The first mistake you'll make when playing as the monster is to assume you should behave like one.
He is, after all, a towering tank of rock and muscle that spews jets of fire and can throw boulders the length of a football pitch.
So the first time I trampled through a densely vegetated forest, leaving wildlife pulverised in my wake, perhaps it was understandable that confidence was high as I searched for my adversaries; a motley crew of four puny, pitiful humans.
Suddenly, a harpoon darted into the beast's shoulder. It was about the size of a small lamppost. Next, from another direction, a tranquilizer dart lodged into his left calf. Clever girls.
The quartet of hunters emerged from the forestry and encircled their foe, now semi-paralysed. Streams of bullets poured in, and as the monster attempted to flee the trap, another was sprung. A huge hexagon-tiled dome materialised, encasing all five combatants inside. With the monster's health already halved so expertly, and nowhere to run, the only choice was to fight back.
New: Gameplay footage
Smothered in red lashes and open cuts, the monster roared out a jet of fire, almost in outrage at being seized upon. He crushed one hunter into the mud by lunging at him, which worked for a few seconds, until the fallen comrade was revived by a teammate. Then the nukes, tremendous energy-sapping bursts of light, fell from the heavens. These are usually hard to land, but not so much when your target is tethered down by a whale-grade harpoon. That did it. Beast slain. A mercilessly embarrassing schooling, but also a fine demonstration of how to coordinate an attack.
Evolve isn't trying to recreate Godzilla or King Kong. It doesn't want the star of the show to be a near-invincible ogre who can wipe out entire platoons with one swing of his arm. The focus of this next-generation shooter is a masterfully balanced four-versus-one deathmatch where the advantage swings back and forth like a pendulum.
Californian studio Turtle Rock, which co-developed Left 4 Dead with Valve, has made levelling up an essential strategic element. The Goliath begins each match on a baseline level of ferocity, underpowered to the extent that the scales are tipped against him. So while it seems counter-intuitive to run away at the outset of each match, especially considering he's the intimidating sort, only highly skilled players can hold their own at level one.
The immediate goal for the Goliath is to find surrounding wildlife (elk, if I remember correctly, or some fictional herbivores) and devour enough of them to evolve to level two. At this stage, the scales are perfectly evened. The Goliath's skin hardens, his health bar fills up, the muscles surrounding his back and shoulders balloon.
When a level-up takes place, an alert message is sent to all four hunters. Anxiety sweeps through the team; not only have they lost their early advantage, but they must also expedite their search before the Goliath reaches his third and final evolution.
At level three, the Goliath is pure danger. His arms grow to a critical mass and a Stegosaurus-like crown of bone pierces through his back. His melee attacks are devastating, the boulders he throws are now one-hit-kills, and the player controlling him can activate an entire list of skill perks.
"Since the very beginning of development, that core idea, four versus one, has always been there" - Matt O'Driscoll, producer
Victory for the Goliath is achieved by killing all four hunters, yet the window of opportunity is small, as his victims will respawn after just two minutes. But Evolve is so deftly calibrated that, while four-versus-one is a balanced fight, the team is at a major disadvantage when reduced to a trio.
The difference between four and three hunters is obvious (yes, fewer) but it's also nuanced. Active communication and masterful teamwork is essential for the unit to survive, and each hunter carries unique weapons that the whole team will rely on.
Slideshow: The Goliath up close
Griffin, a 'trapper', was the one who fired the harpoon gun. He is also responsible for activating a giant mobile dome that entraps the monster. Val, meanwhile, is the medic with the tranquiliser gun. While she also carries the typical team-healing tools, her party trick is a sniper rifle that carves a hole in the Goliath, opening a custom weak spot for others to pour their bullets into.
Hank was responsible for launching the nukes. He also comes equipped with a cloaking device that is particularly useful for reviving teammates. Markov, meanwhile, spearheads the attack with heavy-duty weapons. The combination of a personal shield, along with an energy-sapping lighting gun, makes him the only hunter who can get up close with the Goliath.
All four, bar the medic, carry standard weapons too. However, it is their unique abilities, especially when combined, that creates the most effective offense.
