The Scientific Gamer

jiiiiim asks

Which game has made the best use of physics so far? So, not necessarily the most accurate simulation, but the one that made the best addition to gameplay.

Ah, your qualifier makes this question a bit more difficult to answer than it would be otherwise. No game ever uses a true simulation of physics; that’s something that can take the very best supercomputers weeks to model for one specific scenario and would consequently be far too complex for a mere video game to handle. Games use pseudo-physics instead – in the “best” case, massively simplified versions of real-world physics, but it often turns out that simply integrating a physics system won’t necessarily improve your game, and the best example of this is probably ragdoll physics.

In my view, ragdoll physics is good for one thing and one thing only: modelling bodies being tossed away from explosions like… well, like ragdolls. Early games featuring ragdoll physics were a little bit enamoured with the technology and simply replaced death animations with ragdolls, with the result that bodies tended to collapse to the ground in entirely unrealistic, inhuman ways. Raven Shield is probably the first game I played where unnaturally splayed-out corpses would be scattered around the levels – this is an example, although I wasn’t able to get a shot of the classic “folding in half backwards” death that so often occurred. The trouble with ragdoll physics is that human bodies aren’t ragdolls. They’re bodies, with bones and spines and musculature that is going to resist contorting in the way that a ragdoll does. And so unless you are modelling a body whose skeletal structure has been pulverised by some massive force – like an explosion – this simplified physics system was actually a step backwards from stock death animations.

Ragdoll physics was something that missed an important point: games aren’t real. They’re games; artificial environments created for entertainment, and the physics of this artificial environment can also be artificial so long as it adds some element of fun. For example, bowling people over in Half Life 2 was amazingly good fun even if the implementation wasn’t perfect, but at the same time I remember getting immensely frustrated when I was about two-thirds of the way through the game and I came across yet another seesaw puzzle. I don’t want to be piling bricks on top of a plank of wood when I can be throwing a radiator at somebody’s head instead. The first uses “real” physics, while the second uses game pseudo-physics, and it’s the second one that worked out best.

What I’m trying to say here is, if you’re asking about the best addition to gameplay then I have to tell you that’s a bit of a non-question. There’s no game that’s ever managed a truly accurate simulation of a physical system, and so most games go for the pseudo-physical solution that best aids their gameplay but which is also inherently unrealistic. How is it possible to distinguish between the physics of, say, Super Mario World, which is very, very simple and overtly unrealistic but which is also very easy for the player to understand, and the physics of Left 4 Dead, which camouflages things with some excellent animation but which, at the end of the day, is nearly as fake in the way it handles most of the object behaviours in the game?

That’s my answer as a scientist. As a gamer, there are several games which have impressed me with the way they have utilised their own private laws of physics to improve the gameplay. These tend to be titles which don’t put the physics front and centre, but which instead blend it seamlessly and naturally into the game’s structure. I found Crayon Physics Deluxe to be dreadfully boring because it was an entire game based around nothing but physics, and unrealistic physics at that. Gimmick games like Toribash or Porrasturvat are fun for five minutes but I’ve found them to have little longevity beyond the stock level of attention you’d give to an interesting curio in a museum. There’s three games in particular which I think sum up the different approaches to physics implementation.

1. I-War 2.

Noteable for being the only space sim with truly Newtonian flight physics. Every other sim models spaceships like World War 2 fighters because that’s a concept that’s fairly easy for our tiny human minds to grasp. Not I-War 2, though. In I-War 2, you will keep on going in the direction you’re travelling even if you cut power to your engines. If you’re moving sideways and shooting at an enemy your projectiles will retain their sideways velocity, making accurate targeting an absolute pain. It was an interesting attempt at realism, but one that was incredibly frustrating to actually play, and the developers had to resort to giving you a heavy level of computer assistance in the HUD to help you fly straight and shoot straight. An example of a true physics system which was actively detrimental to the way the game played.

2. Men of War

Here we have the argument for subtlety. Men of War is built around a semi-realistic ballistics and penetration system that was inserted into the game to little fanfare, but which nevertheless alters the gameplay significantly compared to a traditional RTS because it allows you to try real-world battle tactics. Using a very large gun to punch through solid cover becomes possible; similarly, the best way to take out an enemy tank is to flank it and shoot at its weaker side or rear armour instead of futilely smashing away at the thicker frontal armour. It can also lead to some comically emergent gameplay moments, like the time I drove an AA tank up to another tank to try and penetrate it at point-blank range only to have the shells bounce off at a 180 degree angle and disable the AA tank. It’s nowhere near being a perfect simulation of the way real-world ballistics would work, but it’s close enough that you’d have to look fairly hard to spot the seams.

3. Portal 2.

AKA The Best Implementation Of Physics In A Game, Ever. The entire game is physics, but because it’s simple physics like velocity and momentum and the curved trajectories of objects under the influence of gravity it’s all stuff that we intuitively understand, and so we never think about it. Moreover, the game encourages you to step outside the bounds of those “normal” physics systems by using the portal gun. It’s constantly – and very consciously – changing the rules, and it’s understanding how those changing rules have affected the physical system present in the game which leads to finding the solution to a given level. What seals it is the introduction of the gels, as they allow the player to experiment with adjusting the physical parameters of the puzzles themselves. It is, again, not even close to being real physics – the puzzle where myself and my co-op partner collided in mid-air and came to a dead stop instead of exchanging momentum and bouncing off of each other is proof of that – but Portal 2 doesn’t care. Portal 2 knows what it is doing: it’s not the fidelity of the physical simulation which matters, but rather how it best helps the game to play. And that is something that Portal 2 delivers in spades, using nothing more than a series of quasi-physical playgrounds.