"Skill Ceiling"

General Discussion
Yep, It’s Going To Be Long

I see a lot of posts on these forums bemoaning the "lowered skill ceiling" of SC2 compared to BW. Typical evidence for this claim comes as references to the substantially higher APM of professional BW players as well as specific examples of unit control simplification in SC2, such as decreased need to micromanage harvesters, smarter unit targeting AI, and larger control groups. I am creating this thread in an attempt to help players understand the true meaning of "skill ceiling," which I believe might allow the community to appreciate this game more fully and perhaps will lead to fewer complaints stemming from a misguided perspective.

Skill ceiling is actually a very important component of game design. Simply stated, the skill ceiling is the theoretical point where a player has reached perfection within the constraints of the game world. At this point, improvement is, by definition, impossible. Many factors influence the skill ceiling, but typically they fall into two categories: decision-making and execution. Decision-making refers to a player's choice of how to proceed at a given time, and is influenced by the information the player has to shape his decisions, as well as the options at his disposal. A game with many options but little information is essentially flawed and becomes a guessing game (not fun). Likewise, a game with an abundance of information but few options becomes predictable (also not fun).

Once a decision has been made, the next step in game-playing is to execute the decision. This is usually what people think of in terms of the "difficulty" of the game, and is entirely dependent on the mechanics of the game design and the player's ability to utilize these mechanics to the best of his ability. Execution can be made difficult by requiring speed, precision, multitasking, or a combination of the three. The human mind and body have limitations which are tested mightily by a game requiring complex execution. Herein lies the problem: many game players mistakenly believe that physical barriers are the sole determinants of how much “skill” is required to play a game.

Even in its current, imperfect form, SC2 has an impossibly high skill ceiling that will never be reached by a human. The information a player has at his disposal is influence by his ability to scout as well as the current ebb and flow of the metagame. Obtaining additional information is a primary goal of the game. With this information, players have a nearly limitless number of options for how to proceed. These options include: how and where to move units (each unit representing an individual choice to be made and each location on the map being a possible option); how to allocate resources; how to obtain additional information; and what to do if the enemy is encountered. Furthermore, SC2 is a game of execution just as much as it is a game of decision-making. A player must be fast due to the pace of the game (in other words, being slow is a disadvantage), a player must be precise in selecting individual units or groups of units as well as in ordering them to move or attack very specific locations or targets. And of course a player must multitask effectively because input is required in multiple, sometimes unrelated settings within the context of a single game.

The best way to visualize the skill ceiling of a game is to imagine what a game would look like if it was being played perfectly. For SC2, a perfect player would appear to make thousands of actions every second. Each unit would behave as if it was being individually controlled by an omniscient machine. Units with a movement speed and range advantage would literally be untouchable (marines would never be reached by banelings, phoenixes would annihilate mutalisks without being touched). We see some of this occurring even now, but no one is able to simultaneously able to control all of his units in this manner. It is important to note that even with perfect play, a person would not necessarily have complete information on his opponent as he is still limited by his units’ ability to scout. However, he would absolutely have more (if not absurdly more) information than his opponent and he would flawlessly translate this information into a tactical advantage. The world’s current best SC2 player would not occasionally get lucky and win against a perfect player; he would be destroyed every time, without fail.

What I am trying to illustrate here is that the gap between the best professional Korean SC2 genius and a perfect player is massive. We watch replays and gasp in awe at a player’s brilliant marine micro, but even this level of play would be defeated by an army of individually controlled units, all of which were produced at the mathematically exact time their construction became possible. The first couple of minutes may look the same for a pro vs. a perfect player, but eventually there just become too many variables, too many units to control, too much happening at once for a human being to make perfect decisions and execution at a constant rate.

Before I Continue, a Few Examples

Tic-tac-toe: In this game, the information provided is complete; nothing is hidden from either player. The options are very limited, and the execution of a decision is as easy as writing an X or an O. This game has an extremely low skill ceiling, and someone who had never seen the game before could be taught to play absolutely perfectly within minutes, resulting in a draw every time. This is why this game is best played by babies and people who are extremely bored.