So when just one hunter is taken out of the equation, the remaining three must make allowances and readjust their strategy. They are now also preoccupied with reviving their fallen comrade, meaning the monster is often at a strategic benefit if he guards his corpses.
Two-versus-one is desperation time. The group's palette of offensive manoeuvres is virtually halved, as is the damage output. Then, when one hunter becomes lone survivor, their only chance is to flee and hope to survive until respawned reinforcements arrive. This is a particularly satisfying moment for those who play the monster and have a nasty streak in them - you know your opponent is terrified and on the back foot, so why not toy with them first?
"Since the very beginning of development, that core idea, four versus one, has always been there," says Matt O'Driscoll, Evolve's executive producer.
"Things have changed, and we've tried different ideas, but that core gameplay element has survived everything."
Production on Evolve began three years ago, when the revived company's co-founder Chris Ashton began to prototype a blend of cooperative and competitive multiplayer.
"Ashton has had this idea in his mind for a long time. Even before Left 4 Dead he's wanted to make this game," says O'Driscoll.
The first build of the prototype was playable in April 2011, just four months into the project.
"We were using assets from all over the place to get the idea up and running," O'Driscoll adds. "At the time, the monster was kind of like a giant crab, really early temp stuff, just to prove the concept."
The development process from this point onwards was akin to a beta test, albeit one kept confidential within the studio. Numerous ideas were quickly integrated and tested, while other teams put together the assets. Evolve's beautifully delicate balance and complexity, only achievable through thousands of hours of early testing, is perhaps the greatest justification of this development philosophy.
"Once it was up and running, we were playing it every day. That's the whole studio. We need that feedback," O'Driscoll says.
"We have a fairly flat hierarchy at Turtle Rock, and we're always eager to try different ideas. But it was also fairly easy to test new ideas and see how they played out.
"The monster's rock throw, for example, came later in development. As we were playing we started realising that hunters would begin to turtle together, and set up mines around them so they were protected, and it forced a stalemate. The rock throw solved that problem."
What's so encouraging is Turtle Rock's unshakable focus on its principal idea. Key decisions like the above are not made to enhance the monster's lethality, or to add some cosmetic depth to attacks, but to ensure that all matches are as tightly contested as possible.
"We were using assets from all over the place to get the idea up and running. At one point, the monster was kind of like a giant crab, really early temp stuff, just to prove the concept"
Hunting down the beast, for example, is aided by icons on the map that approximate his location. Startled birds, some one hundred meters away, give you a general idea of where to look, but the game design doesn't revert to pin-point radars and giant flashing arrows. It's far more organic than that.
Monsters in expert hands can also, should they decide, enter stealth mode. Footprint trails go cold, the wildlife offers no clues, and naturally hunters tend to split up and expand their search though the dark forest. It's a classic horror movie trope, and an indefensibly stupid strategy that always seems like a good idea at the time.
Even this has a natural internal balance to it. The general trepidation is somewhat eased by knowing that a motionless Goliath isn't one hunting for elk. And while being pounced on by a Goliath ranks as the game's most terrifying scene, spotting his motionless silhouette before he notices you is a uniquely proud moment.
Evolve's two-hour demo build is a testament to Turtle Rock's multiplayer pedigree, and its faith in a core gameplay philosophy. Publisher 2K promises dedicated servers across platforms (PS4, Xbox One and PC), and O'Driscoll adds that if one player drops out, AI takes over.
The big question is whether the team can bring its expertise to a single player campaign. The fact that O'Driscoll nor 2K are ready to discuss it yet, with release set for autumn, doesn't exactly inspire confidence.
"We are going to have single-player, there are a lot more modes, a lot more monsters, a lot more hunters we want to talk about," O'Driscoll says.
The suspicion, perhaps unfairly, is that Evolve's obvious focus means that the single player mode will inevitably feel derivative. This is a retail release, after all, and such padding is apparently necessary to sell a triple-A game. The boxes need to be ticked.
But regardless of concerns surrounding the anonymous single-player, Turtle Rock's first major project since Left 4 Dead is already showing tremendous promise. Just out of interest, I ask O'Driscoll whether he'd be proud of the game if the multiplayer element was released on its own: "Absolutely," he says.