What Can We Learn from the EPICENTER?
The EPICENTER Major was immediately followed up by a small balance patch that will change the meta slightly, however none of the changes were too drastic. We will closely monitor new developments during the qualifiers for the International, but for now will concentrate on what lessons we can learn from the EPICENTER.
Centaur Warrunner emerges as a top tier offlaner
Between Mars, Sand King and Batrider, it seemed there was no place in the meta for another offlaner, but Centaur proved us wrong. The hero can boast almost 63% win rate across 35 games—one of the most impressive stats from the tournament.
Maxed out 
Retaliate allows the hero to deal a lot of damage to both heroes and towers, even with purely defensive early-game items. This alone already makes him a pretty decent option, since most offlaners are generally limited in their pushing capabilities.
More important, however, is the 
Stampede. This ability became a lot better, once the meta shifted towards slower cores that rely on Blink Dagger or drawn-out fights. A well-timed Stampede can ensure low casualties on your side once the enemy Sven uses God’s Strength or Death Prophet uses Exorcism.
Similarly, it can be used aggressively, to quickly initiate and clean up a teamfight. Given relatively low cooldown, when compared to ultimate like Ravage, Black Hole or Chaotic Offering, it allows the team with Centaur to set up tempo and generally have a macro advantage in midgame stalemate scenarios.
While the hero did receive some nerfs in 7.22d, his main selling point remained relatively unscathed and he still has one of the strongest late-game Aghanim’s as well.
Juggernaut is setting the tempo
On several occasions we have pointed out that the meta has slowed down, but patch being as good as it is, it meant that attempts to play tempo lineups were inevitable. They were also quite successful as well.
Juggernaut is one of the most reliable position one tempo cores in the game. The hero has great early game damage, decent scaling potential and the 
Healing Ward, which is one of the most powerful tools for sieging.
Through this incredible sustain for his whole team, Juggernaut can get a decent early advantage—given a good enough start he can go highground as early as minute 18 and apply pressure to potentially much slower opponents, that still require several minutes of farm to come online.
It doesn’t work amazingly well if the rest of the team doesn’t support the idea of an early push, but with support heroes like Chen, Shadow Shaman and to a lesser extent, Warlock, most teams have one way or another to deal tower damage.
Omniknight is slowly turning into a meta staple
Virtus.Pro in their interview stated that they see Omniknight as a position five support, but the majority of teams play him as a position three. Unlike most offlaners, Omniknight doesn’t really have catch, initiation tools or strong AoE disables. What he does have is incredible amounts of save, heal and status resistance.
Turns out [missing skill: omniknight-heavenly-grace-5264] is a lot better than many assumed initially. It might not be as good as Repel was back in the day, but it has a longer duration and extra benefits, apart from protection from disables.
By using it on one of your cores, you can create a monster that will have extra HP, regeneration and is guaranteed to have a window to press BKB in teamfights. And even if that fails, you still have 
Guardian Angel to rely on—in a meta where Sven is the most contested hero access to physical immunity is a massive bonus.
Given how Omniknight received no nerfs in 7.22d and how Sven remains the most contested hero, while position four supports fill the initiator role more and more, we are probably going to see more of Omniknight in both professional scene and in pubs.
Closing Thoughts
Identifying meta staples is getting more and more problematic as we approach the International. In their typical fashion, Valve went with highly experimental patches early in the season and then start refining the formula in the second half. So far it looks like they are doing an excellent job.
TI qualifiers are definitely going to give us more information on what we should expect from the biggest tournament of the year and whether some other heroes need some tweaking, but Sven aside meta looks incredibly healthy: no hero had a 80%+ contest rate and a total of 105 different heroes were picked with 110 contested.
1st Pog
2st lol
тут инфа прешла шо хрус не придет
Jugg is a beast, all of his skills were relevant trough out the game :
He has great early nukes + dispel + magi immunity skill, as of the other skills were scaling into late game
5th for the meta
6th....
7th...
8th....
this number competition is so stupid !
9th
Pls stop this number comments
Btw 10th
I'm not even sure about the purpose of this counting
11th
Time for a real comment
12th
Juggernaut is overrated and stupid af. Delete hero pls
14th join the admin yeah
15st
!!! yyay
16th smileyface :)
fucker carry 17th
centaur ftw
18th
KawaiiSocks: I'm not writing articles for you all to call numbers.
Where is Brook? We need constructive comments~
19th u scrubs
I'm dead inside
23th but why
But if you pick omniknight in sea server as a support your teammates will complain "gg no disable" coz they themselves don't want to pick any reliable stunners either.
time to beat the egg of instagram
24th
pogchimp
25th
:)
peenis
27rd. Yeah support Omni in sea server is dead.
28th POG
Also i am 28 years old PepeHands
Everything in SEA server is dead, and no matter what you pick someone will complain, so who cares.
Last?
Line ends here
no it end here