I want one of those pony titles so bad.[“Play Evolution” is a column by James Lantz that happens sometimes and discusses the changes that games undergo after their release, from little developer patches to huge gameplay revelations, and everything in between. This week: evolution through goals in Modern Warfare 2.]

FPS games have undergone a weird, subtle, and totally pervasive change in the last few years. Every big budget FPS now cocoons the actual shooting part in an MMO shell. It's hard to tell where this trend started, but it's been increasingly present in FPS games in recent years – most notably in the Battlefield series and the Call of Duty/Modern Warfare games. The now combination MMO-FPS nature of these shooters brings up two major questions: what made this MMO-structure so quickly envelop every major FPS, and what effect does it ultimately have on the games that choose to use it?

The simple answer to the first question is that MMOs are incredibly popular. The carrot-on-a-stick leveling system gives people constant short-term goals and a concrete long-term goal and keeps people playing. It also rewards time as much as it rewards skill, so casual players or less skilled players will never actually go down in rating and get discouraged, as they could in Chess, Go, Starcraft, or other games that use a ELO-style rating system (which adds or subtracts points from your rating depending on your rank relative to your opponent's rank and whether or not you lost the game).

In an MMO, you're always making progress, you always have an easy short-term goal and a distant long-term goal. Many of an MMO-system's advantages over a skill-based system are obvious – for instance, it helps keep casual players playing and it helps players create short-term goals that they can complete in a single sitting.

It also has some less-obvious advantages. For one, I would guess that it also helps keep hardcore interest as well, because it helps players who play the game obsessively to have a more concrete goal in mind than "getting good." For some players, getting good works just fine as a long term goal - but it can often create a sort of trap for players who are hardcore enough to take the game seriously and want to play competitively but aren't quite skilled enough or don't have quite enough time to devote to the game so they just stop playing.

These are all good reasons for FPS to adopt the MMO-style system, but where did it come from in the first place? To answer this question, and to shed some light on the other questions I've asked, let's take it out of the specific case of MMO-style goals for Modern Warfare and Battlefield and into the more general case of how goals affect how we play games and how, therefore, gameplay evolves.

Goals and Pseudo-goals

A game's goal provides a greater context for each individual decision, and ultimately is the driving force behind how people play the game. However, people don't always have the game's goal in mind when they make each individual decision – in most cases, people use pseudo-goals which stand in for the game's goal but are much more useful when discussing strategies or making split-second decisions. For example, the goal of Starcraft is to destroy all your opponent's buildings. Moment to moment, however, most people aren't thinking about how they can destroy all their opponent's buildings but thinking about how they can destroy their opponent's army, or their opponent's production capabilities.

These are pseudo-goals: handy stand-ins for the actual goal, because most of the time the actual goal overcomplicates things and isn't useful to consider. In a game where time is an important resource, it's much more efficient to think about how many zealots you need to build to match your opponent's zergling-lurker ratio so you can kill his army and win than it is to think about how many zealots you need to build to match your opponent's zergling-lurker ratio so you can kill his army and then have an army left over and then go into his base and kill all his buildings.

In Starcraft, these two pseudo-goals are usually identical to the actual goal of the game. 99% of the time, when you kill your opponent's army, he's going to surrender because he can read far enough ahead to see that you're going to raze his base in five minutes and he's helpless to stop you. So why even talk about these pseudo-goals? What difference do they make? Well, what makes them important is the other 1% of the time, where you end up sneaking your armies to each other's bases without fighting, and the real goal of the game – to kill all the other player's buildings – actually comes into play. It's these 1% of games that make the pseudo-goals players often use not equivalent to the game's actual goals.

In fact, some of the most clever strategies are those that subvert the pseudo goal and remind everyone what the real goal of the game actually is. For example, Warcraft 3's goals and pseudo-goals are basically the same – you're trying to destroy your opponents base, but you basically win if you can take out his army and keep yours intact, so the pseudo-goal has become to win in a army vs army fight. However, the trickiest strategies bring the game back to the goal of killing your opponent's base. One such strategy involves building strong siege units and making them invisible, and then razing your opponent's base while he still has an army but is unable to see and counter your units.

Another, more popular one involves building a massive amount of defensive towers around your resources and building very few units, essentially turning the game into a test of patience that can last hours. This transfer from pseudo-goal to real goal shows just how much even a slight change in the definition of a game's goals in the player's head can change the way the game actually plays drastically. This is true even if these strategies end up having hard counters – they still affect the metagame.

