Debunking Asymmetry (aka one faction > n factions)

(where n > 1)

http://keithburgun.net/debunking_asymmetry/

It's a fairly good read, if you wanted a long article arguing in favour of PA's current approach.

"I played a ton of Warcraft III back in the day, and I can tell you: arguing about different races, heroes, and units is way more fun than actually playing the game."
/guilty

I'm not sure to trust his distinction between "inherent asymmetry" (e.g. good asymmetry that doesn't cause problems) and "componential asymmetry" (e.g. bad asymmetry that does cause problems).

While some of his points are good, some are pretty bad.

"A great example of how to do non-asymmetrical flavor is something like heroes in Warcraft III. If the game was only Human vs. Human, there would still be plenty of emergent asymmetry in hero choice. If you get an Archmage first, and I get a Mountain King, that’s a significant difference that really feels different. If you get a second hero and I don’t, that makes it even more different. And this game wasn’t even designed to be single-matchup; imagine a game like that that was."

His argument here is that the game is different and has options even when its only one race versus the other, and that difference is good, but then goes to say that the game would be better if we eliminated the other DIFFERENT races? All he's doing is contradicting himself, saying that more is not better, yet he also believes that MORE strategies are better.

I do completely agree with the fact that non-strategic decisions influence the strategy of the game however.

Interesting article. Funny to see my friend Raph Koster posting in the comments.

Anyway we did this because it's cheaper and seemed like a good idea.

I also have some trouble with his Warcraft 3 example. I don't know enough about a 1v1 human mirror match-up to be sure but isn't the hero choice a blind choice? If it really is a blind choice, meaning that the players make their choice of hero before there have been any counterplay, then it is no different than choosing race before the game.
Or take Starcraft for example. Even in a mirror match-up the players build order is very important. If the strategical choice of build order is locked in before any scouting has been done, then there can be no counterplay and it have basically the same effects as a choosing a race.
Now people in Starcraft accept that some builds counter other builds and see it as part of the mindgame and trying to outsmart your opponent. I'd say that it is strategy even though it occurs before the game has started.

I'm not entirely sure what you mean but I think you mean a blind strategic choice like described above and like in rock-paper-scissors.

Personally I think I would prefer the blind strategic choices to have small effects in the start of the game but much larger effect later in the game.
I think that blind RPS in the start of the game which has severe consequences that determine the winner of the game should be avoided.
However later in the game players will always have the option to scout the opponent so any strategic choice that can really decide the game in an RPS manner is okey and probably even good for gameplay.
Scouting in TA and SupCom requires resources. Do you want to be sure what strategy the opponent is performing or cut scouting and expect what strategy your opponent will perform so that your counter will be even more powerful and unexpected?
So in that design, early game scouting is easy and/or the strategical choices you make have less impact on the game while lategame scouting is hard and the strategical choices you make have deep impact in the game.
Would such design goal be good for PA?

Nothing like reading the opinions of a designer who works exclusively on casual mobile games.

The main thing that put me off reading the entire thing is how much it reeked of opinionated arrogance. He came across as someone who knows very little but tries to argue his point by appointing himself an expert.

I'll make sure to avoid buying his book.

Thanks for pointing this out. I quit reading the article before I got the comments and I read Raph's comment and I'll agree that there were SOME good points. I mostly dislike the delivery.

= addendum =

The truth is that very few games are truly asymmetric. A map could be asymmetric, sure! But in any character selection screen, these are choices you must make, there IS strategy in selection.

The issue occurs when you have the "blind pick" scenario where you must pick without any knowledge of the opponent. This is EXACTLY why playing Draft / Ranked in League of Legends is several magnitudes more enjoyable. There is strategy to picking your line-up and countering their choices.

Any game that offers choices, is to that degree, asymmetric. Any game that offers no choices... is a movie.

Planetary Annihilation is NOT asymmetric. The maps are randomly generated and you get to choose your location. The point is that the other team has roughly equal opportunity to you. That makes it FAIR but not symmetric.

I do, however, have an issue with games that create diversity with no reasoning why. If the character, unit, option, mode doesn't offer something unique then why have it?

But these, after all, are just my uninformed opinions.

By that definition symmetry has no meaning in games because they are not interactive unless they have asymmetry.

