Ambiguous developer intention in fighting games

This is a topic that really grinds my gears: Ambiguous intention from developers in fighting games. I believe this sort of thing is very harmful to fighting games, and I'll be using this article to explain what in particular I have against ambiguous intention, provide some examples of ambiguous intention in several fighting games, and how I think those games would be better of without such things.
First off, let me give a solid, simple definition of ambiguous intention. Basically, it is an instance in a fighting game where a restriction the game's mechanics impose on the players can be avoided via use of one or more mechanics/techniques the game provides. There are several issues with such a thing, in my opinion:

1. It adds an artificial execution barrier to the game. Anyone playing the game at a competitive level will have to manually perform something that would otherwise (and should) be done automatically or not at all.
2. It makes the developer's intent unclear - do they want the restriction to be respected or not?
3. It makes the game seem harder or "deeper" than it actually is.

The first example of unambiguous developer intent that comes to my mind, not to mention one of the most generally-accepted in the fighting game community, is nicknamed "chicken guarding" in the vs. Capcom series. Normally in those games, during a short period after you initiate a dash, regardless of what direction it's in or whether it's in the air or not, you are unable to block. Clearly, the developer's intent is to make dashing punishable if predicted.

However, it is possible to instantly cancel a dash into a jump in most games in the vs. Capcom series, and after you do so, to instantly block. This overrides the previously-implied restriction about punishable dashes. Now, I have absolutely nothing against making vs. Capcom dashes punishable or unpunishable; what I do have a problem with is the ambiguous developer intent: Do the developers want dashes to be punishable or not?

If the developers do want them to be punishable, then disallow blocking even after the dash-canceled jump, or if they do not want them to be punishable, then allow blocking directly during a dash, but by no means should they require the players to go through an unnecessarily-difficult input to accomplish a competitively mandatory technique which may or may not have been intended in the first place!

The second example, and one that I think is particularly egregious, is L-canceling in Super Smash Bros. Melee. L-canceling is a "mechanic" - I am not entirely certain it is deserving of the title - that allows you to cut the period of time your character cannot move or attack following an aerial attack in half by pressing the shield/block button a few frames before you hit the ground.

With the previous example of chicken guarding in addition to the 3 problems with ambiguous developer intent I listed earlier in mind, it should be immediately obvious what the problem with L-canceling is. If the developers want aerial attacks to have x amount of landing lag, then they should leave it at that; if they want aerial attacks to have 0.5x amount of landing lag, then they should do so, but they should not require players to press an extra button for something that could easily be done automatically. Artificial "difficulty" like this makes a fighting game seem more difficult and technical than it really is, and oftentimes hides serious problems behind a fake execution barrier.

The last example I'm going to give is of "kara throwing" in the Street Fighter series. Kara throwing essentially involves using an attack that moves your character forward, but before the part of the attack that actually hurts the opponent is activated, you cancel the attack into a throw, giving the throw more range than it otherwise would have. The window of opportunity to cancel the attack into a throw is usually very small - normally about 1 or 2 frames (1/60th to 1/30th of a second).

Again, it should be obvious what the issue with kara throwing is. If the developers want certain character's throw range to be longer, then they should be longer by default, or if they want throws to be shorter, then they should prevent kara canceling, but they should not require players to perform a difficult input for something which was not only unintended in the first place (kara throwing was originally a glitch in Street Fighter 3, but is intentionally in Street Fighter 4), but which obviously could be solved with a much simpler solution. 

I've seen several arguments stating that mechanics like the ones I've listed are good things, the majority of which basically boil down to the opinion that the more buttons you have to press and the more complicated things are, the better. Personally, I believe the simpler things are, the more time and energy you can focus on the mental aspect of the game, which, in my opinion, is far deeper and more satisfying than any arbitrarily difficult "mechanic" could ever be.
It could also be argued that since no-one really complains about arbitrary difficulty and ambiguous intent, it is not a problem that needs to be solved. I believe there are a majority of reasons this is the case, the largest being that since practically all fighting games ever made have had some sort of arbitrary difficulty, with the recent ones intentionally and the older ones unintentionally, means that anyone who has played fighting games for any decent amount of time is quite used to this sort of thing by now. So yes, there is some complacency, but I think many fighting game players have merely failed to realize how deleterious ambiguous intent and arbitrary difficulty can really be, since they have had to live with it for as long as they have played fighting games. I think once more players realize how much simpler and easier to play fighting games would be without this sort of thing, the more they will warm up to the idea and let developers know about it.

I should note that I think there is a definite place for this sort of mechanic in video games; intentionally difficult single-player games like Devil May Cry and I Wanna Be The Guy thrive off this sort of thing, and many people play those sort of games just because of all the arbitrary difficulty - I'm one of them. However, I don't think arbitrary difficulty has a place in fighting games; I'm trying to fight my opponent, not the game, and the easier it is to do whatever it is I should to in a given situation, the better.

