powerful as the other two guys, he's a hero that plays a role in a specific team composition, and in THAT composition, he excels, while Antimage or Spectre are picked regardless of the rest of the team, and find themselves often in hit-and-miss situations.
Ancient Apparition (12-3 80%)
One could argue that Pudge and Void had such success because of the teams who played them. In the case of Ancient Apparition who was played by 4GL, Na`vi, M5 and Virus Gaming, he found non-stop victory with the first 3. The reasons were portrayed before, not much left to say about this very versatile hero, other than: there's a reason why since coming out, he's the only hero that only received nerfs and not a single buff.
Lich (7-3 70%)
Finally a hero that worked quite well for mostly everyone. 4GL, monkey, Orange and BX3 won when using him. The only one of the most banned heroes that really made a difference when picked. Take note, guys!
|Least effective heroes|
| On the other hand, a few heroes suffered from terrible win rates. Some, because they were favorite picks of the losing teams, but some others, more universally picked, generally failed to work adequately. Here are some examples:|
Vengeful Spirit: (3-11, 21.43%)
The curse of the VS carries on from The International. Theories as to why this happens are all over the place. Some people say that the limited hero pool limits her true potential, others think that Nether Swap is a spell unlike no other in the fact that it can automatically lose a teamfight if used wrong.
Puck: (2-5, 28.57%)
The master escapist didn't fare that well in this tournament. No team could use him to its fullest potential, while the other classic lane solo heroes (WR, Clock, Mirana) performed above 45% each
Earthshaker: (3-5, 37.50%)
Now this supports a good theory. Even if it's just 1 match away from 50% win rate, if we pair this to Vengeful Spirit's win rate, probably the two best roamers the game has at the moment, we could think that teams with exclusive roamers might be the problem. Maybe the metagame discourages them? Food for thought.
Weaver: (3-5, 37.50%)
A powerhouse during The International, it's sad to see Weaver as the most underperforming carry of the lot during ESWC. The numbers aren't that terrible though and we can still believe it's a fluke. Weaver will see better tournaments than this one for sure.
|Sometimes a hero by itself doesn't win matches, but combined with one or two others, it is highly effective. In a 25-game tournament with more than 50 heroes in the pool it's virtually impossible to make any conclusions, so in this case we will just point out a few winning combinations as if to say "they worked". But we abstain from forging any conclusions on that behalf. In the same vein, we will focus on heroes that had an average individual win-rate, as naturally, the combination of any hero with a high percentage hero will have a high percentage itself.|
(50.00% win rate) + (50% win rate) = 2-0 (100% combined win rate)
(50.00% win rate) + (43.75% win rate) = 2-0 (100% combined win rate)
(50.00% win rate) + (37.50% win rate)= 2-0 (100% combined win rate)
(47.37% win rate) + (60.00% win rate) = 3-0 (100% combined win rate)
(47.37% win rate) + (50.00% win rate) = 3-0 (100% combined win rate)
(47.06% win rate) + (62.50% win rate) = 2-0 (100% combined win rate)
(47.06% win rate) + (66.67% win rate) = 2-0 (100% combined win rate)
|Certain heroes only see the light of day in a few special situations. Heroes that have a VERY specific role, or counter a specific strategy are often picked as a 5th surprise pick, giving the opposing team little or no chance to counter them. However, either of those heroes are easily countered themselves, and are not useful unless they are with/against a very particular lineup and/or simply don't warrant an earlier pick because other heroes are simply better at what they do. Those heroes are the situationals, and there were a few of those in this tournament:|
Picked at least 5 times as a last pick out of the 8 times it was picked. A good solo hero that can farm against trilanes by its own thanks to his rockets and escape thanks to his cogs and battery. However, Windrunner, Mirana and Puck were heavily favored as solo heroes. Yes, it's a good ganker, level dependant, but it's just sub par compared to the others as it's a lackluster hero when late game comes.
Picked at least 3 times out of 5 as a last pick. Definitely not the best carry at the moment. There seem to be just two reasons why it's picked: There's no other carry available in the pool already or because of its most disruptive silence spell. In both cases, it comes down to getting to that final pick in order to see if any of those conditions apply.