Godde... that doesn't actually make sense. If you take, for instance, Starcraft 2, and only can play the terrans, and you have a multiplayer game with them, that IS symmetry, and it's still very interactive.

Asymmetry, especially in a multiplayer game, has long been understood as a hard to design around problem. It doesn't mean it's bad, but as the article states, it does limit your options a bit. If you know everyone is painting from the same palette, and the only difference is their personal skill level, the game has symmetry, only the players are asymmetric.

And actually, while our units have symmetry, our procedural planets are actually the source of most of the asymmetry in PA - you can't guarantee everyone will have the exact number of metal spots on start, or possibly have some highly desirable cover, via mountains, plateaus or ravines to protect against some types of flanking attacks. But that is asymmetry derived from the system, rather than the designed gameplay. Those players still have the exact same units, and most of the same strategic decisions to make - to turtle a bit to start, or immediately grab territory. But statistically, over the course of 50 games, two equally matched players should have a roughly 50/50 win ratio.

wouldn't it be a 25/25 win ratio then?

garat. I were mostly trying to point out that TheOwns definition of asymmetry is too broad. I can agree that the strategic choices that the players make is a sort of player induced asymmetry.
If you chose to focus heavily on tanks you cannot also focus heavily on air units for example.
I'd say that strategy games are about making choices with uncertain outcomes. Usually it is a mix of blind choices and trying to outsmart your opponent and solving puzzles from which you can emerge victorious.
Even a game with symmetric starting conditions can erupt large emergent gameplay asymmetries and having only 1 factions makes it much easier to allow more diverse strategies that are balanced to each other.
There will always be asymmetry in a game otherwise there wouldn't be a winner.

An interesting read, although i don't fully agree with the premise. When reading the article, the first warning bells went off at the handwave that asymmetry that is somehow "inherent" to the game is fine, but all other forms of "video game asymmetry" are probably bad. The difference between the two is highlighted by a couple of examples, but not really rigorously defined. I tried to think about this a little, and eventually noticed a pattern in how the writer defines good and bad asymmetry.

To simplify this, the author seems to believe that "bad" asymmetry is asymmetry that arises during the pre-game. For those that don't know, pregame is a term that I first heard from GoogleFrog on these forums (although i don't know if he originated it), which means the portion of the game where players must make decisions about how they play, yet have no knowledge about what their opponent is doing. For example, I would argue that the pre-game in most RTS games starts before you even hit the start button - it starts when you pick a faction without knowing which faction your opponent will pick. It continues until your first scout sees an enemy unit. In fact, most games have a pre-game. Even chess has a pre-game. It's called the first move made by white.

The writer of this blog seems to dislike examples of asymmetry such as picking factions in RTS games or picking characters in fighting games or MOBAs. In other words, they are all asymmetrical decisions made by the player in the pre-game. I don't know if the author has noted this common trend, nor do I know if this is a widely acknowledged concept in the industry (although i would be surprised if it hasn't been discussed somewhere). This seems to link with the idea of the difference in decisions in the pre-game and in the main-game. Mainly that in the pre-game, most choices boil down to the question of "should i be a generalist, or should I be a specialist?". A player doesn't yet know what their opponent is doing, so they can pick a strategy that allows them to adapt, or they can trade that adaptivity for power in one specific area of gameplay. A player has the decision of playing it safe and keeping their options open, or doing something risky that promises either a quick defeat or a quick victory. Here, asymmetry would most likely be associated with the large number of high-risk strategies. Eg. a bomber rush vs a tank rush. This is an asymmetrical position, which would contrast with players who adopt more balanced strategies, such as a reasonable mix between bots, tanks and air. The generalists appear much more symmetrical, as they are both adopting relatively formless strategies.
However, the generalist is unlikely to have committed enough resources to the specific counter to the specialist's mad rush, as they have spread their resources. Thus, if the specialist is able to leverage even the slightest advantage, they have likely won in a brief battle. If they are defeated, then their strategy is thrown into chaos, hence they have lost.