Chess: Here we have a game with very simple execution (pick up a piece and move it to a square), complete information, and a fathomless number of options. A 'strategy' game in the truest sense. Player skill is determined solely by his ability to perceive the strengths, weaknesses, and intentions of his opponent and make superior choices. The physical limitations of the human body are irrelevant; only the limitations of the mind stop a player from reaching perfection. Despite being one of the oldest games on Earth, being played by millions of people around the world for centuries, and having a robust professional scene for over two hundred years, no one has reached the skill ceiling for this game. To do so would require foresight that is beyond the capability of human beings and even modern computers (although we do have better-than-human chess-playing computers). Perfect play would require considering every possible outcome of a certain move, and the number of possible chess outcomes is in the neighborhood of 10^120, which is more than the number of stars in the universe times… the number of stars in the universe.

Archery: A game of pure precision. Technically, archery refers to shooting a bow and arrow in any context, but consider the competitive sport. The only information relevant to the athlete is the environment (air speed, direction, and humidity), which has a negligible effect on accuracy in a controlled setting. Could a human reach the skill ceiling of archery? Well, that depends. If the goal of archery was to hit a one-foot wide square from a distance of ten feet, then yes, a person could achieve the skill ceiling by hitting the bullseye every time. However, to actually make the sport interesting, the target is much smaller and much further away. Hitting the target requires exquisite body control and understanding of the trajectory an arrow will follow.

Guitar Hero: Not really a competitive game (I guess you could compete for a high score), but also demonstrates the physical limitations of humans. Can people play this game perfectly? I suppose so, since the songs were written to be performed by humans. Although it requires *insane* coordination, speed, and precision to play the hardest songs perfectly, it can be done. However, I’m sure that I could write a song that would place the skill ceiling well outside the realm of human possibility. You could probably just take the hardest song and then speed it up 2x.

MW2, Halo, Quake, etc.: I’m not implying that these games are created equally, but first person shooters are an interesting case in skill ceiling. A perfect player would never miss a headshot and could instantly dispatch any foe as soon as his presence was detected in any way. The only way to kill a perfect player would be for a team of highly-skilled players to (get lucky?) coordinate an attack in such a way that at least two of them could not be fired upon at the same time. The perfect player could easily take on entire teams of pros at the same time. Clearly, the skill ceiling for games such as these has not even been close to being achieved.

Dammit, Get to the Point!

Let’s go back to archery. The sport was traditionally played using a simple bow or a longbow, devices with limitations for how hard and accurately they could fire an arrow. Recently, the compound bow was invented, allowing archers to fire arrows at a higher velocity, resulting in less time for gravity to affect the arrow’s trajectory and generally improving the accuracy of a shot. Since the bullseye became easier to hit, some might argue that the “skill ceiling” of archery was lowered as a result of this invention (note that I am using this as an example, I do not claim to actually know anything about competitive archery outside of what I learned from watching Robin Hood: Men in Tights). However, unless a person could hit the exact center of the bullseye every time, the distance to the skill ceiling is


Phew, got that out of my system, and hopefully everyone can see where I’m going with this. Sticking with this scenario, the sport of archery has not been corrupted by the compound bow. In fact, the success of an individual is now due even more to his personal skill and precision and is less hindered by the limitations of the bow itself.
This situation is analogous to the position we find ourselves in when comparing BW to SC2. Yes, the newer game automates or intelligently performs many actions that required player input in the older game. However, the gap between the best player and a theoretically PERFECT player is still astronomically high. The skill ceiling of SC2 will never even be approached. Tiger Woods (in his prime) may have been the greatest golfer to ever live, but even he was miles away from the skill ceiling of golf, which would require a hole in one on every Par-3 (etc.).

Due to the less intuitive nature of BW, many people claim that its skill ceiling is higher than that of SC2. I claim that not only do BOTH games have a near-infinitely high skill ceiling, but the entire concept of “skill ceiling” is utterly meaningless as long as it is well beyond being reachable by a human being. Even if someone could mathematically prove that BW had more possible options and required even greater speed and precision to execute, what difference would it make if no one will ever be close to perfect?

In conclusion, I love both of these games, Blizzard, and you, dear reader. But please do not confuse higher APMs of BW with “more skill.” Neither game has a higher skill ceiling, since they are both nearly infinitely high. There is still a huge amount of growing room and perfect gameplay will never be reached.

One More Thing

Also, while I have you, just remember: BIAS is a noun, BIASED is an adjective. A person cannot “be bias.” He is biased. He can have bias though. Of course, I suspect that those who fail to grasp this will not have bothered to read my post. Maybe I should have put this first.
To summarize:

Skill ceiling is meaningless because no player is anywhere near reaching it.

