Games be hard to make, yo

August 14, 2019 12:37 PM

Why does it take so long to make videogames? Polygon explores what goes right and wrong when games like Diablo 3 take a decade and how you can spend 18 months on getting a character to hop a fence.
posted by adrianhon (17 comments total)

The 18 months article is a pretty great writeup on the insane complexity of the modern art pipeline and the tradeoffs involved. Ten years ago when it was a little bit easier that was my daily as a lowly art AP, and currently it’s my wife’s daily as a full producer overseeing a major studio’s art team. As complicated as all this is the programming side is even worse, to the point where I don’t think it can be effectively communicated to perfectly intelligent non-specialists without literal, actual months of priming in advance.

Here is an example of Epic breaking down a single major new renderer feature to better support open world games, for those specialists. I encourage you to just spend a minute scrubbing through for a sense of scale, because this isn’t even the nitty-gritty details of the changes (which are now my daily), just enough of a 10,000-foot view that the coders coming back from GDC have a rough sense of the schedule impact they’ll have to contend with if they want to benefit from the potential massive reduction in draw calls seen at the end (and given those numbers pretty much the entire team will insist on this, for reasons you can appreciate just having read the 18 months article).

Here’s an excellent writeup on a renderer change of similar scope for lighting on Bioshock Infinite, written by one of the best programmers I’ve ever worked with.

So what does all of this actually mean for game developers and game players?

While this growth in complexity continues, AAA team sizes have expanded to compensate for the increased workload to somewhere between 200~300 plus outsourcing, and yet the price of games hasn’t changed since the 90s. If you’re wondering why the modern DLC situation exists, it’s partly to offset this price stagnation, but mostly so the team isn’t immediately axed by shareholder demand the second the game’s out the door: it makes them something other than a source of ongoing costs on some accountant’s balance sheet. Enabling the studio heads to justify not flushing their talent pool down the drain every 3-5 years when the accounting inquisition shows up on their doorstep is a basic survival trait of all major studios.

While a lot of devs may privately agree with their audience that it’s not an ideal state of affairs, salaries are already half of what you’d get for the same work in other industries, and eating is nice.

This increasing complexity is not, counterintuitively, the primary driver of crunch (though it certainly doesn’t help). Crunch generally comes from two sources: 1) poor project planning because the industry is rife with terrible producers, many of them the cream of the crop from last project’s QA team now freshly graduated and utterly unprepared for even the lowest tier of project management (self included, circa 2007), or 2) an arms race between major studios who (correctly) see squeezing more hours out of an otherwise-equivalent team as a competitive advantage. Because the team size vs communication efficacy graph does not change a certain amount of crunch will always be inevitable, but crunch as a persistent cultural phenomenon is mostly down to the usual bullshit of capitalism trying to win on the margins.

What about paid lootboxes? Where do they fit into all this, you might ask? They don’t. That’s just deliberate exploitation of addictive behavior patterns that have been a mainstay of gambling since the dawn of civilization, worming their way into modern PC/console gaming via cross-pollination with the mobile side of the industry.

Also accountants realizing just how much money they can milk from the whales and breathing so heavily the shareholders can hear it.

Anyways, back to crunch.
posted by Ryvar at 3:23 PM on August 14, 2019 [29 favorites]

When it's a little indie team crunching for years, it's probably just the scale of the job. When it's an AAA studio, it seems like it's hubris and churn. The Diablo 3 story basically confirms this--

senior staff kept second-guessing themselves, which meant that the rest of the team had to spin its wheels by making several different versions of the game, all of which "dead-ended."

—but it seems that Blizzard, unlike most game companies, can deal with the design churn. (Overwatch was rescued from another failed game.) The Destiny 2 story is a lot sadder.

(As a player, I wish the big games would dial it down. I have a bunch of games that I haven't finished because they're just too damn big. The open world where every square inch has a quest marker is attractive for a month or so, and then becomes a chore I don't want to face.)
posted by zompist at 3:38 PM on August 14, 2019 [3 favorites]

senior staff kept second-guessing themselves, which meant that the rest of the team had to spin its wheels by making several different versions of the game, all of which "dead-ended."

The same thing came up in the Anthem post-mortem. They spent four years on various random concepts that they ended up tossing out, then the actual thing they shipped took about 18 months.
posted by tobascodagama at 3:44 PM on August 14, 2019 [1 favorite]

It seems apparent that AAA games should cost a lot more at retail than they do now. Like $200, if not more.
posted by rhizome at 8:56 PM on August 14, 2019

I guess I'm an outlier, because all the focus on highly detailed rendering (and other purely cosmetic things, such as deformable chairs and destructible trash cans) doesn't do much for me. I mean, sure – Skyrim is very pretty. But I'm ultimately way more interested in the game systems and mechanics.