The problems raised in the article with asymmetry are all ones based on the case of asymmetry without information of what the opponent is doing. This is effectively "how much do you want to push your luck?". Fundamentally, i don't think there is anything wrong with using luck as a game mechanic, although games that do so too egregiously don't appeal to me too much. Risk/Reward Evaluation, which is perhaps a more technical term for "how much can you push your luck?", is a common source of game mechanics, and plays a part in a huge number of games. Thus, my analysis is that the key issue that the writer has isn't really a problem with
asymmetry, but a problem with Risk/Reward Evaluation as a game mechanic. Pretty much all of the negative points highlighted can be boiled down to statements along the lines of "risk/reward is hard to balance". On that point I agree that it is difficult, but I disagree that this difficulty necessarily makes it a mechanic to avoid. Being able to correctly use and balance mechanics is part of what separates a good game designer from a bad one. Sadly, a rigorous definition of that is unavailable, but it doesn't mean that mechanics should be avoided because they are difficult to do right.

From my perspective, arbitrary examples picked throughout different genres is a bad choice to "debunk" asymmetry because seldom asymmetry is just used for the sake of using it.
The asymmetry has to fit the envisioned game design in terms of mechanics and story. I think, most agree that using asymmetry in an efficient and balanced way is hard, at least in the time and money spent on it, which is the reason why I can understand the choice of not using it in PA in favor of a common unit pool. Besides, in PA it's easy enough to blend lore-wise into the common "machine" background and RTS games don't necessarily have an inherent need for asymmetry.

In other genres, the game mechanics tend to be more story driven, e.g. RPGs, where the asymmetry arises from the different backgrounds and it often seems an awkward compromise if classes for example are arftificially symmetrical across factions. (That doesn't mean, the corresponding forums aren't filled with screams for balance, but hey everyone has his cross to bear. )

I have to say I was a little disappointed with the article. I was hoping for some light on the benefits of symmetrical design over asymmetrical design. Instead it was a pure attempt at debunking of the idea that asymmetry is inherently good.

Just wanted to put a few comments in myself, based on my play of the latest iteration of Splinter Cell, of all things. The two basic modes in multiplayer are Blacklist and Classic. Blacklist is your asymmetric game in the sense that Keith is talking about. You pick your character class, down to the individual abilities before the game starts. This creates a wide-range of different abilities and skills that you could be playing against. Worse still, you don't actually know what it might be until it's been used on you, because you do not get to see people's loadouts before the match. This means that my first forays into Blacklist mode have been incredibly disorientating.

Contrast this with Classic mode. In Classic mode you are either a Merc or Spy. You do not get to customise at all, but you know without a doubt exactly what you and your opponent can do. Here I can be confident that I know the rules and the way that the game can be played and can begin to formulate strategies instead of learning all of the combinations of equipment and abilities in the Blacklist mode. Incidentally this is also why I dislike MOBAs. I have tried to get into DOTA, LoL and HoN and each game has too much to learn before you can begin to get good at the game.

Going back to the article the reason I was disappointed with Keith for not giving the benefits is I wanted to hear how he would deal with ensuring a game he designed was not degenerate. A degenerate game is a game where there is little strategy because the dominate strategy cannot be beaten. This is the issue that Uber have to deal with very carefully. If a degenerate strategy is found then the game boils down to who can mechanically implement the strategy more efficiently. This is a little bit like the early strategies in Starcraft 2 where it was all about who could get units out on the field the quickest.

I am looking forward to seeing the rest of the variety that Uber put into their game, but for now the path of least resistance for me to win matches is to expand and build more tanks than my opponent. This is more due to my computer not being able to handle matches and crashing out a couple of times near the end of the game than anything else, and Uber's efforts to make Bots viable is a sign that they know what they are doing.

The above feels like a ramble, but hopefully you've stuck with me. If not, Symmetric and Asymmetric designs are not inherently better than the other, but they both solve problems the other design has. I enjoy both types (Starcraft 2 & PA) and think that often asymmetric design is used for ideological reason rather than design reasons.

yourLocalMadSci I am not sure if the article is referring to pregame RPS or not. It's at least a similar idea but technically the concepts are orthogonal.

The article only talks about asymmetry in terms of choosing a character or faction at the start of the game. From there asymmetry can develop through play, it does not rule out pregame RPS.

Pregame RPS is when one player has a distinct advantage over the other before any information on the other player is available. If you have no information about your opponent then you're basically playing singleplayer until you discover them, I don't recall using the term 'pregame' by itself but it works as a description of this phase of the game. A reduction in pregame RPS corresponds to a reduction in viable strategies which hardcounter each other which players have to lock themselves into before anyone has a chance to scout.

