Erik Wolpaw to Valve on Portal 3 — 'we should just do it'

Valve should just do some jams. You don't need 75 people to make a game, get five, make some gameplay. It's the artsy stuff that takes time - graphics, sound, that stuff. And then all the programmer time spent messing with the engine to support the artsy things.

Gameplay though? Gameplay goes crazy fast when you cut the turnaround times of asking for animations and stuff to go with your experiments out of the picture. Mess with some gameplay, design some levels, go all untextured and programmer art and recycled assets from previous games and you're golden, five guys who know what they're doing can make a whole huge game like that in a couple months.

Get your five guys to make the whole game start to finish like that. It'll look like crap but that's not important, it just has to play good. Then, when it's already fun, and the levels are basically designed, and the story is written, then you bring in the big team for the final stretch on the expensive bits. Replace your boxy CSG test levels with real art. Animate new models. Bring in the voice actors and get the sound guys doing their thing. Screw with the engine to push boundaries, like you do. But the key here is that the game part of the game is already made, there's a very clear picture of what work needs to be done.



Heck, you're making Portal 3, right? Portal is 90% puzzles. And Valve made that whole slick fancy level editor for fans to use, right? The one that's so ridiculously easy to use that a braindead monkey can make Portal test chambers in a few minutes a pop?

Build the first version of Portal 3 in the Portal 2 test chamber editor. Do you even need five guys? You could have a comprehensive prototype in a snap and bring in the big guns to go from there. Bro if you really wanted to make Portal 3 you absolutely *could* solo it, forget all the polish, make the scaffolding. Save the set dressing for later, that's where the time is spent.

Or say you don't solo it. Say that you get all 300 people at the company... for one, single day. Everyone, business people included. Sit everyone in front of the portal 2 test chamber editor, even the accountants, remember, brain dead monkeys can do it. Have everyone shoot for 5 levels, tell them to go wild, make anything they want. See what they come up with. Make it a party, get everyone in a room together bouncing ideas off each other with the level editor right there to realize them as fast as you can think it up. Then have everyone play each other's levels, try to sort out the best stuff. Boom. Just like that, you've got a huge amount of content to jump off with. 300 x 5, we're talking 1,500 levels here, no way you can't find two dozen or so ideas worth exploring further.

And if it's a party? A party with 300 of the most creative people on the planet hanging out together to just make something? No risk, no strings attached? You're gonna get a lot more than just levels. You're going to get jokes out the wazoo. Story ideas. New mechanics, people going "wouldn't it be cool if...?"

You don't have to commit to a 3 year development cycle with all hands on deck, you can start something good just having a little bit of fun.

Quoting: SkyGuyWhyValve should just do some jams. You don't need 75 people to make a game, get five, make some gameplay. It's the artsy stuff that takes time - graphics, sound, that stuff. And then all the programmer time spent messing with the engine to support the artsy things.

Gameplay though? Gameplay goes crazy fast when you cut the turnaround times of asking for animations and stuff to go with your experiments out of the picture. Mess with some gameplay, design some levels, go all untextured and programmer art and recycled assets from previous games and you're golden, five guys who know what they're doing can make a whole huge game like that in a couple months.

Get your five guys to make the whole game start to finish like that. It'll look like crap but that's not important, it just has to play good. Then, when it's already fun, and the levels are basically designed, and the story is written, then you bring in the big team for the final stretch on the expensive bits. Replace your boxy CSG test levels with real art. Animate new models. Bring in the voice actors and get the sound guys doing their thing. Screw with the engine to push boundaries, like you do. But the key here is that the game part of the game is already made, there's a very clear picture of what work needs to be done.



Heck, you're making Portal 3, right? Portal is 90% puzzles. And Valve made that whole slick fancy level editor for fans to use, right? The one that's so ridiculously easy to use that a braindead monkey can make Portal test chambers in a few minutes a pop?

Build the first version of Portal 3 in the Portal 2 test chamber editor. Do you even need five guys? You could have a comprehensive prototype in a snap and bring in the big guns to go from there. Bro if you really wanted to make Portal 3 you absolutely *could* solo it, forget all the polish, make the scaffolding. Save the set dressing for later, that's where the time is spent.