For example, a player might always buy invisible-revealing items after the invisible-seige strategy becomes popular, which makes it completely pointless. However, the popularity of the strat has made buying invisible-revealing items an interesting choice – the strategy is no longer popular because it's so easily countered, so do you really need to buy that invisible-revealing item, or can you save the money and take the risk but end up ahead if your opponent buys one and you don't?

His K/D is going to be so highStrategy games aren't the only games with pseudo-goals, however. First person shooters (FPS) have them too, and they're equally important. In a strategy game, you can measure individual skill by wins and losses because it's usually only played with two players. Most strategy games extrapolate those wins and losses into a fairly accurate ranking, which becomes an explicit, overarching goal for most players. FPS games, however, are often team games – so it becomes much harder to give an accurate measure of individual skill. An excellent Counter Strike player might lose several games in a row because he's on a team with several bad players, or a bad player might win several games in a row on a team with good players. So unless you take an average of your total wins and losses over a long period of time, it's hard to get an accurate measure of individual skill. However, most people like to see how good they are, or if they're getting better. They want a goal equivalent to ranking in a strategy game. Enter the Leaderboard.

In Deathmatch modes, the leaderboard is equivalent to the game's real goal. Because the game is based entirely on kills, and the leaderboard places people in order of kills, being on the top of the leaderboard at the end of the game always means you've won. So the leaderboard bring a way to measure individual skill into the game. It gives players a tangible pseudo-goal (topping the leaderboards) that is usually equivalent to their actual goal (helping the team win). However, the two goals aren't always equivalent in non-Deathmatch modes.

In Team Deathmatch, for example, the leaderboard is a pseudo-goal that's very close (but non-identical) to the actual goal of the game. In Team Deathmatch, the goal of the game is to reach a certain score before the other team does. You score by a point when your team kills an opponent and vice-versa. In most games, leaderboards are arranged by pure kill count – that is, the player who has killed the most enemies is the highest. However, your net value to the team in Team Deathmatch is usually not your pure kill count but your kill/death ratio. For example, a player who has 12 kills and 1 death has scored his team 12 points but lost his team 1 point by giving the enemy team a point for killing him, so he has a net value of about 11 points. A player who has 23 kills and 30 deaths, on the other hand, has a net value of about -7 points, but is ranked much higher on the leaderboard than the other player.

So, most of the time, your kill/death ratio is a more accurate measure of your individual skill than your pure kill count. However, even kill/death ratio is a pseudo-goal – a skilled player might play a scouting role, figuring out where the opponents are by risking himself and therefore ending up with a negative kill/death ratio but giving his team valuable information. Another player might get one or two kills without a single death, and have an infinitely large kill/death ratio but not actually help the team out that much. Despite these irregularities, leaderboards are often a fairly accurate measure of individual skill. The real goal of any given game of Team Deathmatch is still to have your team to win, and a loss is still a loss, the leaderboard still gets wiped at the end of the round.

However, what happens when we introduce stat-tracking across games, and make your total kills and deaths public and permanent? Suddenly, instead of leaderboards being a rough attempt at measuring your skill, they are a measure of a second explicit goal itself – and not only that, but a second goal that can contrast with the first goal, as in the examples talked about above. So, instead of hierarchy like in Starcraft or Chess, where there is one overarching goal and a win is always a win despite possible pseudo-goal losses, you have two real, explicit goals that often conflict and end up with players playing what are really two completely different games in the same space.

Some players are playing and trying to maximize their kills or kill/death ratio – playing the leaderboard game – while other players are just trying to do everything they can to help the team win each individual game – playing the "original" game. In fact, one player can often switch between playing the two – sometimes trying to help his team win, maybe in more competitive games –and other times simply trying to improve his overall stats, even at the cost of his team's victory. One interesting thing is that even though these goals will sometimes produce contradictory optimal strategies (getting one kill and hiding might be great for your kill/death ratio but doesn't really help your team), neither one is really wrong. Both are explicit, real goals. Neither is a pseudo-goal. Basically, these people are playing two entirely different games.

Why Goals Matter

This argument raises several questions. First, how much do slightly different goals really change a game, and second, what are the ramifications of having two games being played in the same space? To the first question, I would say absolutely yes. A small modification of the game's rules may look superficial at first, but can lead to cascading consequences in the metagame that ultimately change the way the game's played entirely. I think it's easier to see if you imagine an even smaller change than changing the goal. For example, imagine changing a marine's health in Starcraft by 5 points. At first, this change seems so small. However, it might allow a new strategy that was previously not viable before, which would then create another strategy which was ridiculous before but countered the now-viable strategy, which would lead to a new counter-strategy to that, and so on, creating a completely different possibility space of strategies.

