Shanghai Major Group Stage Stats Recap
With the group stages concluded it is now time to have a closer look at the meta during the group stage of the Shanghai Major. Compared to the previous tournaments, the group stages at the Shanghai Major were more diverse, with only 23 heroes neither picked or banned. For Frankfurt Major group stage the same number was 29. The “core” pool of heroes has also slightly decreased—there are 14 heroes which were picked or banned in >50% of the matches, compared to 15 in the Frankfurt Major group stage. While seemingly insignificant, coupled with a higher number of persistent situational picks (10+ games), the patch pick distribution becomes substantially more uniform, making games more interesting for the viewers. Overall, despite questionable production quality, the tournament has been really enjoyable with good and varied professional Dota.
Team Draft Diversity
Despite a significantly more diverse environment, there are still teams which tend to cling to their prepared strategies or strongly adhere to the overall meta. Interestingly, this statistic has some correlation with the results teams have shown during the group stages:
With the exception of EG and [A]lliance successful teams tend to have lower versatility rating. This can have both positive and negative consequences on the team’s performance in the main event: on one hand, lower versatility can be attributed to the team’s ability to dictate the flow of the game starting in the drafting stage. Teams are comfortable with the heroes they get to play and excel on them. On the other hand, it can be an indicator of inflexibility.
The high versatility score of teams in the lower bracket is likely due to them being forced to adjust and experiment in their games. They are less likely to beat the opponent in a conventional manner. It is especially clear in case of Archon and Complexity. They are strong teams, but they are very much outclassed by their opponents in most aspects of the game and are more likely to be forced into drafting unusual, high risk-high reward lineups.
Of course there are many other factors which contribute to the drafting preferences of the team, but in general it is possible to speculate that both EG and [A] are slightly more prepared for the tournament, with larger successful hero pools. It is also worth noting that neither of these teams has been in significant danger of being forced into the lower bracket, hence they were less likely to use their “last-minute” strategies.
Heroes Statistics
Unsurprisingly, 
Earth Spirit has been the most contested hero throughout the group stage. Despite several nerfs he remains a top priority for most teams.
The hero has received a nerf to 
Aghanim's Scepter upgrade in the aftermath of MDL, where the hero had a ridiculous win rate of 82.76%. This has made the hero substantially weaker as a 3rd/4th position, since the return on investment became much lower. This nerf, coupled with experience the players have received playing in the MarsTV tournament have resulted in hero only having 50% win rate. It is important to note, however, that the hero has been only played 8 times and by teams which might have not been able to unlock the full potential of the hero.
The two junglers and roamers 
Chen and 
Enchantress have also been valued highly. Both of these heroes boast a respectable win rate across a significant amount of games. Teams were more likely to let Enchantress through, whereas Chen was usually banned. It seems somewhat surprising though, considering the fact that Enchantress scales better and is more versatile than Chen. We may see a changing trend in the playoffs with more bans towards Enchantress.
Lone Druid has had a very good showing of 71.43% win rate across 7 games. There is a reason the hero has been a ban priority and this time around it is not only [A]lliance who are interested in the hero.
The best surprise pick of the tournament has to go to 
Slark, who won 5 games out of the 5 he was played in. We have previously discussed how the hero is an excellent fit for the meta and it is surprising the hero didn’t get more attention. A weak laning stage and early-game insignificance might be why teams have been hesitant to pick him so far.
Tidehunter and 
Lion have also been good situational picks. The former has won 7 out of 8 games he was played in, whereas the latter can boast a 75% win rate across 12 games.
[missing hero: outworld-devourer], 
Sven and 
Disruptor are the most consistently well performing heroes of the tournament. With a decent number of games played, they all managed to maintain a win rate of over 60%. Interestingly, Disruptor has been rarely contested. Given 
Io’s rather high popularity in the groupstage,, it is quite surprising that teams do not ban Disruptor, a natural counter to Io, in the second phase.
Night Stalker, 
Morphling and 
Ember Spirit are the most consistently underperforming heroes. They have been played in more than ~15% of the games, but their win rates are below 30%. Morphling was expected to make a comeback in this tournament, but it seems the hero has been greatly overvalued or the players are yet to recall how to effectively play him.
