Aztez – A Game of Conquest and Brutality

The Mash Flow

Mash flow is the very simple concept of what exactly your fingers are doing while you're fighting a group of enemies in a beat 'em up game. There are a handful of integral types, and while there are many subtle strains of these types, I'm going to break down and scrutinize the big boys.

Type 1: Super Traditional

I call this Super Traditional because the very first generation of beat 'em ups (specifically Taito's Renegade and Technos' Double Dragon) took a two button approach. The input scheme was identical; one button was for punch and one button was for kick. Either button would start a very straightforward combo that you could either see to the end, cancel by ceasing input, or get hit out of. It is very important to note there was no interplay between the two basic flows, it just seemed like it because you could start one combo in the middle of the other. What's fascinating to me about the manifestation of beat 'em ups and their early use of this very rigid style of mash flow is that in all fairness, every attack in the combo past the very first attack was just a satisfying formality; you could design the same game without even having combos. But they did it anyway because it was incredibly satisfying to players. So much, in fact, that we've been iterating on this basic formula for 25 years.

Type 2: Traditional

It didn't take very long for game designers to realize they were misusing an entire button with a trivial redundancy so they smashed the game's important mechanics into one button. Thanks to this decision, the second generation of beat 'em ups (Final Fight, Streets of Rage, Golden Axe, all the way up into the later era of Aliens Vs Predator, Battle Circuit, and Armored Warriors) were much more elegant and had the players navigating the game environment, staying clear of attacks, and then pinpointing the right moment in which to execute one very useful string of attacks by mashing one button. Eventually the genre developed a "special attack" button that would execute a unique mechanic that was very powerful but also very limited in use. Before the implementation of this ability on its own button, most beat 'em ups let players perform this by pushing the attack and jump buttons at the same time. This was too difficult for the average player to do reliably, hence the move to its own button.

Type 3: Gap Timed

One of many concepts the first Devil May Cry introduced into the world of beat 'em ups is the idea of branching live combos into completely different combos by waiting at key points in the combo for a split-second before hitting the attack button again. For example, pressing A four times in a row would get you a fundamental standing combo, but pushing A twice, waiting for the second attack animation to approach completion, and then pressing A two more times would get you a completely different combo. The benefit of this method of mashing is that it's varied but elegant. The downside is that most players struggle with the execution of this type of combo and give up on it, which is understandable.

Type 4: Circuit Jumping

Heavily popularized by Team Ninja's Ninja Gaiden and Sony Santa Monica's God of War, beat 'em ups eventually returned to a two-button input scheme that offered "light attack" functionality on one button and "heavy attack" functionality on the other. Admittedly, this can feel a lot like the Super Traditional mash flow but there are two key differences. The first key difference is that the attack types don't just look different, but they feel different and have different effects. For all intents and purposes, attacks on one button in Super Traditional mash flow games are functionally identical to attacks on the other button, they just look different. The other key difference is that modern beat 'em ups have formalized the dynamic use of the two buttons so that it is still considered one combo if the player decides to jump into the other circuit. This was certainly not the case in Super Traditional games, which is another reason the second button was ultimately useless. Circuit Jumping mash flow allows the player to switch mid-stream between the light and fast combo they have started to the slower and more powerful combo they want to finish with, or vice versa. Once they have jumped, however, they must follow through to the end of the new circuit.

Type 5: Super Flexible

This super modern and very newly employed mash flow is an incredibly flexible and intuitive flow that allows the player to switch back and forth between attacks types on two different buttons whenever they want as a dynamic response to whatever is happening around them. As far as I know, the only games that have implemented a mash flow like this are Bayonetta and Dante's Inferno. To be fair, Bayonetta is much more nuanced than this diagram would suggest, but for the most part the idea is on the mark. While it does have the potential to be very exciting and expressive, the danger with this mash flow (as evidenced by both Bayonetta and Dante's Inferno) is that your attacks can lose meaning. I say "danger" and not "problem" because I don't believe this issue is inherent to the flow type, but to the game. In Dante's Inferno, none of the standard attacks on these buttons are particularly meaningful or different from each other. In Bayonetta, the problem is fascinating; there is such a vast amount of meaningful attacks and combos that many of them are rendered meaningless by our human habit of "sticking to what works". I firmly believe this idea can be wrangled in to create something that feels dynamic but doesn't get turned to mush by the player's brain when they are knee-deep in an encounter.

I wrote this post to create new language. If you started using these terms to describe the types of combos you like and want to implement it would inject joy directly into my biological joy parser. :)

  • Pingback: Beat ‘Em Up Sheet Music « Aztez Development Blog()

  • Pingback: Combat Analysis: Bayonetta « Aztez Development Blog()

  • Pingback: Combat Analysis: Lollipop Chainsaw « Aztez Development Blog()

  • Billy

    The problem with “Super Flexible” may be more than people sticking with what works. There is also the factor of memory.

