No Two Games of Dota are Ever the Same
No two games of Dota are ever the same. We have all heard this bold statement. Dota is a brilliant game with a very open meta, and the continuous stream of patches keeps the game fresh, but no two games ever the same? Surely, that can’t be right.
Dotabuff is in a unique position to test this claim: we looked for matches where the same five heroes were facing the same five heroes of the enemy team, and the findings were quite astonishing, to say the least.
After analyzing 236854 esports matches across 2321 league including qualifiers, group stages and LAN playoffs of amateur, semi-professional and professional leagues we found only 24 instances where the same hero lineups would be played twice. That’s 0.02% of our sample size. Moreover, there wasn’t a single instance where the same 5v5 match was played three times.
Certainly, the amount of potential different 5v5 matchups in Dota is massive; however, not all of them are going to be viable even in the amateur leagues. Different patches also favor different heroes, and the likelihood of seeing a particular set of heroes several times within a single tournament is quite high.
It certainly shows from what we managed to find: in all 24 cases the difference in time between when the matches were played is minimal, and most of the time involves matches from a single BO3 series. Moreover, in all cases, at least one team would remain the same, playing the same lineup of heroes.
The Little Guys
Most of our findings involve “the little guys” — teams from amateur and semi-professional leagues trying to make it big. Speaking from personal experience, these smaller teams generally have smaller hero pools and only a handful of strategies, which are also generally quite rigid.
In fact, in several instances the picks didn’t deviate even when the first pick priority was changed between games of a series: teams were content with trying the same thing once more, even if the first game didn’t go as planned in a bo3 match.
There is a whole discussion to be had on why most established professionals teams prefer the first pick, while many less experienced teams and pub players favor the second pick, but when you have trained a single lineup of 5 particular heroes the pick priority doesn’t seem to matter as much.
It is also hard to call any of these strategies “cheesy.” There are two Drow Ranger strategies, but none with Broodmothers, Techies or the likes. For the most part they are just solid, all around lineups, most likely selected because players feel comfortable playing them in the meta.
On the way to glory
There are two cases that were of a particular interest, since they involved high-profile teams of quite experienced players.
First one is a BO3 series between TNC Pro Team and TongFu.WanZhou in 2015. Back then TNC was only getting to the status of a SEA powerhouse and could easily be classified as a semi-professional team. The same can be applied to their opponents.
In fact, of the two rosters involved in this BO3, only two players currently play Dota at the highest level: Sam H for TNC and Boboka for TongFu.WanZhou. It was also the first team for a current member of iG, who is probably best known for his impeccable position 4 plays on Monkey King and Earth Spirit.
It also adheres to the rule mentioned in the previous section: these lineups are simply comfort heroes for the players involved. The first pick priority remained the same in both matches, as did the sides, however the pick and ban orders were quite different.
The TNC lineup for this game can be also classified as “cheesy,” since it involved a Drow, Clinkz and QoP for the core roles, with Dazzle and Earthshaker supports. It was a very easy-to-execute, pushing strat, with an incredible tempo and amazing pushing and split-pushing potential.
It is quite surprising to see a professional team respond with a late-game oriented draft without safe outpush and low early teamfight potential twice in a row: a Bounty Hunter and Spirit Breaker duo, backed up by Ancient Apparition and Gyrocopter sounds good on paper, but only for when the enemy team is split up and only after the supports get to level 6. Otherwise, the lineup doesn’t have enough skills to effectively take the fight or at least slow the enemy down enough for Storm Spirit to come online.
The definition of insanity
This series is a very unique occurrence in the Dota world and one that should be remembered: doing the same exact thing and expecting a different outcome is very unlikely to work, even if you have the confidence, spirit, and skill to try it again.
The second case of high-profile teams picking the same exact lineups is much more recent and involves some of the biggest names in the Dota scene. The DPL Season 4 match between NewBee and VGJ happened less than a year ago, and all of the players are active in the scene, even if the VGJ itself is no more.
It is also interesting because it was a rOtk vs. Faith drafting matchup, and both of these captains are respected in the scene for a reason: while in the TNC vs TongFu.WanZhou series mistakes could be attributed to inexperience and rigidness, these two captains have proved their worth time and time again, often leading new players in the scene to great success.
The team has been together for over four months at the time, so “playing it safe” and “testing the new squad” claims can also be put aside. All of the players were and still are top tier players, with a potential exception of Freeze, who nevertheless proved himself already, currently playing for VGJ.Thunder.