Phoenix, 
Batrider and 
Vengeful Spirit are also on the losing side. Their showing wasn’t as poor as with the previous trio, but it still leaves a lot to desire for.
Batrider has been underperforming for the second tournament in a row, potentially indicating that it might be too early to play the hero and he still needs some buffs or that teams need to play around him better..
Phoenix, it seems, was supposed to be the Chinese trump card, but the results on the hero have been mediocre at best. It might have been the fact that the Chinese teams, who historically train among themselves, have converged on Phoenix as a good meta hero in the Chinese scene, but it has failed to deliver against international contenders.
Finally, the most picked hero of the tournament— Vengeful Spirit has managed to win only 41% of her games. Being jack-of-all-trades type of hero, she was definitely outperformed by more specialised supports.
Players
Loda, MVP.QO and TSpirit.Iceberg lead the charts for most average kills. While the first two players are unsurprising, given the performance of their teams as a whole, the latter statistic is extremely impressive.
Team Spirit has played a total of 7 games, with only 4 victories and despite that Iceberg has managed to accumulate an average of 9.14 kills per game—only marginally lower than the first two spots. From the looks of it, this young mid player has a very bright future—kill statistics are obviously not representative of the skill of a player, but it certainly points towards his worth for the team, especially considering their success with him.
The KDA top 3 spots are dominated by the Korean team. QO, MP and FoReV have KDAs of 16.25, 15.75 and 8.43 respectively. The first two players also have the lowest average deaths per match—1. This statistic does not only give hope for all the Korean fans, but should also alarm the other regions. The Korean overlords have awoken. Perhaps the most impressive part of this statistic is the fact that MVP.Phoenix has played two very strong teams—Team Secret and EHOME. These teams might have been out of shape recently, but it certainly doesn’t take away from MVP’s astonishing performance.
Miracle- is once again dominating the GPM/XPM chart with an almost 50 points difference to the next contenders—Matumbaman and Loda for GPM and XPM respectively. This consistency as well as high overall team win rate make for a solid case that there is place for Pub Stars in the professional scene and personal skill of each individual player matters a lot.
Finally, it is worth noting that the EHOME’s offlane player old eLeVen has one of the lowest KDAs across all matches. The reason why this had to be brought up, is that the player is generally in top 5 for KDA statistics of offlane players in big tournaments and this result is very uncharacteristic, as well as an overall result of the team. In case you have missed the news, old eLeVen had to undergo a tooth surgery in-between matches and is currently recovering from it. For the sake of EHOME’s fans, one can only hope that he will recover on time to boost team‘s performance - they will certainly need it for their upcoming bo1 challenges.
Final Thoughts
As discussed in our previous blog post the results of the group stage have been incredibly surprising—the rise of the Phoenix, Fnatic playing excellent Dota and a single Chinese team in the upper bracket are things which were rather hard to predict. That said, in terms of hero statistics, most of the things went in line with what was expected.
The patch is more versatile and the viable pool of heroes is bigger and more spread-out. The top contested heroes have remained relatively unchanged since MDL, with a new gimmicky focus on Phoenix and Vengeful Spirit. The performance of the popular heroes has also been in line with what could be predicted. This could be indicative of two things—either this is the final form of the patch meta, adjusted for small in-patch tweaks, in which case we will see more of excellent Dota, or alternatively, the teams have saved some interesting strategies for the main event and we will see new heroes on the rise. Either way, the main event promises to be very entertaining.
Image Sources:
....................
My bugatti's gone because of phoenix
MVP played Secret and EHOME not CDEC
Aftermath ? Kappa
Im 5th
Disruptor is hardly a 'natural counter' to wisp. Relocate ganks is at most half the point of wisp, glimpse doesnt work against wisp making another carry hard to kite and hard to kill.
@kokoroko how is disruptor not a natural counter if u make his relocate ganks useless (and that is as u mentioned it at least half the point of wisp)
navi will win this major,im most certain of it
почему он во сне улыбается?!
Iceberg is a great player but not the top even with those statistics. Obviously Miracle is alpha here since gpm and xpm matters the most on your carry. As an anti-mage you can have average KDA but still you can do much more than kill some supports early game...As Puppey said..dota is about money(gold and don´t forget the leves=exps).
MVP Phoenix