For a more practical example where changing the goal slightly actually does effect the game significantly, look at the difference between strategies in tournament poker and cash-game poker. Everything besides the goals of the two games is exactly the same, yet playing tournament poker is very, very different from playing cash-game poker. For another practical example, visit the message boards of any game that has stat tracking and see how many arguments you can find about whether win/loss record or kill/death ratio or assists per game is the most important stat in determining individual player skill.

After each match, the game rewards you for your achievements with 'acolades,' which you pretty much always getBefore we discuss the question of ramifications, let's bring this back to our specific example of the modern FPS. In particular, let's talk about the biggest modern FPS out there that embraces the MMO-style system like no other, Modern Warfare 2. Modern Warfare 2 has stat tracking separate from both the MMO-style leveling system and the individual games, so it still creates the two often contradictory goals of having good stats or winning games, and has, therefore, two games being played in the same game space. However, the MMO-style leveling creates a very large third goal which is not equivalent to either of the previous two.

In fact, the MMO-style leveling is, in some ways, an evolution of the original leaderboard – it mostly takes into account pure kills and does not subtract points for deaths at all. So, players who are ultimately trying to optimize their climb up the ranks will be playing an entirely different third game, where the goal is to get the most kills as possible at the cost of everything else, which often clashes with both the kill/death ratio goal of permanent stats and the individual game-instance goal of having your team win (even though there are bonus points for winning, the two goals aren't always the same).

So what are the implications of these three different games being played at the same time in the same game space? Well, there are several problems with it. First, it splits the community and makes it hard to transition from one sub-community to another. In Call of Duty 4, competitive play is radically different from casual play, because all competitive play is about winning individual matches, while most casual play is about maximizing your kill/death ratio. Someone who wants to go from casual to competitive has to actually relearn a lot of things to understand how competitive players play. By contrast, being competitive in Chess or Starcraft is a smooth transition - casual players play the same game as competitive players, so if you get good enough playing casually you simply become a competitive player by definition.

Second, it severely limits the practical value of trial and error in become a good player. When you start playing Chess, you can become better much more quickly simply by playing possible moves that look good over and over and seeing which ones end up with you losing and which ones end up with you winning. In Modern Warfare 2, players will often have several different goals in the same game space, so it's difficult to actually see whether what you're doing works or not towards a particular goal. Finally, it's just frustrating if you're trying to play the game in which the goal is to have your team win the individual game instance and other people are playing poorly towards that end but very well towards another end, like having a high pure kill count and leveling quickly, and vice-versa.

Even aside from the dissonance of having several games played in one space simultaneously, the MMO-style unlock system creates other problems, most of which can be boiled down to its time-based nature (as opposed to the skill-based nature of other goals). Because one of the main goals in the game is time-based, players aren't rewarded as much by the game for improving in skill as they are by simply playing. Whether this is really bad or not, however, is a debate for another article.

So what are the net effects of the MMO-style system on the evolution of a game? It creates a strong sense of progression – people in general definitely seem to like it – but maybe it affects a game's evolution problematically. It's good because it's addictive like an MMO – it gives you bite-sized chunks of what would otherwise be a difficult learning curve to competitive play, it lets you feel accomplished after every session, and it makes you feel good like unlocking an achievement makes you feel good – after nearly every game. How can more stuff to do be bad? But it's possible that Infinity Ward is sacrificing a lot of clarity about the actual goal of the game for these flashy, constant rewards – and with that clarity, maybe, ultimately, deeper play.

If you own Modern Warfare 2, and you play the online multiplayer, consider your own experiences. When a game is over, you probably feel good when your team wins and you probably feel less good when your team loses. But that's probably only a part of what really matters. You're probably also thinking about whether you got a good kill/death ratio and where you are on the leaderboard. You're probably hoping you got a sweet title for that kill to put on your callsign. I know I've often been excited to get a really cool title – much more excited than I would have been just to win the game. So, in the end, what's the real goal? To win the game? To get a good kill/death ratio? To earn titles? Which game are you actually playing?

[James Lantz is a starving writer/designer who has played more MW2 in the last few days than is healthy. He also writes a blog, of course.]