Or say you don't solo it. Say that you get all 300 people at the company... for one, single day. Everyone, business people included. Sit everyone in front of the portal 2 test chamber editor, even the accountants, remember, brain dead monkeys can do it. Have everyone shoot for 5 levels, tell them to go wild, make anything they want. See what they come up with. Make it a party, get everyone in a room together bouncing ideas off each other with the level editor right there to realize them as fast as you can think it up. Then have everyone play each other's levels, try to sort out the best stuff. Boom. Just like that, you've got a huge amount of content to jump off with. 300 x 5, we're talking 1,500 levels here, no way you can't find two dozen or so ideas worth exploring further.

And if it's a party? A party with 300 of the most creative people on the planet hanging out together to just make something? No risk, no strings attached? You're gonna get a lot more than just levels. You're going to get jokes out the wazoo. Story ideas. New mechanics, people going "wouldn't it be cool if...?"

You don't have to commit to a 3 year development cycle with all hands on deck, you can start something good just having a little bit of fun.

Are you working in project management?

(I hope not )

Quoting: SkyGuyWhyValve should just do some jams. You don't need 75 people to make a game, get five, make some gameplay. It's the artsy stuff that takes time - graphics, sound, that stuff. And then all the programmer time spent messing with the engine to support the artsy things.

Gameplay though? Gameplay goes crazy fast when you cut the turnaround times of asking for animations and stuff to go with your experiments out of the picture. Mess with some gameplay, design some levels, go all untextured and programmer art and recycled assets from previous games and you're golden, five guys who know what they're doing can make a whole huge game like that in a couple months.

Get your five guys to make the whole game start to finish like that. It'll look like crap but that's not important, it just has to play good. Then, when it's already fun, and the levels are basically designed, and the story is written, then you bring in the big team for the final stretch on the expensive bits. Replace your boxy CSG test levels with real art. Animate new models. Bring in the voice actors and get the sound guys doing their thing. Screw with the engine to push boundaries, like you do. But the key here is that the game part of the game is already made, there's a very clear picture of what work needs to be done.



Heck, you're making Portal 3, right? Portal is 90% puzzles. And Valve made that whole slick fancy level editor for fans to use, right? The one that's so ridiculously easy to use that a braindead monkey can make Portal test chambers in a few minutes a pop?

Build the first version of Portal 3 in the Portal 2 test chamber editor. Do you even need five guys? You could have a comprehensive prototype in a snap and bring in the big guns to go from there. Bro if you really wanted to make Portal 3 you absolutely *could* solo it, forget all the polish, make the scaffolding. Save the set dressing for later, that's where the time is spent.

Or say you don't solo it. Say that you get all 300 people at the company... for one, single day. Everyone, business people included. Sit everyone in front of the portal 2 test chamber editor, even the accountants, remember, brain dead monkeys can do it. Have everyone shoot for 5 levels, tell them to go wild, make anything they want. See what they come up with. Make it a party, get everyone in a room together bouncing ideas off each other with the level editor right there to realize them as fast as you can think it up. Then have everyone play each other's levels, try to sort out the best stuff. Boom. Just like that, you've got a huge amount of content to jump off with. 300 x 5, we're talking 1,500 levels here, no way you can't find two dozen or so ideas worth exploring further.

And if it's a party? A party with 300 of the most creative people on the planet hanging out together to just make something? No risk, no strings attached? You're gonna get a lot more than just levels. You're going to get jokes out the wazoo. Story ideas. New mechanics, people going "wouldn't it be cool if...?"

You don't have to commit to a 3 year development cycle with all hands on deck, you can start something good just having a little bit of fun.

I will simply say that if it were really that easy, then that's how everybody would do it.

Quoting: sub

Quoting: SkyGuyWhyValve should just do some jams. You don't need 75 people to make a game, get five, make some gameplay. It's the artsy stuff that takes time - graphics, sound, that stuff. And then all the programmer time spent messing with the engine to support the artsy things.