I think I should also make it clear that the presence of an arbitrary mechanic does not automatically ruin a game. I enjoy playing and watching many games that have a number of arbitrary mechanics, including the games I used as examples. All I'm saying is how much better those games would be than they already are without the arbitrarily difficult mechanics, in addition to a larger userbase, since a (somewhat correct) public notion of fighting games is that you have to spend a lot of time practicing arbitrary stuff in training mode before you can become good at them.

And you know what? Eliminating ambiguous developer intent and arbitrary difficulty would go a long way towards eliminating that notion.

Super Smash Bros., Combos, and You - an article on a theoretical combo system for SSB

Super Smash Bros. has something of a mixed history as far as combos go. The original SSB64 had combos galore - most of the characters had combos that could take an as-of-yet untouched opponent and KO them without a single chance to dodge or counter - the definition of a "true" combo. Super Smash Bros. Melee significantly reduced the stun the average move does to your opponent, so combos (true combos anyway) became weaker and shorter compared to SSB64.

Now, with the latest release in the SSB series - Super Smash Bros. Brawl - true combos are practically non-existent, meaning the guaranteed damage the average character in the cast can do is typically limited to one or two moves in a row. This may not sound important, but it has several ramifications:

• Lack of combos means mistakes aren't as costly. In the typical fighting game (SSBM and SSB64 included), 2-5 mistakes might mean a stock (or the game). SSBB, on the other hand, typically lets you get away with more than a dozen mistakes before you lose.
• The skill ceiling is significantly lowered. SSBB pros can punish mistakes no harder than intermediate players, with the possible exception of what's called "gimping" - KO'ing your opponent with relatively weak move, usually used away from the stage edge.
• Comebacks become a lot harder, since you'll typically have to repeatedly use one-off attacks, while only a few attacks (many times just one well-placed one) from your opponent will cost you the game.

What I suggest is to give SSB a combo system/mechanic. I'm not going to describe what I have in mind for the system upfront - rather, I'll evolve it over the course of this article, which I think will lead to a better understanding of how a "true" combo system would significantly help SSB and alleviate the aforementioned issues Brawl has. So let's get started!

The first step is to see what other fighting games use. The combo mechanic you most often see in other games is this thing called canceling. Basically, canceling let's you literally cancel one move's animation frames into an entirely different move altogether. Depending upon the game, you can only cancel a move if it touches the opponent (regardless of whether or not it does damage), and for the purposes of this article, we'll assume that's the case for our combo system. Additionally, you can typically only cancel regular attacks into heavier regular attacks (think SSB's standard attack/auto combo) or special attacks.

So now we have our base mechanic. How do we apply it to SSB? Let's assume we can cancel a standard attack into a tilt attack or special attack, we can cancel a tilt attack into a smash attack or special attack, and we can cancel a smash attack into a special attack. So now, every character has this basic combo: Standard attack > tilt attack > smash attack > special attack.

This is pretty good already, but this system suffers from the same issue Brawl has: There's little - if any - room for improvement over the simple combo I just mentioned. The skill ceiling issue remains basically untouched. We need to figure out a way to allow creativity within our little combo system. What can we do to fix this problem?

We have two options: We could further extend the boundary of what can or cannot be canceled, or we can come up with another method of extending combos. The first method will simply lengthen our previous combo, while the latter's benefit will be dependent upon how versatile the introduced sub-mechanic is. We'll try the latter approach, since it helps alleviate the skill ceiling issue Brawl suffers from. So without further ado, I give you: The launcher.

Simply put, a launcher is a move that, like the name implies, launches the opponent up into the air. You can also typically cancel the launcher's animation with a jump if the launcher hurts the opponent, and for our combo system, we'll also allow launchers to be canceled into a dash, which should allow for a bit more creativity as well.

We have to integrate launchers with the little cancelling hierarchy we defined earlier now. Let's insert launchers between smash attacks and special attacks, so that our standard combo now looks like: Standard attack > tilt attack > smash attack > launcher + aerial attack or special attack.

It's not much, but this combo system is already shaping up to be relatively open-ended. However, since we've been focusing on adding new features to it, we've been left with a lot of loose ends to tie up:

• Knockback (not to mention DI) will effect the viability of a combo - i.e, an opponent might be knocked back far enough that even if you cancel, the new move will not connect. Something needs to be done about this.
• Hitstun - the amount of time the opponent is stunned and unable to do anything after being hit with one of your moves - will probably need to be increased on average, at least relative to Brawl. Otherwise, even if you perfectly cancel a move into another move, your opponent might recover before the second move's active frames activate.
• It's conceivable that the combo system could possibly be abused to the point of infinite combos. We obviously can't think of all possible combos, so we should put in some sort of mechanic that will prevent combos from becoming excessively long.

