Friday, March 1, 2013

Dota 2 in Depth: Drafting

Since I still haven't managed to complete any new games (Xenoblade is rather lenghty!), I want to write more about Dota 2. For my topic I have chosen different drafting (hero selection) modes that are currently available in the game. I have chosen this topic primarily because hero drafting is one of the most entertaining phases of the game to watch. The irony here being that not much happens on-screen during drafting. Nevertheless, it was indeed drafting that sparked my interest in Dota. The most interesting drafting mode by far is the one that is used in tournaments, Captains Mode. A couple of other modes are however noteworthy as well.

1. Captains Mode

This mode is the de facto tournament mode in MOBAs in general. The drafting is divided into four phases: first bans, first picks, second bans and second picks. The name of the mode comes from the fact that it is the team captain that does all hero bans and picks for their team. In the first phase, a total of four heroes are banned. This banning phase could be called a balancing act in a sense because usually the heroes that are currently considered most useful get banned. These bans tend to give away very little information, and they are fairly standard. There is typically a pool of less than ten heroes that get banned in this stage. These bans are therefore the most dependent on the contemporary meta game. Some additional information does go into this process though: match history. Knowing what heroes the enemy team is particularly strong with generally allows a captain to make the most out of their first bans.

This first ban phase is interesting because it can give off the impression that the game is imbalanced in the sense that some heroes are simply better than others. This does ring true but one thing to consider here is what we discussed last time. Sometimes the meta game shifts because of balance patches that buff some heroes and nerf others, but they also shift simply because teams come up with new ways to use heroes. Besides, banning four heroes out of the pool does have one huge advantage: it prevents the game from becoming too repetitive. From the captain's point of view, the first bans also ask a very important question: "which two heroes we are the least prepared to deal with?" Answering this question does require some general idea of what kind of strategy the team is prepared to go for. The team that picks first also has an advantage here because they can leave two nasty heroes into the pool and force the opposing team to give them one of those heroes while banning the other.

After four heroes have been ejected from the pool, teams pick their first three heroes. The process has four steps: 1) the first team picks a hero; 2) the opposing team picks two heroes; 3) the first team picks two heroes; and 4) the opposing team picks a hero. This is the phase where teams generally pick heroes that afford a wide variety of strategies because giving away too much information at this point can be disastrous. Just like first bans, these first picks are fairly standard in competitive play. It can be tricky to not give away too much information here, because after these picks, three out of five heroes in a team have been picked. The last picks in this phase tend to be more interesting than the first ones. This is where teams have to start playing some of their cards and they also have more information. There are a lot of things to consider here. There's generally 5 roles in a team, and although some heroes can comfortable fit into more than one, this is often where three roles will be filled.

That is why the second ban phase is much more context sensitive than the first one. Both teams get to ban three more heroes and the goal of the captains here is to get rid of heroes that would be the best fit for the opposing team's lineup. This asks for quite a bit of strategy sense because now is the time where a captain really needs to have a very solid idea of what kind of strategies both teams are going for. This in turn requires pretty deep understanding of every single hero in the game. Although a lot of the same heroes do show up in matches, some really wild picks have been also witnessed in tournament games. The three bans are what make this possible: the most obvious best choices are gone, so a captain has to fill their last two slots with heroes that are not the most optimal choices. This is interesting design because taking away heroes from the pool actually makes the effective pool larger.

By the time teams get to the last picking phase, a total of 16 heroes have left the pool due to being banned or picked. The last picks are done in 1-2-1-2 order, and here the team that picks last has the advantage, counterbalancing the first picker advantage from earlier. In a sense they can counter both of the first team's last picks because the last two picks need to fill different roles. The countering ability of the first team is weaker because their very last pick needs to both fulfill a role and at the same time counter the previous pick (if necessary). Of course, the very last pick of the second team cannot be countered at all. This makes it the only pick in the entire process that is done with complete information available to the picker. If you have played any decent strategy games, the advantage given by complete information in decision making should be pretty obvious.