    Crimson Tears (PS2, made by DreamFactory) used two buttons with combo trees (so the third hit of AAA might differ from ABA which might differ from BAA. Or they might not.) Scattered within those trees were a few moves that either acted as launchers or guard breaks.

    I tended to stick with what worked. If for some reason that chain didn’t work, I found I might have to experiment a bit to find the right combination for some other effect.

    To be fair, there were several reasons for this effect in this particular game. First, there were three characters and each weapon type had its own tree. Some trees were short (Kadie’s Buster Sword was maybe three moves long) while others were long (either of Amber’s dual wield trees). And the game used a weapon degrade/break system, so you sometimes found yourself out of necessity using a weapon type that you didn’t like and didn’t remember. And, bluntly, most of the attacks in the chains were effectively interchangeable. The only things that tended to stand out were the few special property moves. Maybe different attacks did different damage or had different wear or produced different heat, but if they did, the game never communicated that info to the player. And mashing worked. The player didn’t really have an incentive to memorize trees. Heck, you could even afford to experiment to find some other chain if you ever needed to. (For example, you could keep a guy blocking until you stumbled across a guard break. He might even open himself up in the meantime.)

    As for “Circuit Jumping,” would you include Dynasty Warriors combat system in that category? Normal and Charge attacks, and you can jump from Normal to Charge (but not the reverse.) The specific Charge move you get is based on how many Normal attacks you did before it. (This system was changed for the worse with Dynasty Warriors 6, being converted into a simpler system with less risk and no button counting.)

    • First of all, I am familiar with Crimson Tears and I’m delighted you’ve referenced it to make your point. I didn’t think anyone else knew it existed!

      Unfortunately, I’m not quite sure what you’re trying to say. But just so we’re clear, I do NOT like having to memorize lots of arbitrary combos. From what little I recall of Crimson Tears (while it’s a fascinating game, I don’t particularly enjoy playing it and so only put a couple hours into it), you are definitely incentivized to memorize trees because specific sequences had properties that were very useful, like you mentioned. So while I believe having attack mechanics that are useful because of their special properties, I believe in making them accessible at all time, which is why in Aztez you can link from any attack in a sequence into these other distinct attacks.

      And I’ve mostly unfamiliar with the Dynasty Warriors series; I’ve dabbled for brief moments in the distant past but didn’t find them very fun. Which one would you recommend? I will gladly play the one that best represents the series. :)

  • Mr. Strange

    Dynasty Warrios 2 was the cleanest in combat execution – the others only make sense in reference to the combat system DW2 introduced. (DW1 was a fighting game – pretty much an entirely different game, though it shared some thematic elements.)

    Your description of different combo types is very interesting to me, as a designer of fighting games, because it’s almost completely upside-down from the way I actually build input and animation trees.

    Also, check out the Godzilla series (Godzilla: Save the Earth on XBOX is the best) for an good example of what you call “Super Flexible” combos from as early as 2002.

    While you’re at it – might as well point you at my kickstarter for a new Kaiju game…

  • Glowyrm

    Dynasty Warriors was just way too hack and slash. That caused my brain to disengage and just get completely bored. I haven’t come across a game that I think successfully tackled the large one person fighting against an army combat system in a meaningful way. Though I have always held out hope.

    Just for clarification, many games, such as Devil May Cry, have done a good job with combat for groups of enemies, but I feel that it’s a different game in scope than a game where you’re fighting in the middle of two armies.

  • Pingback: A Combat Accessibility Fork « Aztez Development Blog()

  • Pingback: Combat Analysis: DMC « Aztez Development Blog()

  • Pingback: Combat Analysis: Metal Gear Rising « Aztez – A Game of Conquest and Brutality()

  • Pingback: Combat Analysis: Ryse Son Of Rome « Aztez – A Game of Conquest and Brutality()

  • Gregory Livingston

    Been rereading this one a bit lately, and it just hit me. A friend of mine, playing Streets of Rage 2, came up with a strategy that had never occurred to me; he punched an enemy multiple times, but instead of going for the finishing hit, he paused. The enemy would stay stunned in the meantime, and then he’d start punching again before the enemy left the stunned state. By pausing instead of landing the combo finisher, he kept the enemy close by (whereas the finisher would hurl the enemy far away). It seems like this is a gap timed combo, right?

    • Not formally, no! That’s just an improvisational exploit that exists in most combo systems. Gap timed combos are implemented (i.e. intended) to provide access to additional options at key points in pre-designed combo.

      It’s a legitimate tactic, though! But the onus is on the designer to make sure that it’s not an unbreakable technique by properly tuning the stun states on enemies and/or making sure surrounding enemies can break that potentially infinite string of attacks.

  • Pingback: Canto: O Monster Hunter do Canada - BlogMMO()

  • Pingback: PAXtravanganza 2014! - Aztez | Theology Gaming()

  • Laura Marx

    Sengoku 3, developed by Noise Factory and published by SNK, had the ‘gap timed’ style, and was released precisely one month prior to Devil May Cry. Ofcourse, I don’t know when they both began development, or who implenented it first during development – I wonder who was peeking into who’s office?