Gameplay though? Gameplay goes crazy fast when you cut the turnaround times of asking for animations and stuff to go with your experiments out of the picture. Mess with some gameplay, design some levels, go all untextured and programmer art and recycled assets from previous games and you're golden, five guys who know what they're doing can make a whole huge game like that in a couple months.

Get your five guys to make the whole game start to finish like that. It'll look like crap but that's not important, it just has to play good. Then, when it's already fun, and the levels are basically designed, and the story is written, then you bring in the big team for the final stretch on the expensive bits. Replace your boxy CSG test levels with real art. Animate new models. Bring in the voice actors and get the sound guys doing their thing. Screw with the engine to push boundaries, like you do. But the key here is that the game part of the game is already made, there's a very clear picture of what work needs to be done.



Heck, you're making Portal 3, right? Portal is 90% puzzles. And Valve made that whole slick fancy level editor for fans to use, right? The one that's so ridiculously easy to use that a braindead monkey can make Portal test chambers in a few minutes a pop?

Build the first version of Portal 3 in the Portal 2 test chamber editor. Do you even need five guys? You could have a comprehensive prototype in a snap and bring in the big guns to go from there. Bro if you really wanted to make Portal 3 you absolutely *could* solo it, forget all the polish, make the scaffolding. Save the set dressing for later, that's where the time is spent.

Or say you don't solo it. Say that you get all 300 people at the company... for one, single day. Everyone, business people included. Sit everyone in front of the portal 2 test chamber editor, even the accountants, remember, brain dead monkeys can do it. Have everyone shoot for 5 levels, tell them to go wild, make anything they want. See what they come up with. Make it a party, get everyone in a room together bouncing ideas off each other with the level editor right there to realize them as fast as you can think it up. Then have everyone play each other's levels, try to sort out the best stuff. Boom. Just like that, you've got a huge amount of content to jump off with. 300 x 5, we're talking 1,500 levels here, no way you can't find two dozen or so ideas worth exploring further.

And if it's a party? A party with 300 of the most creative people on the planet hanging out together to just make something? No risk, no strings attached? You're gonna get a lot more than just levels. You're going to get jokes out the wazoo. Story ideas. New mechanics, people going "wouldn't it be cool if...?"

You don't have to commit to a 3 year development cycle with all hands on deck, you can start something good just having a little bit of fun.

Are you working in project management?

(I hope not )

Gonna have to quash you're hopes! Yep. Works damn well actually, way better than you'd think. I don't do AAA though, it's a money pit that, pushing the boundaries. I aim for Double A. Fast, cheap, get a whole lotta content done real quick. Nothing new of course, the idea is do the established things really good, things someone else has already figured out and written an instruction manual for. Preferably multiple instruction manuals, you want lotsa perspectives to point out the speedbumps you might run into. GDC vault is my favorite thing. XD

Maybe I wasn't very clear with how I wrote that - that's definitely not going to get you something you can ship, especially not at Valve quality. But it's a start. A strong start. The idea is that they're having trouble starting something because they're not willing to commit the resources to knuckle down and bring in the team for a significant amount of time. A few really solid prototypes can help alleviate that and get the ball rolling.

Valve knows how to develop a game. They don't need me telling them how to do that. Valve's problems are all politics, the place is run like a democracy and you have to campaign to get anything done. A game jam? Well that's a sick campaign stunt.

Quoting: Mountain Man

Quoting: kuhpunkt

Quoting: sub

Quoting: BeamboomImagine sitting on a *guaranteed* multi million seller game franchise and choose NOT to release a sequel.
And they even have several of those. They're livin' the good life over there at Valve.

Well, I guess we're easy to overlook the numbers.
While it seems to be big money, it's actually significantly less compared to Steam sales
and putting all your resources in there.

The estimates for Valve's annual revenue is like 7-10 billion USD.
Compare that to the total of sold Half Life copies (~ 9 million) over many, many years -
even if you assume they were all sold at max price (they weren't).
In particular, if you consider how much pressure it must be for Valve always trying to keep
or surpass their standards for their games (I can only imagine). So there is a risk for reputation loss.

