CartDota: On the subject of roster locks

by Corey "CartDota" Hospes Jun 29 2016
Thumbnail image courtesy of SL i-League

Valve introduced some big changes to competitive Dota after the International 5, but few were as controversial as the roster lock system. For most of the history of the game, teams were free to add and drop players as they wished, and they made liberal use of that freedom. By the time TI5 rolled around, the slightest gust of wind could shake a top team apart. Valve decided something had to give, so they introduced a series of seasonal deadlines when teams would have to lock in their lineup.

At first, almost everyone accepted that teams would have to stay together for a few months before each of the new Majors. More stability! I can keep cheering for the same team when they're not swapping stars every two months! Aside from a few dissenters, the majority of fans seemed to be satisfied with Valve’s decision to take a more direct hand in ensuring roster stability.

Amidst the general jubilation, however, a handful of cynics saw a problem on the horizon: Valve's deadline schedule included a single lock period after the second Major of 2016, which would last until after The International in August. A six-month long roster freeze, compared to just three months for the winter and fall events. Could teams really last that long without making any changes? Six months is a long time to commit to anything, especially when millions of dollars and your dream is at stake.

Now in the middle of that six-month lock, it's patently clear the system hasn't worked as intended. As most observers could have anticipated, teams that performed poorly at Manila decided to change their rosters after the event ended, and the threat of facing open qualifiers proved a poor deterrent. Shockingly, the two teams to violate Valve's rules included the winner of TI5, making this the first International where the previous year's winner has not been invited. It's easy to see this as the teams' fault for breaking the rules, but did the rules really make all that much sense, given the realities of competitive esports?

Lock, stock, and two multi-million dollar tournaments

By decree of Valve, a team that breaks a roster lock before a Major can't be directly invited to the event, and the only way for the perpetrator to earn a seat there is to go all the way through the open qualifiers. Open qualifiers, initially seen pre-TI5, are the punishment meant to scare teams into cooperating, since a run through opens means facing eight best-of-one matches in a single day, then two more on the second day, followed by two best-of-threes (without even considering all the matches in the main qualifiers that come immediately afterward).

Every match in the opens is an elimination match, and if you lose you have to try again on the second go-around and do it all again. Fail again, and you’re out of the running for good. Many players would agree that Dota is not a best-of-one game, and nobody wants to be potentially cheesed out of the big-money Valve events.

The decision facing EG and Secret after Manila was whether their internal problems were dire enough to risk the open qualifiers. Put yourself in their shoes for a moment: you’ve been playing with Universe or Bulba the past three months, things aren’t working out as planned, and there are clear internal issues within your team. You’ve just finished at the bottom of the biggest event before TI. You can either keep trying to fix your problems, or you can try a fresh start — from the very beginning — with a new player. Should you really take that failing team to TI, where you're almost guaranteed to finish outside the Top 8? Or should you face the qualifier gauntlet with a new one?

As it turns out, the choice EG and Secret ultimately made wasn't the one Valve set up the roster lock system to encourage. Both teams realized that with the experience and TI-winning talent on their lineups, the opens weren't really a serious challenge, even if they were performing at less than 100%. To these teams, even Regionals aren’t that scary — irritating, perhaps, but not a viable enough threat to force them to stick to a lineup that isn't working.

Getting cheesed is still always a possibility, but both Secret and EG made the bet that a second-tier team running a cheap start wouldn't be able to beat a world-class team, and they turned out to be right.

From a distance, that doesn't seem so bad. EG and Secret ended up getting their hands slapped, paid their penance by playing through the qualifiers, and ended up at TI6 as they would have anyway. Unfortunately, that neglects what happened to the other teams that went to the open qualifiers, including the handful of mid-tier teams that broke roster lock.

When the invites for The International were announced, there were, as usual, two open qualifier spots at the main qualifier. With EG and Secret competing in two different regions, both were essentially guaranteed one of those open qualifier spots. That meant that for everyone else, the number of qualification spots was effectively reduced to one, and that there was little to gain from playing in the first open qualifier other than exhaustion. The same was true at regionals — Secret and EG were more or less guaranteed to occupy one of the two regional qualifier spots at TI6, meaning there was effectively only one spot for everyone else.

One of the reasons direct invites exist is because sending top-tier teams through qualifiers wastes everyone's time. Sending a top-tier team like Liquid or OG through regional qualifiers is a recipe for boring, one-sided games, that discourage aspiring b-tier teams and end in predictable outcomes. There's no question Secret and EG in their current state are not Liquid or OG, but they’re still strong enough to win qualifiers with (relative) ease — as they showed by, of all things, winning the qualifiers with ease.

The core problem here is not that EG and Secret had to face an extra six days of games they didn't really want to play; it's the collateral damage that was done to up-and-coming NA and EU teams that didn't break Valve's rules, but nonetheless were robbed of their shot at getting experience in the regionals and TI6. Teams like Shazam and Kaipi, both of which were eliminated in the open qualifiers after having broken roster lock for similar reasons to EG and Secret. What happened to those teams showed the regressiveness of Valve's the roster lock: it hurts less-skilled teams much more than top-tier ones. And after all, Kaipi didn't have much choice but to break roster, when zai returned to EG.

Even internationally proven teams like Complexity have suffered because of EG getting sent through open qualifiers. Ineligible for a direct invite, EG's presence in NA has now forced them to fight for tournament life in the wildcard stage.

A better way(s)

I decided to ask Joe Eastham, Kaipi's manager, for a few suggestions about how to improve the roster lock system. The first thing he suggested was having one less Major each year, to avoid the back-to-back double-lock period with TI. "Having only two Majors would even out roster lock length," he said. "When normal locks are three months apiece, a six-month roster lock is incredibly daunting for most players to actually commit to."

He also thinks there's a way to avoid punishing other open qualifier teams for EG and Secret's infractions. "While I think it’s good that EG and Secret were forced to go through open qualifiers since it means Valve is sticking to their guns, I think there should be an additional open qualifier spot added. One in each region is essentially forfeit to EG and Secret," he said.

Creating separate lock deadlines for dropping players and adding new ones, something that's well established in other esports, is another solution that's been widely recommended. "Separate drop and add dates would prevent players being dropped from their teams right before lock date," Eastham explained. "This would give players dropped by their team appropriate time to find a new team.

Separate dates would also allow tryout periods if spaced enough: have a drop date, and give teams two weeks to tryout new players before rosters lock again. If teams can’t find a suitable replacement, they could ask their former teammate to rejoin if they so chose to.

It’s good that Valve is taking a more direct approach in the professional scene. Their hands-off attitude has at times gotten incredibly frustrating for us as fans, and trying to influence roster stability is a welcome change. However, like with any new idea, roster locks and the rules around them need to be tweaked to find what works best. For Valve's system to work as advertised, it will need some significant debugging, especially in light of what's happened this season.

Corey "CartDota" Hospes is a freelance writer, Dota 2 addict, lover of numbers and techies picker.​ You can follow him on Twitter.