Tourneys and Ladders: A Response to David Sirlin

Overwatch Logo

David Sirlin just published an article on his design blog regarding the newly released competitive mode for Overwatch. In it, he highlights some of the problems with the format, problems that result in a lot of draws. Sirlin approaches them from the perspective of tournament play, which is his forte. He asserts that the fundamental issue with the game types in question is that the number of rounds they require is even when it really should be odd. It’s a simple solution, but when you analyze the dynamics of that game type, it really isn’t so simple. And there’s a bigger problem than that because of one of Sirlin’s underlying assumptions: that the competitive mode needs to have a tournament format. But Overwatch competitive mode isn’t a tournament – it’s a ladder. And there’s a huge difference.

The Challenge of Asymmetry

To start with, here’s what you need to know: Overwatch has two game types in particular that are causing us issues. The first is “Assault,” a game where the two teams are split into an attacking team and a defending team. The attacking team must take and occupy a sequence of control points, in order, before time runs out. Every point they take flips to their control permanently. That means that defense can only prevent the attackers from taking a point – they can’t capture it back. As such, their goal is to hold the attackers off of the current contested point for as long as they can, hopefully until time runs out.

The second game type is “Payload,” a similar game where the teams are split the same way. In Payload, the attackers are still trying to capture points, but they do so by moving a cart forward on a track. Any time there’s at least one attacker and zero defenders present near the cart, it moves forward. When the cart reaches a control point, it flips permanently, just like in Assault. Defenders must keep the attackers off the cart, and the cart away from the last control point, until time runs out.

In both game types, both teams get a chance to play both attack and defense, and their performance is scored by the number of control points they took during their turn on offense. The problem: there are only so many control points on the map, and if both teams’ attack runs end up pushing all the way and capturing the final control point, they’ll both get maximum score for the map. Hence the problem with draws.

The whole issue can be framed as a problem with a truly asymmetric game – the goals of Team A and Team B are different goals. It’s really hard to know if offense and defense are balanced against each other, and if one side has it easier, then it’s kind of a coin flip who gets to win (i.e. whoever gets the easier side). This is especially true if matchmaking is any good, because ideally you’re striving for a 50% win chance for both teams. But if one team gets the advantaged side, that undermines all your matchmaking efforts. This dynamic is the entire reason why we switch sides at all: to give both teams a shot at both sides. That way it doesn’t matter if offense is balanced against defense – you get to do both.

It’s important to have a clear understanding of the problem so that you solve the right thing. Right now, Blizzard is solving draws by having a sudden death round where one team has a shorter amount of time to push only the first point on the map. Sirlin seems to believe that the reason people don’t like the tie-breaker round is that it’s not a proper round three, but an abbreviated pseudo-round.

But the real problem with the tie-breaker round is that by deciding the attacker with a coin flip, we’re right back where we started. We just got done giving both teams a chance at both sides in the interest of fairness, and now we’re going to undermine that in the worst possible way. Not only does one team get the advantaged side (whichever that happens to be), but if the advantaged side is attack, they’re maximizing the advantage by having them only capture the first point. If this is what we were going to do all along, then we shouldn’t have even played the first two rounds – they have no bearing on the sudden death round and were a huge waste of time.

But wasting time is exactly what Sirlin wants to do. He has an obsession with tourney-style rules and as such is a big believer in having the third round – in other words adding time to the match. In the TLDR, he even suggests tacking on some other game in the event of a tie, something like King of the Hill. Well gee, Sirlin, why did we even play this whole game of Assault if the game that actually mattered at the end was KotH? Why don’t we settle it with an ARAM match of League of Legends? Maybe a round of Calvinball?

Truth is, there’s a completely acceptable solution that’s staring everybody right in the face, and nobody’s willing to admit it: tie-break by comparing time taken or distance traveled. It’s a better solution than anything suggested so far for two big reasons. First, it doesn’t change the logical definition of the game. The game never stops being assault or payload. It doesn’t turn into something weird like mini-assault or KotH. It’s two teams comparing who can better accomplish the goal of a unified game. It doesn’t get any more honest than that.

