Over the past week, I've laid my soul bare and opened myself to blunt public criticism by documenting my first halting attempts at creating level designs (rather than simply reviewing them) via Super Mario Maker. As you've all seen, the ability to analyze games doesn't immediately translate into being able to create them.
It's been a bumpy road full of learning, experimentation, failure, and refinement. Today, however, I put my work to the true test. Rather than simply doing trial runs and playing through my stages as I went along, I handed the Wii U Game Pad to my wife Cat to see how a normal human being approaches my work. Cat is what you might call an occasional game enthusiast; she enjoys word and puzzle games, but when it comes to action games she only dabbles on occasion. The only action games she's ever loved enough to play to completion, fascinatingly enough, are the DS Castlevania games, which she likes because (1) they have a moderate pace; (2) they let her kill stuff; and (3) the more stuff she kills the stronger she gets. While Cat's played around with some of the New Super Mario Bros. games, she's never made it very far in them; their pure action approach just doesn't click for her. So while she's no stranger to Mario, she's also not what you'd call a seasoned player. In other words, a perfect test subject for my purposes. Yesterday I talked about wanting to make truly accessible World 1-1 stages that anyone could master — now here's where that ambition gets put to the test.
Probably the most important thing about watching someone else play through your work is the way it helps you see things you've made in a different way. I wouldn't be a very good QA tester; I tend to fall into grooves with gaming, and don't easily switch tracks to take different approaches. So, with these levels I've made, I have a hard time doing anything but playing optimally. Someone who doesn't know the ins and outs of the stages, though? Who plays Mario casually rather than with casual ease? Totally different.
What I noticed during our play test session is that Cat doesn't approach levels by tackling the most efficient route to the end; instead, she explores, tests, moves tentatively. Rather than dash to the end, she makes use of the full clock ("What's that music? Why did it get faster? I should be worried, shouldn't I?") and really interacts with the various stage elements that, for me, are simply background details. It's interesting to see.
"Gambolin' Goombas" works really nicely for this approach thanks to the inclusion of the SMB3-style "plate" platforms, which don't create an obstruction for efficient play but do create the presence of interesting structures to wander through more methodically.
I also confirmed what a few people have mentioned in comments: Springy note blocks have no business existing in a World 1-1. Cat recognized their purpose and importance, but she couldn't master the timing of her spring jumps. Much too fussy for a starting stage.
The Fire Flower also proved unreasonably difficult to collect — you need to either make a very precise running leap at it, or stand directly beneath it and bend your jump in mid-air... both of which require far too much finesse for a beginning stage.
I also realized that the interaction between this gap and the automatically placed background element creates too much visual clutter, and the pit doesn't "read" well. Also, the hidden Starman past the small gap actually caused her to bump her head on the invisible block and fall into a pit — that's some Lost Levels cruelty right there, which has no place here.
Based on these observations, I filled in the pit, moved the invisible block to the left, and widened the platform beneath the Fire Flower to make for an easier jump.
The verdict? Much more approachable.
In fact, after considerable struggles before the edits, Cat managed to clear the stage on her second attempt once the bumps had been smoothed over. Success! If not 100% novice-friendly.
My other 1-1 attempt, on the other hand, proved far more manageable from the get-go. The biggest hang-up actually appeared right at the beginning of the level, with the first gap to be cleared. It's just a little too far to clear without running, and the small divot at the lip of the left side of the pit creates a sort of psychological hazard that caused her to leap too early every single time.
I tweaked the ground — raising the divot and lowering the landing area — and it worked perfectly without changing the structure of the level.
There were a few bumps that followed, but she bested each of them on her next go-around — which is kind of the point. Live and learn.
Some near-misses, too.
I loved her reaction to the Muncher bridge, though. While they're simply there as scenery, Cat didn't realize the bridge was solid ground and made this sort of flinching jump, expecting to land on the Munchers. She responded with happy disbelief when Mario survived after all — but it made her much more conscious of the Munchers to appear throughout the rest of the stage.
Also, surprisingly, while I had determined the lower path here was the more appealing one, she took the straight upper stretch because it involves less jumping over open pits. This gave her the opportunity to indulge in wanton turtle abuse.
And an easy win.
My older level designs, unsurprisingly, proved much less accessible. The underwater level (no, not the Ninja Turtle level — I wouldn't make her play that! I'm not a monster!) proved manageable, but the rest were well beyond her abilities. That's no judgment on her skills; those other levels simply need some reconsideration. Especially the second level I built last week, which needs reconsideration from top to bottom, even after I gave it a massive overhaul over the weekend. There's nothing like watching another person struggle with your work to help you realize how terrible a designer you are!
Anyway, this was a valuable exercise. Hopefully it means fewer totally awful levels going forward...