Let's tackle knockback first. We obviously can't take knockback completely out of the system - it's too much a part of SSB to remove it now. So I propose that the hitlag system normally reserved for dramatic, powerful attacks be applied to all moves in the game, with about 5-10 frames of hitlag per hit. Those frames will be the "cancel window" during which you can cancel to another move and be guaranteed that it will connect - no knockback will be applied until after the cancel window has passed. Launchers will have set vertical knockback to make comboing easier, similar to Falco's shine in SSBM.

For those not familiar with the hitlag (also known as "freeze frame") system, hitlag is basically a period of time after which a move connects with an opponent during which the game is, for all intents and purposes, frozen - neither the attacker nor the defender can move during the hitlag period.

This doesn't address the tricky problem that is DI (Directional Influence), however. It's possible in SSBM and SSBB (SSB64 didn't have DI) to use DI to shift the direction in which you get launched and thus escape combos. Whether or not that's a good thing probably comes down to personal opinion, but DI is here to stay in all likelihood, so I'd prefer to design our theoretical combo system to incorporate DI.

The way I see it, the easiest way to incorporate DI into our combo system is to allow DI only after the cancel window on each hit has passed. This brings up a dilemma: What about multi-hit moves? I want to keep comboing as offense-oriented as possible, and thus have relatively few options - if any - to break out of a combo, so I think having a cancel window after each hit, but no knockback and DI until after the cancel window for the final hit has passed is a good idea. Launchers probably shouldn't be DI-able for similar reasons, so let's assume that we can't DI out of a launcher as well.

Now we only have to deal with infinite combos, something a lot of fighting games don't get right. Even the borderline combo-less SSBB had a few infinites here and there, so we'd better take extra caution with our combo system, since it's a lot easier to chain multiple moves back-to-back. Again, taking a look at what other fighting games typically do to alleviate this, I now introduce: hitstun deterioration.

In a nutshell, hitstun deterioration gradually lowers the amount of hitstun of each successive hit in a combo. Eventually, the hitstun of each hit will be practically 0, and the opponent will be able to easily escape the combo. It's obvious that hitstun deterioration could possibly limit legitimate non-infinite combos, as well as put characters with lots of multi-hit moves at a disadvantage, but it's probably the best defense against infinite combos out there right now.

Let's take a quick review of what our combo system's main mechanics are (this is a good TL;DR paragraph if you skimmed over the other parts of the article):

• Canceling. We can cancel moves into certain other moves in this order: Standard attack > tilt attack > smash attack > launcher + aerial attack or special attack.
• Cancel windows. There's a small window of opportunity after each hit of a move (about 5-10 frames), during which you can cancel the current move into another move, as per our canceling mechanic defined earlier. No knockback is applied (and DI is impossible) until after the cancel window is over.
• Launchers. We can perform a launcher, which is basically a move which launches the opponent into the air with set vertical knockback, to perform an aerial combo. If the launcher connects with the opponent, you can cancel the launcher into a jump or a dash. You cannot DI out of a launcher.
• Hitstun deterioration. The hitstun of each successive hit in a combo gradually decreases to almost 0, thus effectively preventing infinite combos. Note that this doesn't prevent a zero-to-death combo, just infinite combos.

So there you have it. We evolved a theoretical combo system around SSB's unique mechanics, ironed out some minor kinks, and even introduced a few unique mechanics of our own along the way. It's probably not perfect, and might have some errors I've overlooked while I wrote this article, but it's a simple, solid combo system that should be enjoyable by newbies and professionals alike. So that said, it's time for closing statements. :)

• Project M, a very popular mod for SSBB, includes a combo system for Lucario very similar to the one described here. The most notable differences is that the cancel window system described here is missing, and Lucario's launcher (his up smash) does not have set knockback, meaning it's better as a standalone kill move, but somewhat hard to use at times as a launcher. Hitstun deterioration is also missing, and all his combos are DI-able as far as I know, but the system is still quite similar to the one in this article.
• It would be feasible to make a mod for SSBB that implements our theoretical combo system, but although I'm not lacking in programming skills, I'm still quite unfamiliar with modding SSBB in general, unfortunately.
• SSB4 will be coming out sometime next year (2013), and Masahiro Sakurai (SSB's creator) has stated that the game will be taking a "new direction" compared to previous SSB titles. Namco - who develops the popular fighting game series Tekken - is heading development of SSB4, so who knows, you might see something like this in SSB4.

Thanks for taking the time to read through this giant wall of text!