Second, and more importantly, it keeps the matches as short as possible. It’s already kind of onerous that we even have to play twice at all. We have to remember that switching sides isn’t just another wave of the designer’s magic wand. The role reversal isn’t cost-free – it’s a compromise that we made. We sacrificed players’ time in the interest of a fair match. Ideally we wouldn’t want to make the match twice as long, but it’s a trade-off that we made to even play this type of game at all. Trying to fix the resultant ties by tacking on even more rounds or, in Sirlin’s case, entire other games is the opposite of what we should be doing, because every second the match goes on is precious. Striving for the match to take up more seconds takes an already non-ideal situation and makes it even worse.

The Difference Between Tournaments and Ladders

So why does Sirlin want to do add more time? It’s that fixation on tourney-style rules. A tournament has to be relentlessly obsessed with “Who is the best one here?” Time is a factor there, too, because people at the tournament want to go home someday, but no self-respecting tournament would end any match in a draw, so time necessarily takes a back seat. As such Sirlin expects that time always takes a back seat any time you hear the word ‘competitive.’

But competitive mode isn’t a tournament. It’s a ladder.

Competitive ladders and tournaments are two different things. While both a tournament and a ladder are interested in who the best players are, the ladder is only interested in player ranking as a means to an end: player matchmaking. Whereas a tournament eliminates players as they prove they’re not the best, a ladder never eliminates players. The whole point of the ladder is to let players play the game whenever they want and have a reasonably skill-balanced match when they do. Overwatch positions competitive mode as SRS BZNS, but let’s not let that completely take over our brains. Time is the Number One concern in ladder play for two reasons.

First, shorter matches mean faster match results. Faster match results means people get rank adjusted more often and play more games. More frequent rank adjustments means that people get to their appropriate rank faster, which means better games for everyone. And last I heard, Blizzard was having some trouble with that – no less than 50% of the entire player population falls in a very narrow window of ranks. Shorten the matches, get as many of them in as you can, and you let the ladder do its work. You can’t obsess over the winner of every single match – you have to consider the whole season.

And here’s a twist: in that vanishingly small probability that a game results in a draw in time or distance, let it be a draw. I’m sure that’s an auto-reject for Sirlin, but let’s remember what the whole point of the ladder is. If two evenly matchmade teams deliver completely equal performance, that sure looks to me like the ladder is doing its job. Moving them farther apart on the ladder for the sake of someone having to win is counterproductive, especially when it takes more time to do it.

The second reason time is our priority in ladder play is because, god dammit, it’s time. I shouldn’t even need to justify this, but having a game take up more time than it has to is the ultimate cost. We only have so much time in our day, in our lives, that flippantly upping the time requirement for a game is a special brand of cruelty. We already have so many games that take stupid amounts of time, and games whose whole purpose is to take up as much of our time as possible, and it drives me bonkers. I’m drawing a line in the sand here and now because someone has to. Even concerns of fairness or balance don’t constitute a license to take as much time as you want. Players giving you their time is just that – a gift. Cut down on the time requirement, and you give players back control of their lives. That’s the least you can do.

What is a Win?

Sirlin won’t like using time taken or distance traveled as a tie breaker. He says it’s “game-warping.” “A win is a win,” he says, implying that it doesn’t matter whether you win by a lot or a little. This is the obsession with tourney rules talking again. There’s a curious dichotomy in his thinking: he’s aware that the reason we’re playing more than one round is because the asymmetric goals are more or less forcing us to, but he also chooses to see each round as completely isolated from the other. In that way, we’re playing more than one round not out of fairness, but simply because that’s how tournaments are typically run. And as such, he doesn’t want to see strategies that are runaway victories be seen as superior to strategies that take a long time to work. In other words, he doesn’t want to privilege some strategies over others, because he wants them to be equal in viability. I guess that’s “game balance?”

