No matches

All about stats is a monthly column that discusses trends in Dota 2 based on recent patches.

If this Dota 2 season can be described in a few words, it is one that has seen colossal changes in the meta. There hasn’t been one team that dominates or a few heroes that stand out and dominate all tournaments (although Gyrocopter could stake a claim for the hero ever present). The reason for that is, no prizes for guessing, the biweekly patch system. Granted that sometimes the changes do seem to come a bit too fast, but with so many tournament and so many games being played, it definitely helps to keep the games fresh and the teams on their toes.

From patch 7.13, IceFrog and Valve have devoted their time to make it a slow game and that made teams go for a whole new approach. Roaming supports have been all but eliminated from the game. 2-1-2 or in some cases 3-1-1 lanes are seen. The mid matchup, majority of the times, is the good old 1 v 1 matchup it should be. No Earth Spirit to roll in and harass you as you farm away. And I for one totally agree with what is being done. The past year saw a heavily roaming favored meta. Position 2 and position 4 were the most important positions as mid had the extra creep in 7.06 to get more XP and snowball. It also lead to shorter games. It’s time things were slowed down a bit and teams’ composure in the late stages of the game was tested. Here, we take a look at how the meta has shaped up from 7.13 to 7.15 culminating in what we saw at the most recent Major, MDL Changsha.


Most played heroes from 7.13 to 7.15

This era consists of the games from Epicenter XL, GESC Thailand and MDL Changsha along with a few qualifiers for ESL One Birmingham and the Supermajor. It’s no surprise Sand King is the most played hero in this time line. Well, he is the most played hero right from 7.07 till now with 1320 games. The next in line is Disruptor. Tusk, Gyrocopter and Death Prophet are pretty much in the same group. The group that dominated the meta post 7.07. Most of the names seen here aren’t a surprise. What is a surprise though, and a happy surprise at that, is that Terrorblade is not there in the list. Ther nerfs finally caught up to the overpowered demon and sent him back to the portion of hell he belongs in.

Most played heroes in the professional scene from 7.13 – 7.15

Another interetsing thing is that a lot of heroes seen here did not manage to do well in the tourneys and come the MDL, most heroes of the heroes on that list disappeared and were replaced by others. Well there are certain ones like Gyrocopter, Death Prophet, Dragon Knight and Tusk that are still there, but a lot of new faces can be seen as well. Clockwerk, Kunka and Jakiro and Bane, to name a few, have started making regular appearanced in games. A lot of this is attributed to the cnages made in 7.12 and 7.14. The nerfs also targeted Gyro, Death Prophet and Tusk, but those just seem to be heroes that the players seem to favor so much that no amount nerfs can banish them to the nether realms. The more surprsing one amongst them is Tusk. Roaming heroes have been nerfed to the ground and Tusk was very  much a position 4 roaming support. But even with those nerfs, teams have managed to pave a way to fit Tusk into drafts and with a win rate of 71.43 % (whattt?!) at MDL Changsha, he seems to be doing pretty well for himself! I am really glad to see Sand King not in there. That hero has been in the meta since before TI 7 and needs to be kept in the sand for a good year.

Most played heroes at MDL Changsha

The professional scene is a strage one though. Out of nowhere, Kunka has becomre a tier 1 mid. A hero which was mostly picked as a position 4 support and one the odd occasion as a position 2 is now a regular mid hero. It’s baffling, isn’t it? He didn’t even receive any buffs. A fair explanation could be that as we head into a slow meta with longer games, you need stronger heroes at the 50 and 60 minute marks. And a hero that can wipe out farmed enemy heroes with one swipe of his sword that late in the game surely has some credibility. I mean, look what Sccc and Maybe did at MDL Changsha. Also, usually at that stage in the game, Kunka is bound to have a decent chunk of farm thanks to his Tidebringer. Another one is Clockwerk. Clockwerk was ignored to a great extent for a lot of the small patches. Now out of nowhere, it’s first pick Clockwerk days! To be fari, the fact that he can use Cogs to block offlane creeps and get the enemy creeps close to the offlane tower for the first wave does help his case.


Most Contested at MDL Changsha

Often the entire story is not portrayed through the most picks in the game. Some heroes are banned in nearly every game which does not give them the opportunity to shine. The image below is testimony to that. Io was banned in 58 of the 90 games played at MDL Changsha. In close second is Night Stalker with 57. Teams just do not want to deal with these heroes. Especially with the game turning into a lane focused game, a baby sitter for Io can make a huge difference for the core. Or put Io in the offlane with a Doom or a Beastmaster and they can harass the living daylights out of the enemy safelaner. Of the 19 games Io has played, the Wisp has a 73.68 % win rate, which makes it surprising that he got a buff recently instead of a nerf. Phantom Lancer and Slark are two other who aren’t picked a lot because they are banned in most games where they can’t be countered. It feels good to see Slark there after a long time in Dark Reef. Slark isn’t a Gyrocopter who comes online really early. He needs some farm to ensure survivability in fights alonf with the Shadow Blade for initiation. The long games allow him to not be involved in fights in the early stage and farm instead. Also, the added base regen he received coupled with Mangoes gives him a lot of sustain in lane. Teams are warming up to the hero and there will be more of him in the tournaments to come.

Most contested heroes at MDL Changsha

A stat that stands out is Leshrac’s win rate. 14.29 % in 21 games. That 3 games won in the 21 the Disco Pony was picked. This when he was dominating tournaments before MDL Changsha as either a support or mid. It will be interesting to see if teams still opt to go for Lesh.


Game Durations

I keep talking about game durations increasing due to recent changes, but has it really happened?! The images below show the distribution of games based on time. The first comparison is from patches 7.08 to 7.12 and the patches 7.13 to 7.15. There is a clear decrease in the percentage of games in the 0-20 minute duration categories. If the 90 games of MDL Changsha are analyzed separately, the long game theory is further bolstered as we see more of a migration to the right side of the bar graph.

Making the barracks harder to kill, decreasing tower bounty, making it easier to buy back for the losing team and removing the buyback penalty, easing up a bit on the mid game death timers and changing bounty rune timings from every 2 minutes to every 5 minutes are the prominent changes responsible for this. IceFrog and Valve have received a lot of Slack for this, it’s just a way to change up things. It has been a relatively fast meta for a long time and some long drawn games are just the breath of fresh air we need. Thanks to that, we are beginning to see Spectres, core Naga Sirens and Slarks among the ranks. It’s not that these heroes are picked every game, but an odd pick here and there makes things exciting.

Distribution of games based on time based on time for patches 7.08-7.12 and 7.13-7.15



Distribution of games based on time based on time for MDL Changsha

There are only two tournaments to go before the decision for the direct invites to TI 8 are made. ESL One Birmingham will be played on 7.15 while the Supermajor will be played on 7.16. The next patch, 7.16, is a hero change patch. The question is, after two big hero change patches, how big will IceFrog go this time? Will it be just slight balance changes or will the Supermajor have a new meta of its own?!


Cover image credits: Shannon Lau (


Related Articles:

All About Stats: Post 7.08 Meta Analysis



Do you like the current meta?

Share on FacebookShare on Twitter