You typed way too much just to say that.
See also: my thread from yesterday


Yep. I thought I would try to really flesh it out though for those who had not previously considered this point.


I thought about discussing the skill floor too but I had already spent 30 minutes typing; thankfully, your post is great. Thumbs up.

or you could have simply posted a link to this youtube video...

Yeah good stuff Willywaas and Cogwheel. This is something people need to read before they comment on this issue.

Having things like, only 1 building per hot-key, limiting control groups to 12 units, no auto-mining from command structures for workers, bad unit pathing/ai and all the other aspects from Brood War do not really raise the skill-cap/ceiling but rather the skill floor. Not to mention they were merely limitations of the hardware/software at the time and were never intended to make things more difficult.

I think we can all agree that having a high skill floor is bad for everyone and wanting a higher skill floor is a pretty elitist viewpoint. Using the Chess example, its pretty damn easy to understand how to play chess, but damn challenging to get to pro level.
08/30/2011 03:50 PMPosted by Shiori
i think the best way to frame the argument about skill ceilings is in the manner that you might call an infinite series convergent or divergent. it's not simply that broodwar converges at a higher value of n (lim n-->infinity) such that all values above differ by essentially zero, but that the variation between n and n+1 is much larger than in sc2. if you were to visualize the bw skill dynamic as an oscillating function that converged at essentially infinity, you'd see what i mean. conversely, the skill dynamic for sc2 looks a little bit more like f(x) = +/- 1/x. the difference between players near the convergence is very small, whereas in broodwar it's huge. a tiny slide toward the origin represents a tremendous change in skill.

Although I haven't had a math class for quite a while, I see your point. In fact, I'm not even going to try to argue with you about it although it would probably be educational for me. The reason for my post was because I take offense when people imply that SC2 play requires less skill than BW play. I believe what you are saying is something like "BW doesn't actually require more skill, but a small increase in BW skill results in significantly greater achievement, whereas a small increase in SC2 skill produces much smaller gains in achievement and is therefore less compelling." Is that right?
08/30/2011 03:50 PMPosted by Shiori
i think the best way to frame the argument about skill ceilings is in the manner that you might call an infinite series convergent or divergent. it's not simply that broodwar converges at a higher value of n (lim n-->infinity) such that all values above differ by essentially zero, but that the variation between n and n+1 is much larger than in sc2. if you were to visualize the bw skill dynamic as an oscillating function that converged at essentially infinity, you'd see what i mean. conversely, the skill dynamic for sc2 looks a little bit more like f(x) = +/- 1/x. the difference between players near the convergence is very small, whereas in broodwar it's huge. a tiny slide toward the origin represents a tremendous change in skill.

With the higher skill ceiling, it seems like you're narrowing the field of competent/skilled players. Makes sense. However, read the OP again. This does not mean the skill cap is higher, it just means there are fewer pro level players. Sounds like a bad thing to me.
People complaining about "skill ceiling" are almost invariably whining about the "skill floor".

As verbosely stated, :) Sc2, specifically, and RTS, in general, have virtually infinite, and thereby unobtainable skill ceilings. Making some things easier doesn't noobify the game it enhances what can be done in a limited time.
I'm pretty sure nobody who complains about SC2 having a lower skill ceiling, while they are indeed using poor terminology because technically the skill ceiling in SC2 is impossibly high, is really talking about what you're explaining here. What they really mean is that SC2 is effectively an easier game in terms of mechanical execution than BW: that, in other words, SC2 has a less steep skill gradient. Because BW has a steeper skill gradient than SC2, in SC2 practicing mechanics more than somebody else confers fewer benefits than it does in BW. In other words, some "noob" who practices less than a pro will be more likely in SC2 to beat the pro than would be the case in BW.

In this sense, one could say that SC2 is less competitive. I choose to think that SC2 will simply end up revolving more around strategies than mechanics, since the mechanics part is much easier, thereby freeing up more options in terms of strategy.

What this means for SC2 is that we are less likely to see consistency in terms of players that excel in mechanics and more consistency from players who are able to perceive their opponents' strategies and to make good decisions. While mechanics are of course still critical in SC2, I think it will be apparent that once everyone achieves more or less top-notch mechanics (which, again, will be much easier to achieve in SC2 than in BW), we will see strategy start to trump all. This is partially why I'm unconvinced whenever people say that if Flash or Jaedong switched they would simply be untouchable.

Join the Conversation

Return to Forum