The whole line of thinking is a departure from how games and game design actually work. Privileging some strategies over others is the fundamental act of game design. You can’t avoid doing it. Every rule you make, every number you tweak, every card you write up, every little bit of map geometry you place are all in the service of defining what successful strategy looks like. As the designer of the game, you are the decider of right and wrong, of good and bad play. Not, “You should be…” You are. You cannot get around that fact, so you may as well embrace it.

Embracing it means accepting the unavoidable fact that some strategies are just off the table to start with. The fact that you have to push a payload at all means that six snipers is a terrible idea. Attacking with any combination of defensive heroes is a pretty questionable choice – Overwatch says as much when it suggests “Too many defense.” Strategies where you ignore the control points – off the table. Healing chains that focus on survival over eliminating enemies – off the table. All that seems pretty head-bonkingly obvious to any Overwatch player, but only because the game has already winnowed available strategy down to a razor’s edge. Who knows, maybe those off-the-table strategies might “barely win” sometimes. We’re still discriminating against those strategies, because that’s just what games do.

Sirlin obviously doesn’t mean that all strategies should be viable. But then which ones does he mean? He wants some strategies to be “just as viable” as others, but which? What makes them viable? What makes them equal? When does a strategy suddenly become not worthy of viability and gets banished to off-the-table land? Why does the slow strategy deserve any special attention just because it wins sometimes? He talks about “strategies that happen to win by a lot sometimes,” but why does that happen?

And why is he talking about strategy like it’s a series of immutable monoliths that teams adopt with unshakeable conviction? Strategy isn’t some stone block stubbornly grinding against another stone block. It’s fluid, it’s adaptable, and that’s especially true of a game where you can switch roles whenever you want. Strategies don’t have “barely wins” as a quality of the strategy. There are only strategies that barely won in this specific circumstance, and it’s the designer who gets to say what it means to “win” in the first place.

If that happens to include “Do this faster than the other team did it,” that informs, privileges, and extinguishes strategies no more or less than any other victory condition. If anything, encouraging players to do something “fast” creates a positive play dynamic that discourages teams from dragging out a match. We’re going to make people play twice anyway, so we can make up for it by putting the match into overdrive.

Sirlin has a lot of experience in competitive tournaments, and has some valuable insight to add to them. I don’t mean to denigrate his opinions. I don’t even think Sirlin is wrong to think of competitive tournaments in the way that he does – the fact that draws are possible, even in my solution, is a big problem for tournaments, and he has valid input there.

But he needs to remember that the “tournament” isn’t the only competitive format and that different formats have different goals. On the surface, they both look like their goal is to see who the best players are. But tournaments and ladders rank players very differently, and each has a place when you consider their underlying goals. A tournament is ideal for spectators where time is less of an issue. A longer match creates a larger narrative arc and makes for a memorable event. Tournaments are also good at finding Number One, and playing many matches makes sure you’ve got him.

For the rest of us, there’s ladder play: a mode that lets us play on our own schedule, lets us play as much as we want, makes sure the games are evenly matched, and gives us good feedback for our change in skill over time. Ladders don’t want long games, because it works against everything competitive ladders try to do. Sirlin has tournaments figured out, no doubt about that. But he should reserve his judgement for things that are actually tournaments. And Overwatch competitive mode just isn’t one.

4 thoughts on “Tourneys and Ladders: A Response to David Sirlin

  1. Lauri August 5, 2016 / 7:15 am

    You got it right. It baffles me how stopwatch mode is not preferred over a coin toss. Team Fortress and others have been doing stopwatch for years and there it works just fine. It’s cool to try to iterate but if you can’t come up with something better, just stick with the old version.

    I don’t get the warping strategy argument either. What is it about Overwatch that demands stalling strategies to be so much more desirable? Is it the ultimate charging mechanic? If anything, a slowly charging super power would guide my intuition to reward the teams that don’t have to wait for it to advance.