You could have a game with one faction and some very stark choices at the start of the game as well as difficult scouting. The stark choices could create pregame RPS but this article would call that a symmetrical game. A game could have many factions and just be balanced well enough to have no pregame RPS problems, early scouting is the biggest help in fighting pregame RPS.

The Warcraft III Humans vs Humans example clarified this for me. I didn't play competitive WarIII but I know the mechanics fairly well. From the way Humans can start I assumed that people would choose their first hero before they scout their opponent, in the pregame. So it's not saying that asymmetry which arises before you scout your opponent is bad. I don't think anyone thinks that is bad as long as it is balanced. Pregame RPS is bad but only when it creates a blind counter system. Diverse options create the potential for this to occur which just makes it important to balance these options.

Also there are degrees of asymmetry. Zero-K is technically a symmetrical game but you can start with several factories which are each fleshed out enough to be treated as a distinct faction. Factories are expensive enough that in a 1v1 you're unlikely to make a second for several minutes so your starting factory is like a faction. Even as the game progresses you are unlikely to have exactly the same units as your opponent because nobody builds every factory. It's sort of like hero choice in Warcraft III but for your unit set.

50/50 = 50%/50% = 1 / 1

i know, i meant, 25 games won/ 25 games lost (since your sample size is 50).

I guess statistics are not very good for jokes.

Personally I don't see the difference between choosing race and choosing the first hero in WC3. The differences between race and hero choice is more profound in Warcraft 3 but it is basically the same pregame mechanic in my opinion.
If you had a game where you decided the faction/race 1 minute into the game it would still be pregame choice if you are unable to scout your opponent before the choice is made. Worse yet, you might not even know what faction/race you are facing making the pregame RPS even worse if there is one.

Thanks for your views, Godde.

My point is that why should we vilify one choice (i.e. picking a faction) but not another (i.e. which tank to build first). It is a complete arbitrary.

If we say, "Well the problem with factions is that they are often a blind choice" then could we not also say that any game that has fog-of-war creates these blind strategic choices? Yes, we can mitigate those issues but the same CAN be true for faction choice.

But this is exactly my point. While we can understand that more choices = less balanced play, we can understand that taking it to its extreme, fewer choices creates fewer outcomes, to the point where if we reduced Rock-Paper-Scissors to just Rock, then we'd always tie.

Extremes aside, choices create asymmetry. To say that picking your faction is somehow less strategic than deciding on a build order is just idiotic. Just as one faction can be superior to another, so can one strategy, or one build order.

His argument that if one strategy becomes dominant then the game requires a balance patch just points out his idiocy. The same is EXACTLY true if one faction/character is dominant. So you fix nothing.

Less choices = easier balance. Less choices = easier to learn.

More choices = more diversity. More choice = more replayability.

If anything, the article misses the point. Giving the player a choice which isn't based on strategy and is purely a blind choice detracts from gameplay. It's like playing Chess blindfold.

- QFT -

^ This is the point I was trying to argue.

Godde that was my point. The article seems to take a narrow (and pragmatic) view about asymmetry. It's not really theoretical game design, it's talking about the realities of making and balancing a game. In this context many factions creates a lot of work for little gain but asymmetry within the mirror match is good so attempts should be made to balance it. It's less work overall because many units are reused.

Creating two races is a lot of work in balance and art. Creating different heroes for the same race reuses most of the art and does not 'throw away' balance work because most unit-unit matchups will occur in any given game.

The simplest point to make is that asymmetric games are some of the most popular games on the market today.

League of Legends - over 100+ different characters
World of Warcraft - multiple races + classes combinations
Call of Duty - multiple classes
Guild Wars 2 - 8 classes, 5 races.
Starcraft II - 3 factions
Diablo III - 4 classes (5th coming)
Battlefield 3/4 - 5 classes(?)
Latest Halo Games - added multiple classes
GTA V - Several characters

The list goes on.

Clearly, contrary to Keith Burgun, people ENJOY asymmetry.

- As a side note, I'm stepping away from this conversation. The article author is way too arrogant for me to waste my time discussing his opinions. If someone wants to create an unrelated discussion about the topic then I'll join in. Laters all, enjoy!

Also, don't get me wrong. I think the concept of having a single shared pool of units is a perfectly reasonable direction for strategy. Especially in this circumstance. Lets be honest, who ever played CORE in TA?