Given the option, I'd prefer for developers to spend their finite time and energy working on those. Imagine how immersive a game you could create if those CPU cycles were spent modeling in-game social relationships, economic systems, etc. (Substitute whatever non-cosmetic element is appropriate for the genre of game that you're making.)

This is basically why I lost three years of my life to Minecraft (I'm exaggerating, but only slightly). (Well, that and the great community over at MeFightClub.) The graphics engine is laughably primitive by AAA standards – but it's full of engaging mechanics that interact in complex ways. That's what makes a game feel like a medium for creative play, and not just a series of quest-marker hoops to jump through.

Games aren't movies. They're games. Let them shine as games. Chess didn't need next-gen shader algorithms, or a corny savior narrative grafted clumsily on top, to become one of the world's most popular games.
posted by escape from the potato planet at 4:41 AM on August 15, 2019 [7 favorites]

The thing is, this relentless chase after bigger and bigger gameworlds with higher and higher fidelity is primarily driven by the publishers themselves wanting something to brag about on the back of the box. The biggest game in the world right now is fucking Fornite BR, which has a cartoony art style, runs on any smartphone from the past five years, and focuses on a single map. Before Fortnite, it was a tie between Minecraft and League of Legends.

The whole "MORE RAYTRACING, BIGGER OPEN WORLDS" thing can lead to some success, but building a small game that does one thing well leads to runaway blockbuster success. (And here's where I further note that Minecraft, at least, was a single, budget-tier up-front purchase with no microtransactions until fairly recently.)
posted by tobascodagama at 5:30 AM on August 15, 2019 [3 favorites]

If you didn't experience it at release, you probably don't know that Diablo 3 was an abysmal game when it came out. It was grindy, it was not fun, it had a real money auction house that would instantly boost your power astronomically, the highest difficulty was unplayable and stupid. It just wasn't fun. At all. Not a bit.

Then they went back to the drawing board and it's quite good now. Their last two "seasons" which are around 3 months of start over and level up and race up leaderboards have had as much population as the season where they introduced a new character. Next season they're dropping new armor sets for each class (basically like getting a new class for each!) and that's neat.

So yeah, that was even a decade ago from now when it came out. Sheesh. But yeah, Diablo 3 was clearly just not a coherent vision when it was first released. It was bad.
posted by OnTheLastCastle at 7:34 AM on August 15, 2019 [1 favorite]

As a programmer who has tinkered with making games, and who has a vast catalog of game concepts I wish I had time to explore, I recently enjoyed this short video about the ways the indie game Dead Cells interprets user input to provide a satisfying experience. The takeaway is that one mustn't take input too literally; whenever possible it is better to detect the user's intent and then give them what they want.
posted by Hot Pastrami! at 8:19 AM on August 15, 2019 [2 favorites]

But yeah, Diablo 3 was clearly just not a coherent vision when it was first released. It was bad.

I love these retrospectives because it shows diverse "players" are as a community, and how subjective the experience of a game is. I absolutely loved Diablo 3 on release and have many fond memories of it. What you saw as "grindy" and being too difficult to the point of being "unplayable and stupid" I saw as "awesome, finally a meaningful challenge".

I mean compare the two experiences:

Original Diablo 3: Ok I just beat Hell difficulty, time to try Inferno. Omg even a wasp one shots me lol how am I suppose to beat this, everything does 10x more damage??? Then you actually got good at avoiding damage, farmed some decent gear either by grinding or sniping good offers from the AH, then finally beat it, wow what an achievement, now I'm done with the game. Remember this was before the RMAH.

Today's Diablo 3: Ah my best G.Rift is level 89. I want to go for a personal best of level 90. The mobs on that difficulty have about 15% more hp and damage. But the main determinant to "beating" it is really mob density, mob type and favorable rift boss. So if you wanted to beat 90, the main way you'd do it is just "fish" for a good rift by repeatedly opening and closing them until you found an easy rift then run through it and pray for a favourable boss type that you'd geared yourself for. Even if if you beat 90... well there's 91. And 92. And it doesn't end, there's always a higher difficulty with 15% more hp and damage. You're never finished. It's all meaningless numbers in the end. Every season they just buff your damage and hp numbers so you get to a higher number than the last season.

I mean don't get me wrong, I love current Diablo 3 as well and it's probably a much better "game". But I'll always remember being one shotted by an Inferno wasp and feeling how impossible it was and then finally getting good enough to beat Inferno... I won't really remember fishing around for +1 G.Rift level...
posted by xdvesper at 4:35 PM on August 15, 2019 [1 favorite]

Isn't it more satisfying to play chess with a weighted set on a smooth, level surface, with pieces that have felt on the bottom, and designs that clearly communicate the different pieces in an aesthetically pleasing way, than on an unfolded piece of cardboard with cheaply-made plastic pieces?