Picked 3 times at the last pick, and other 2 as the fourth pick. Never picked before that. The reason why Rhasta is picked last is not because of its possible counters (there aren't many things that SPECIFICALLY counter him, he's just another squishy hero that has difficulties running away from more than 1 foe. The thing with Shadow Shaman is that he tells a lot about the team strategy. He's a pusher and you can build a pushing team around him easily without any other telltale signs.
|The winning team.|
|Now, let's take a look at Na`vi. They were the best team from start to end, finishing with a 6-1 record. What was the key to their victory? We'll go through the players, one by one. We are adding a new stat too: Team Participation which is (Kills + Assists) / Total team kills. This gives us an idea about how active that hero was in the game. Heroes with global presence and gankers will have naturally high participation, while other heroes' participation will only depend on the player teamfight awareness and positioning|
|The players: |
LighTofHeaven played three different heroes with different roles. He was a positive influence with those 3 heroes, but we can clearly see that Windrunner is the odd one out here. With a participation of only 24% of kills in their victories, it shows that the team didn't need him much, and that's not necessarily a bad thing, but also that he didn't take a predominant role in those games. A bit of a warning, but then again, WR seems to be a hero with which lots of people had problems with being in teamfights, during the tourney. The high number of assists and participation of Furion on the other hand speaks of his awareness when using that hero. He's the blue collar worker of the team and he gets things done.
Dendi was THE ganker by excellence, and it shows on his Storm Spirit, BM and Pudge. When using them he was all over the place. Granted, he couldn't kill as much with his BM in the monkey game, but he was active and all around the place. And how many pro players can average 15 kills and 13 assists with a Pudge?
Puppey rocked the world during The International with his Chen, but they didn't let him use it this time. Chen was either banned or picked first by every single one of Na`vi's opponents. That didn't stop Puppey from performing great. With huge numbers for a support during their victories, he more than fulfilled his role during the tournament. And it's hard to dismiss his amazing Enigma during the Na'vi vs. Monkey match.
XBOCT, Na`vi's carry, showed that he isn't a one trick pony. Playing four different carries with very different playstyles, he was great all around. His death average with the Weaver could have been a bit better, but that's nitpicking, and his Faceless Void was crucial in outcarrying a 600gpm Antimage and he reminded us all that Drow is not a force to be ignored.
ARS-ART, the man in charge of replacing ArtStyle had all eyes on him, and he delivered. Opting to switch between ganker and support, he fared extremely well at both, playing a superb Vengeful Spirit against EHOME, a hero with which most people struggled, and showing that he can gank too with 2 great Night Stalkers and a Slardar. All in all a very good tournament for him.
|The Bans and picks:|
|Analyzing what Na`vi picked and banned, we can figure out two things: Their core (picked 5 times each) was trying to play Windrunner and Ancient Apparition. This is a really interesting move because both heroes can be played in very versatile ways. Na`vi knows that getting these 2 heroes on board gave them a lot of flexibility at the time of picking. On the other hand, they seemed to be intent on banning Lich (5 bans, banned by the other team on the other 2 games) and Tidehunter (4 bans). Both heroes with great teamfighting presence, and in the case of Lich, very dominant in the lane too. Seeing as Na`vi favored a teamfight oriented playstyle during this tournament, the bans make a lot of sense.|
|Bans and picks against Na`vi:|
|Not much can be said about what the other teams picked to play Na`vi. With the current metagame and small hero pool, the game seems more about picking from the 10-or-so pool of great heroes without much consideration about counter picking, so with 3 picks, Antimage and Tidehunter were the most picked heroes, but then again, both were fairly used picks, and the fact that Na'vi didn't favor any of them, made them available to the opposing team.|
Banning against Na'vi, was a different story however. Aside from the 4 Furion bans (In the other 3 games, Na`vi either picked him or banned him) which can be attributed towards the power of that hero and not Na`vi itself, there were 5 Chen bans against Na`vi. Almost half the total Chen bans in the whole tourney. The reason for that? Puppey's amazing play during The International, when Na`vi adopted a more push-oriented strategy. Surprisingly enough, Na`vi didn't pick him even when they could, as their strategy facing this tournament was completely different, albeit equally sucessful.
So that pretty much wraps up this feature about the tournament. If you want to make your own inferences, following this paragraph we attach the general tournament info. If you would like to see all the data, and/or a further analysis of upcoming tourneys, let us know!
General Tournament Info
Na'Vi's picks and bans
Picks and bans against Na'Vi