In a way the captains mode might have risen out of necessity. It is simply not possible to completely balance a game like Dota 2 without making it dull. The captains mode is a clever design that prevents the game from becoming too stagnant. Bans force captains to look for unexpected solutions, and this way the meta game slowly changes. I think there is something more interesting going on here though. I think that Dota 2 is so interesting because of its imbalance. The captains mode is not a monkey patch to fix an imbalanced game - it's a mode that capitalizes on the imbalance as an advantage. The myriad of ways heroes interact with each other makes the hero pool a highly interconnected structure. Taking out even a single hero from the pool can massively impact the usefulness of another hero. Therefore, although the total number of bans sounds quite small, their impact on strategy is huge.

2. Random Draft

This mode is the other mode that can be considered almost tournament worthy. Like captains mode, this one also reduces the hero pool size, forcing teams to pick alternate solutions. As a measure, random draft is far more drastic: only 22 heroes from the entire game are made available. This limitation is naturally even better at preventing stagnancy. The biggest reason this mode is not used in tournaments is its random nature. The draft can be rather lopsided, sometimes giving a huge advantage to the team that gets the first pick. This is especially true if there is only one good hero for one role in the entire selection. What this mode is good for however is playing with random internet strangers. Captains mode is tricky because one player makes all the decisions and poor communication can lead to very disastrous picks (such as four heroes that no one knows how to play).

Random draft on the other hand does two things that make it suitable for quite serious play on public servers: it forces a pick order which makes counter picking a more strutcured activity than the chaos that is all pick. Second, it restricts the hero pool so drastically that every player in a team needs to carefully consider their choices against their skills. It also acts as a better catalyst for communication, because after the first picks a clearer image of available strategies starts to form in a way that everyone can perceive them. Furthermore, the most popular heroes often don't all make it into the pool. This creates some much needed variety. Seeing certain heroes come up in every other game does get a bit old eventually. Of course you can get very unlucky, and have none of the heroes you actually know come up. This is the disadvantage of random draft. However, I do consider it the best mode for playing with random internet strangers. The only thing it's not good for is practicing a particular hero, because the odds of them actually being available are quite small.

3. Least Played

This mode is not mentioned because of its tournament worthiness, but because of its ability to encourage people to practice new heroes. The system is quite simple. Every player has their 40 most played heroes banned (but only heroes that the player has three or more wins with can get banned, so effectively this can be less than 40). What this means is that every player in the game is playing with a hero they are just not very good with. It gives players a friendlier environment to get the hang of new heroes and still allow them to play against other people instead of bots. Everyone knows that everyone else is new to their hero so people generally wont get flamed for not knowing how to play their hero (there are some asshat people who choose to flame in this mode, but they are very rare). It is the mode where players are allowed to screw up because the punishment for mistakes is generally lower (because the opponents don't know their heroes very well either).

I like this mode because it allows me to play some heroes I would not even dream of trying out in any other mode. I also like the fact that people care much less about winning or losing in this mode. Furthermore, most players have played the most popular heroes too much to get them in this mode, so hero picks tend to have a lot more variety than in most other modes. The matches can also get pretty weird in this mode, far more often than in any other mode. It is entertaining, more relaxed and even a source of new ideas. Currently my biggest lament is that this mode is not nearly as popular as it was after its launch, It often takes up to five times more time to find a least played match (compared to all pick).


Different drafting systems in general are an interesting topic, and they have been used quite effectively in Dota 2 to improve the game. What is really interesting about drafting in Dota 2 is its impact. Like I said, it is the most exciting part of a match between two professional teams. Seeing how it plays out is of course entertaining too, but the draft is where it is really easy to get absorbed. As the captains are doing their job, couch strategists can in a sense take part in the process by trying to guess what is going on in those great minds. The most fun of course is discussing the draft with other players who are watching the match. Another important way in which drafting affects the game is its impact on how games with random internet strangers roll out. The game experience is always affected by the draft mode. It allows Dota 2 to be played in more than one mindset.

For improving one's play with a given hero, there is all pick; for the best team play experience, there is random draft; and for the most relaxing gaming experience, there is least played. Captains mode is there for those moments when a full team gets together to really play some hardcore Dota. That is how I roll anyway. There are yet other modes in the game but they are less noteworthy.

No comments:

Post a Comment