It is hard to pinpoint the exact mistake made by rOtk in the draft: NewBee pick and ban orders remained almost exactly the same, with an Ursa for Lifestealer last ban switch-up and the same can be said about VGJ draft. Soon after the event the team disbanded.
Conclusion
Dota is certainly a beautiful game. The openness of the meta, the amount of different interactions, the undeniable differences between seemingly similar heroes — all of this forces Dota players to constantly evolve, think of new strategies and experiment, otherwise they will fall off. It is especially true for the highest levels of Dota.
The game gives both teams unbelievable amounts of options even before the horn blows, and the process of exploration of new interactions, new ideas and new builds is hard-coded into the DNA of dedicated Dota players. At this point in time, the NewBee vs VGJ series in DPL is the only instance of a high-profile Dota series having the same exact lineups and let it stay that way. Because no two games of Dota should ever be the same.
nice one
0.02%? ..... so you’re saying there’s a chance....
I’d be interested in seeing the number of times the same 15 heroes were played in a series. Not same game but similar games. Bucharest appears to have a very closed meta. I mean Gyro, Sand King, TB, DP in every game isn’t exactly exciting.
Amazing blog
0.02%? that's way higher than my chances of bashing when playing sb
@Nemesis i guess it got to do with tournament format many teams wouldn't risk at all as it's always bo1
no your wrong... PubG is laifu!
i dont understand the point of this article tbh
@Mr Poopybutthole that’s a good point. But you would think that some of these teams that are underdogs would have tried to cheese one out. I guess I’m kind of surprised how little change there was between Katowice and Bucharest even after patch.
You didn't say how the games that had the exact same drafts were different! i.e. did they pan out in the exact same way??
Yes. Every single time, the team that lost and went for the same pick would lose again.
Did they lose in the same way though? Did the teams make the same moves at close enough minute marks? Were the differences in xp and networth close to equal? Did the laning stage go the same way, and evolve into the same midgame? Did these games have the same tempo? Because, if you simplify, then you can say 50% of the games go the same way, and 50% the other, either Dire or the Radiant destroy the enemy Ancient :D
Valve pls fix yr servers!!! Just tried to play ranked in SEA and within 10mins, everyone dc because of poor network. Its early in the morning there should not be much traffic. My next closest server is India that does not have ranked. (When u look at the server status its 5 ping.) Shame on u.
Valve pls fix yr servers!!! Just tried to play ranked in SEA and within 10mins, everyone dc because of poor network. Its early in the morning there should not be much traffic. My next closest server is India that does not have ranked. (When u look at the server status its 5 ping.) Shame on u.
-Dreadnought325 on Dotabuff.com
@KawaiiSocks
I'm assuming there are far more instances of double matches in pub scene, correct? More to the point, does it happen with the same .02% frequency in pubs? I'm guessing it's closer to .2 or even half a percent.
I am confused when you say at least one team would play the same lineup. If one team played the same five, the other team would have to play the other five.
@Brock Hall: there was an amateur team. Let's call it Team A. And Team A picked a certain lineup of five heroes against Team B. They won. Later in the tournament they were playing Team C and Team A picked the same exact heroes, while Team C picked the same lineup Team B did. )
When it comes to pubs, it is hard to make speculations. On one hand in pubs 4- or even 5-core liuneups are more possible and certain hero combinations that would be counter-intuitive in the pro-scene would actually work or at least get picked. So the diversity of picks is even higher. On the other, some pub players aren't too bothered picking strategically and have a pool of comfort heroes, which generally tend to consist of 5 or so easy-to-execute heroes and are frequently the same for a huge subset of players. On top of it, 3k+ meta is very responsive to the pro-scene, so there are times when all you get is people playing the most meta heroes from the latest tournament. So the short answer is we don't really know, can't make speculations and can't really test the claim, since there are billions of matches)
Сбалансированная игра? Наличие меты уже говорит о том что кто-то сильнее другого, и речи о балансе вобще не может идти.
Любая моба игра это постоянное взаимодействие персонажей разных мастей, по принципу камень-ножницы-бумага, и говорить о балансе когда влетает селедь или троль и не выпускает из стана , смешно разве нет , где тут баланс когда ты нечего не можешь сделать ?
I love calling dota a beautiful game, if only the community was anywhere near as good XD
Наверное, под словосочетанием " все игры разные", подразумевалось , что нет таких игр,которые бы были похожи не только по драфту,но и по самим действиям в игре.
@KawaiiSocks
Thanks for the explanation. Shame that we can't get more info for the pub matches but I very much get why.
Definitely like the concept of this kind of article. I'd love to see more of these in-depth researches that go beyond the winrate articles (although those are very interesting as well.)
a revelation