Why are games with destructible environments so rare?

[Damned Registrations]

Enjoy your minecraft then.

 

[Ash]

Not so much this (unless you mean ancient hardware like XB 360 and PS3). We are currently getting diminishing returns by throwing more processing power and RAM on visuals. Spending a good chunk of the budget on destructibility and physics shouldn't be a big deal.

I revisited Vigilante 8 the other day (old '90s Playstation vehicular combat game) and was surprised to be reminded that nearly every mesh in the environment was destructible. Geometry could be reshaped dynamically (crater maker weapon) too.

And yeah, reasons above as described are the reason it is rare. It's a lot of effort, cost on performance can be large due to spawning in loads of explosions and debris, and it presents new problems with AI. Plus since levels are twice as much detailed these days, that's twice as much effort making all that detail destructible.

 

[sullynathan]

Ubisoft is doing destruction better than everyone right now.

 

[sser]

Casual late night thinking. Bold, possibly tautological thinking.

I really do not think it is that big of a technical hurdle, seeing as how some powerful environment engines have been coming out of non-AAA and AAA devs alike. I think it's more of a design decision. And when I say "technical hurdle" I mean in the macro sense -- the suggestion that we're incapable of it as a whole. We clearly are not with plenty of examples around (X-Com style destructible terrain can also be found in Laser Squad, for example). And I don't think a game has any technical hurdles in this question if it was never designed with the question in mind -- basically, putting the cart before the horse scenario.

The most basic example I can think of is that of DotA vs. LoL. In League of Legends, you have very static environments which you must go around in some way or another. In DotA, you have forests which can be cut down, used for hiding, and other gameplay elements. It's not that Riot Games doesn't have the means to make some destructible environments, it's just that it is clearly not a part of their design format.

Another example would be a game called Soldiers: Heroes of World War II. That game came out in 2004 and had real-time physics and destruction.

See here for basic destruction of buildings, including shooting guys out of specific windows, https://youtu.be/DGfwt_seUno?t=175

See here for real-time bullet physics where a tank shell nails a gunner out of an assault gun/APC without actually hitting the gun platform itself, https://youtu.be/DGfwt_seUno?t=357

Again, 2004. You might know it better by its continuation in the Men at War series. This series of games lets players take direct control of individual units and there is a modicum of actual aiming skill required.

Company of Heroes came out in 2006. It was a very sharp looking game -- yet, very little procedural destruction. Outside of airplanes slamming into ground units, I can't really think of anything substantial like we see in the above examples. Not due to technical issues, but because CoH is designed around the player building small bases, moving squads around, and not actually directly controlling any units' aiming.

Even destruction itself is tailored to design -- and the real nuts'n'bolts of these differences is found in Silent Storm/XCOM. In Silent Storm, destruction is free form because bullets and ricochets have physics of their own. In XCOM, shots are not free form -- they're pretty much glued onto targets (which is why morale breaks leading to friendly fire looked so unbelievably stupid). So we do have destruction in XCOM, but it is specifically tacked onto the battle map's grid, more or less like X-Com. However, there is also a distinct lack of good verticality in XCOM, and every time it is there you can almost feel the presence of designer ghosts still arguing not to put it in. I bring up verticality for a reason, though, which is that in Silent Storm you can actually angle your shots including through ceilings and floors. I'm not sure there is ever a scenario in XCOM where verticality allows for such a play.

But wait, there's more. XCOM utilizes a grid for its movement, too, meaning you essentially need a static form of cover to literally square up to. This keeps controls very refined and simplified -- it also allows for the 2AP system without many bumps or bruises. Silent Storm utilizes a free-form system for movement and utilizes the proper tools for it (such as strafing). If you swapped these two move styles, the destruction systems in both games would become instantly incoherent. (Want a simple example of design incoherence? Take Hatred, where you have a high-level of destructible terrain and physics except for when you bump into a moving vehicle, in which case the game retrogrades itself to Frogger, 1981.)

FPS games seem to avoid destructibility because they're often pursuing multiplayer aspects. A game like Counter-Strike would lose its razor sharp 'skill' nature if everything was destructible. The rigidity of its design focuses the players on aiming, timing, and routing skills. Counter-Strike is perhaps the most refined version of this, but it's more or less the same for any other competitive FPS. Rainbow Six Siege is designed with both destruction and multiplayer going hand-in-hand; it also seems to be the only FPS in awhile to try and go outside the box. Stranglehold is a pretty wicked third-person game with lots of destruction/terrain manipulation of a sort.

Think way back to Populous II. That game has very malleable terrain and a pretty decent modicum of destruction while utilizing a base RTS format. Its engine is more or less primitive 3D. Are RTS games of the next 20+ years just incapable of replicating this? Nah. Populous II utilized a very unique globe setting for its maps. Most other RTS's go with the flat maps and focus their design energies elsewhere. It's not that they can't do Populous's crazy world manipulations, it's just that they don't have the gameplay focused around it -- so it's not in the game.


TL;DR,

By the way, I think the Source2 engine will have some emphasis in this territory. Just a hunch.