How does one create combos?


#1

Picking up Margaret (P4U2) the day she was released really made me realize how little knowledge I possessed in general. Without a bunch of threads detailing the most optimal combos or her or help with her general gameplan or in learning her, I was lost, completely unsure of what to do with a character I intended to have as my sub. Basically, I want to know how you attain the knowledge, or what sort of knowledge is required, to be one of those people who come up with these “Day one” combos, optimal combos, and so on. And beyond that, even things as simple as coming up with frame traps, or counter-hit set ups, pressurestrings, etc.

I know it takes a lot of experimentation, but I feel I lack the know-how to understand how to properly do so and it gives me a feeling of inadequacy and I feel it’s a hurdle I need to overcome to become a better player. I appreciate any feedback. Thanks.


#2

A lot of it is just understanding how combos work. Here’s a couple things I’ve written once upon a time:


In anime games normals combo into each other pretty much exclusively via chain cancels (called gatlings in GG at least), so you have to explore what chains into what to build combo strings. With specials, likewise. Once you understand how combos work mechanically start exploring the properties of the special moves. See what normal canceled into what special combos on the ground up close, what still works at longer distances and so on. You’ll find pieces of 1-3 moves that string together. Then put the pieces together and bam, a combo.

With juggles it’s a bit harder, but it’s pretty much just having an eye for it and testing a ton. The more you test, the better you can just feel what should work and what shouldn’t. Some people are called lab monsters for a reason.


#3

As one of these lab monsters, maybe I can break down my process a little bit when messing with a new character.

1.) Get a feel for what all of my moves do. This is basic whether you’re making combos or not. I usually start in training mode and just do stuff, keeping an eye out for any moves with large untech time/juggle potential/freeze state, and also recording the proration (damage scaling) for all of my normals so I know which ones are good starters and which ones will kill my damage if I use them in a combo.

2.) After I’m comfortable but before I’ve done any real work, I usually go into the character’s challenge mode. The combos there are usually fairly practical, but more importantly, they give you an idea of what the character can do, or what routes the developers intended. There’s a lot of weird stuff you can find in challenge mode that you might not expect to work.

3.) Training mode, set the dummy to never air tech so I can see whether something is ‘almost’ working or not even close. Start with midscreen routes always, because corner routes usually involve some of the more damaging parts of midscreen routes.

4.) This is where it really breaks down per character, but I guess I would suggest starting in the middle, looking for ‘pieces’ of a combo that deal a lot of damage and can be utilized from multiple starters. Usually this involves getting them off the ground somehow, juggling them, freezing them in place, and dealing as much damage as possible within that time. Then, I look at enders, I figure out the best oki I can get and try to link my damage route into that ender (for example, trying to end the combo in hard knockdown is usually a good idea).

5.) Finally, once I have a damaging route into a good ender, I look at starters, and just test whether I can do my new combo from 5A/5B/AA/everything. If it doesn’t work from some starters, I try to make minor adjustments until it does, or come up with an all new combo from that starter. Always do starters last. Even with weird starters like throw or random AA’s, you can find that pieces of your optimal route will often work, and will yield the best damage.

6.) This is obvious, but read and contribute to combo threads, and watch (usually japanese) videos. Most of the best ideas won’t be your own, and a lot of the optimal combos that I’ve helped with on Dustloop and TYM have been the result of multiple people bouncing optimizations back and forth.

7.) Or you could just, like, wait a week and copy other people’s stuff.