Video games are not movies it's true, but they are not simply games (i.e. a collection of rules, strategies, outcomes, etc.) either. It's fine to ignore aesthetics and focus on mechanics, and there are video games that cater to that, but it's perfectly reasonable to want to play a game for the story, or simply to see what kind of images can be achieved with the most modern hardware and rendering techniques. Happily there is a tremendous variety of them available.
posted by subocoyne at 4:54 PM on August 15, 2019 [1 favorite]

when i’m really deep in a game of chess the pieces stop seeming like pieces and start seeming like abstract icons representing the origin point of the imaginary lines across the board that indicate the squares that that piece-icon might occupy in subsequent moves. when i’m really on and the player i’m playing against is really on, it feels like we’re knotting those intersecting lines together in increasingly elaborate patterns.

i haven’t played chess seriously in years, and i miss being able to get into that headspace.
posted by Reclusive Novelist Thomas Pynchon at 7:53 PM on August 15, 2019 [1 favorite]

i guess this is all to say that graphics are often orthogonal to what i’m looking for out of games.

i’ve often low-key fantasized about throwing together a quick 2d fighting game where the characters are represented by nothing but abstract hitboxes and hurtboxes of varying sizes, shapes, and velocities.
posted by Reclusive Novelist Thomas Pynchon at 7:56 PM on August 15, 2019 [2 favorites]

Remember this was before the RMAH.

What ? Diablo 3 had the RMAH from release. Removal of the auction house in 2014 is the reason why today's D3 is a much better game than on release.
posted by Pendragon at 11:09 PM on August 15, 2019

Yeah, Diablo 3 had the RMAH from the start with a slight delay from release May 15, 2012 to June 12. I will choose to believe you because hey, miracles are fun and maybe you did run around for 3 hours trying to chip away at a single elite mob. Enrage timers on bosses + broken math from game makers + no intelligent drop system so the odds of you having even one ok item just dropping... it's like they intentionally made a game to piss people off and THEN asked you to pay money to buy good items.
posted by OnTheLastCastle at 7:28 AM on August 16, 2019 [1 favorite]

i guess this is all to say that graphics are often orthogonal to what i’m looking for out of games.

And this is totally valid. It is a sign of a well-designed game that players can enter a flow state and perceive the underlying relationships between game elements and the rules that govern their interactions apart from their representations. Sometimes the graphical representation can interfere with that if it's not implemented well.

And in some video games, it is necessary to render a 3 dimensional world in high detail to allow that. Chess works on an 8x8 grid which just needs two colors. Hide and seek wouldn't work in that space and benefits from detailed, complex physical spaces.

Just as some players are strictly concerned with gameplay, there are many people who engage with video games primarily through their narrative and aesthetic elements. When players put the game down, they may think about gameplay strategies, or they may draw fanart of the characters, discuss plot points on a forum, or try playing the pause menu theme song on piano.

Someone may want to put the computer they built through its paces and play the most graphically demanding game available to see what the latest rendering techniques can produce.

It's wonderful that the medium of video games (which yes, is a terrible name for it) can support all of these different ways of engagement, and I'm much more interested in how the elements inform and relate to one another rather than trying to determine which element is the fundamental one.
posted by subocoyne at 10:12 AM on August 16, 2019

Chess didn't need next-gen shader algorithms, or a corny savior narrative grafted clumsily on top, to become one of the world's most popular games.

No, clearly it needed Generals.
posted by sneebler at 9:28 AM on August 17, 2019 [1 favorite]

What ? Diablo 3 had the RMAH from release. Removal of the auction house in 2014 is the reason why today's D3 is a much better game than on release. posted by Pendragon

Not true, Diablo 3 launched on May 15 and I had more or less finished Act 4 Inferno by the time the RMAH came out on June 12.

When the RMAH came live I didn't see any point in playing further - I had already defeated the game, and now it turned into a money grind? I immediately sold about 2/3 of my gear on the RMAH when it launched and got paid about $200, and I gave the rest away to my friends who hadn't beaten the game yet. I still have the payment receipts from Blizzard. So I guess I benefited from the system - I never bought anything from the RMAH, only sold items.

Enrage timers on bosses + broken math from game makers + no intelligent drop system so the odds of you having even one ok item just dropping... it's like they intentionally made a game to piss people off and THEN asked you to pay money to buy good items.

Haha, not just that, there was also the extreme repair costs, and the fact the bosses would heal to full when you died or leashed them. Some people enjoy difficult stuff, and there's clearly a market for games like that (Dark Souls, etc).
posted by xdvesper at 11:22 PM on August 17, 2019