"We finish the best stuff," Ashton said. "A lot didn't work. We experimented with different prototypes of monsters. Some of those ideas worked really well and some of them had a bunch of potential which made us go farther with them, but ultimately they didn't work. That’s normal game development, that’s what happens. You shoot for the sky and then once you start getting into reality you clear away the fluff."

It's a turn of phrase the two made use of often. If "the fluff" is the pie-in-the-sky, throw-everything-at-the-wall concepting phase, one can understand how those ambitious ideas might cloud your ultimate vision for the game. Hitting against functional realities will help shape what a game is, and perhaps more importantly, what it isn't.

"At the end of the day we had three monsters that we felt really, really good about, and we had just enough time to finish and polish those three," he continued. "It works well because we had three tiers of Hunters. So if you like medic, there’s three medics, and if you like being a monster, there’s three monsters. It just worked out in the end."

Monsters are the stars of Evolve, and in any iterative design process, plenty of ideas will naturally be left on the cutting room floor. Given some of the rejected concepts, though, it could have been a very different kind of game.

Parasite

This failed concept was particularly appealing, but the team struggled to strike a proper balance between automation and interactivity. It would engage in combat indirectly, using parasitic minions that rode on its body. To hear them describe it, it was almost like an injection of real-time strategy in the middle of a co-op shooter. Minions would be sent out to fight the Hunters, but players would find themselves fighting with what Ashton called "an A.I middleman." On top of that, the minions raised a technical issue. Pathfinding was difficult for the parasites, and they would interfere with the wildlife. It reached the point that the studio even had to consider reducing the level of wildlife, just for the sake of one type of monster. 

"At the end of the day it just wasn’t worth it, all the problems, all the things that weren’t working," Ashton said. "It started getting real geeky. Players wanted control over what the parasites did so they had a bunch of commands telling them to guard, or to go attack, or to move here. It sort of started becoming a different game and it wasn’t very approachable."

Grabberclaw

More a revised prototype than an entirely discarded idea, Ashton and Robb said that the Goliath monster was once a bipedal crustacean, which they described as a "big crab monster." It had a set of interchangeable claws with different abilities. One would snap so quickly it produced a concussion blast, while another would stick a poison tip in opponents to slowly kill them. The one that caused the most trouble by far was a grab claw, which could snatch and hold a Hunter.

"In playtests, he wouldn’t throw the hunter, he’d just hold onto him," Robb said. "One time this guy just grabbed a Hunter, held onto him, and walked into deep water just far enough that the Hunter was underwater and just stood there and drowned him. That was one of those abilities that people started abusing, where it was fun for about a week and then it started getting really un-fun."

Worse than dooming Hunters to a watery grave, however, was the ability to chuck them great distances. That was the intent of the grab claw mechanic, but it didn't take long for playtesters to exploit it.

"The Hunters didn’t have jet packs back then, so the idea is you could throw them," Ashton said. "If you were up on something high and threw him it might kill him when he lands. If another Hunter shot you it would force you to drop the guy you were holding, so it was all about trying to pick good spots to throw the player which was kind of a game in and of itself. Then one day--it was all based on physics--and a guy picked up a Hunter, aimed straight down, and spiked him into the ground. Killed him instantly because of the velocity. So it turned that ability from kind of a fun game of figuring out where to throw people to, oh, standard procedure is just to spike a guy into the ground. It took the fun out of it. Instead of seeing a guy fly hundreds of meters away, now you couldn’t even tell what happened. You’d just wind up dead."

Around that time they were experimenting with the jet pack. It had been a piece of optional equipment, but the team found everyone was favoring it over other the other pieces of equipment like bionic legs. That prompted them to simply give all the Hunters jet packs, which rendered the throwing ability pointless. "I think that was the last nail in the coffin for Grabberclaw," said Robb.

It was their first attempt at a monster, and many of its concepts lived on in the other creatures instead. 

Concept art for Cthulhu-like beast, Kraken

Velvet Worm

As opposed to A.I. hurdles and exploits, some monsters simply didn't fit into the game as they envisioned it. The Velvet Worm, a giant creature seemingly straight out of Dune, would be able to skitter over the ground. To pull off the effect convincingly, though, the team would have had to compromise its vision for the environments. They wanted the world to feel lush and organic, with overgrown trees and irregular rock formations. That presented a heavy hurdle for animation.

"We had it concepted and designed and we had its abilities mapped out," Ashton said, "and then when we started thinking: okay, what do we need to do to make a worm move through the environment convincingly? It would be one thing if we were playing in a city environment where the terrain was flat. But our environments are very organic and just getting a worm’s body to flow over the rocks--that’s where reality started clearing away the fluff."