    • The Ludite August 5, 2016 / 10:31 pm

      The way Sirlin talked about strategies really didn’t jive with me because it was way too simple. He writes about certain strategies being “X powerful already,” and if you make a certain change their “power level increases quite a bit.” But why is that bad? Do we want all strategies to be equal in all circumstances? Are they like moves in Street Fighter where you have explicit counters in a rock/paper/scissors sort of game? Clearly not since strategy is only partly about your composition and partly about the way you execute it.

      It also doesn’t account for all skill levels. For a strategy to be “X powerful,” you also have to define “against who?” Being X Powerful is not an inherent property of a strategy, because not everyone sees the strategy in the same way. Some people see right through it, other people will have a tougher time against it. Maybe you could call the former group “good” and the latter “bad,” but the situation might be reversed for the same two groups when you consider a different strategy. Certain things are VERY powerful in low-skill matches, and non-existent in high-skill matches. And hopefully what we consider to be “high-skill” won’t always be, as we discover still more strategic depth. That won’t cause what we currently regard as “high-skill” to stop existing, and at every single skill level that exists, strategies will have a different value for X Powerful.

      The point is that he’s looking at strategy too narrowly. If you want to discourage a particular behavior across the board so as to make the game play better as a whole, that’s worth doing. But trying to preserve “strategic balance” by making strategies equally powerful is a fool’s errand. That will never happen at any skill level, nor is it desirable. The act of forming strategy is to cut through the realm of what’s possible, and arrive upon what works. And what works is in motion all the time. You can’t simplify it down to “this strategy works 45% of the time.” There’s so many factors contributing, like skill level, what parts of the game each team has learned, how much depth has been explored at large, that there is no acceptable number.

      Often times people get too caught up in what’s happening at the top tier of play, and regard everyone below that level as muddling about in no man’s land before they finally get to the level that “matters.” If anything, the average player’s experience is the one that matters most since the broadest band of players plays there. It doesn’t make top-tier play unimportant, but it’s only a small slice of game balance in the grand scheme of things. We shouldn’t declare only top-tier play to be the “real game” – there’s a lot of people in the middle, and the game is very real to them. They’re climbing and mapping the heuristic tree individually, each in their own way, and that’s what makes games fun. Blowing off every layer except the top one makes no sense.

      Thanks for reading!

  2. anonymous July 28, 2016 / 8:43 pm

    I think you’ve misunderstood Sirlin’s blog. He wants to have competitive play that preserves the game’s identity. This is what he means when he says he doesn’t want competitive to be “game-warping.” If you start judging players based on time taken it fundamentally changes strategies and play style. This is acceptable if you want to create a new game mode, but as you say: competitive is a ladder. Optimally competitive should act as a way for players to compete against other players of their same skill level in Overwatch as it exists, not an alternate version of Overwatch.

    Disclaimer: This is coming from someone who hasn’t tried competitive Overwatch.

    • The Ludite July 30, 2016 / 6:19 am

      I think that the hand-wringing that people are doing about changing viable strategy is a huge overreaction. I’m not saying that changing the match format won’t alter strategy, but I am saying that every single change you make alters strategy. It’s just as “game warping” to limit the characters to one per team or to give D.Va a toggling shield with a recharge. Game designers make it their job to warp the game. Trying to preserve the game “as it exists” isn’t even desirable. Overwatch is a first draft in the grand scheme of things. It’s going to take quite awhile to get to the point where we even have a game worth preserving, and the proof is in every patch note.

      So warping the game is good. It takes us one more step down the road of improvement for the overall game, and cutting down on the match time so that people can get in more matches or manage their own time better is solidly in the “improvement” camp. If we end up altering strategy in a BAD way, we have other knobs we can tweak to fix it. In fact, Blizzard proved it when they limited characters to one per team – it directly fixed the “lame” strategies Sirlin named in his article, and we didn’t have to resort to lengthening the match structure to do it.

      In short, don’t be afraid of change – it’s inevitable. The younger a game is, the farther it has to go to reach a state worth preserving. So warp that game! It needs it.

      Thanks for reading!

Comment on this article

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s