Don't get me wrong. I'd love to see Valve getting back to their franchises.
And there IS room for hope. And I'd also claim there is way more value for Valve having those franchises alive
with recent high-class releases than "just" what they directly earn from them.
I just hope they see this.

Yes - but if they weren't developing games anymore, why employ them? They earn good money at Valve and they aren't doing nothing.

Positions are not as clearly defined at Valve as you might expect, and employees can freely move from one project and position to another depending on their interests. So just because someone might have the job title "level designer" doesn't necessarily mean he spends 40 hours a week designing levels for games.

Point is, you can't look at job titles at Valve and get any idea of what is happening inside the company.

Yes, but a level designer will not have the qualifications to work on say Steam.

Quoting: Mountain Man

Quoting: SkyGuyWhyValve should just do some jams. You don't need 75 people to make a game, get five, make some gameplay. It's the artsy stuff that takes time - graphics, sound, that stuff. And then all the programmer time spent messing with the engine to support the artsy things.

Gameplay though? Gameplay goes crazy fast when you cut the turnaround times of asking for animations and stuff to go with your experiments out of the picture. Mess with some gameplay, design some levels, go all untextured and programmer art and recycled assets from previous games and you're golden, five guys who know what they're doing can make a whole huge game like that in a couple months.

Get your five guys to make the whole game start to finish like that. It'll look like crap but that's not important, it just has to play good. Then, when it's already fun, and the levels are basically designed, and the story is written, then you bring in the big team for the final stretch on the expensive bits. Replace your boxy CSG test levels with real art. Animate new models. Bring in the voice actors and get the sound guys doing their thing. Screw with the engine to push boundaries, like you do. But the key here is that the game part of the game is already made, there's a very clear picture of what work needs to be done.



Heck, you're making Portal 3, right? Portal is 90% puzzles. And Valve made that whole slick fancy level editor for fans to use, right? The one that's so ridiculously easy to use that a braindead monkey can make Portal test chambers in a few minutes a pop?

Build the first version of Portal 3 in the Portal 2 test chamber editor. Do you even need five guys? You could have a comprehensive prototype in a snap and bring in the big guns to go from there. Bro if you really wanted to make Portal 3 you absolutely *could* solo it, forget all the polish, make the scaffolding. Save the set dressing for later, that's where the time is spent.

Or say you don't solo it. Say that you get all 300 people at the company... for one, single day. Everyone, business people included. Sit everyone in front of the portal 2 test chamber editor, even the accountants, remember, brain dead monkeys can do it. Have everyone shoot for 5 levels, tell them to go wild, make anything they want. See what they come up with. Make it a party, get everyone in a room together bouncing ideas off each other with the level editor right there to realize them as fast as you can think it up. Then have everyone play each other's levels, try to sort out the best stuff. Boom. Just like that, you've got a huge amount of content to jump off with. 300 x 5, we're talking 1,500 levels here, no way you can't find two dozen or so ideas worth exploring further.

And if it's a party? A party with 300 of the most creative people on the planet hanging out together to just make something? No risk, no strings attached? You're gonna get a lot more than just levels. You're going to get jokes out the wazoo. Story ideas. New mechanics, people going "wouldn't it be cool if...?"

You don't have to commit to a 3 year development cycle with all hands on deck, you can start something good just having a little bit of fun.

I will simply say that if it were really that easy, then that's how everybody would do it.

You would be amazed to realize how many parts of this world are run the very hard way simply because people don't know anything different. The human spirit is incredible, many people are massively successful doing things the worst possible way just because they grit their teeth and force through anyway, coming up with a wonderful product through sheer stubbornness. The game industry at large, and in fact, all of software really, are definitely done the hard way. It's a new industry, very much in its infancy, naturally nobody's really any good at it. What's 50 years in the grand scheme of things? Barely a single generation, two if you push it! Do you know how many generations it took us to figure out "put seed in ground, plant grow"? We're the forefathers those whippersnappers in 2200 will laugh at for doing things so stupid while they know better, because we passed on our knowledge learned through painful trial and error and they're standing on our shoulders.

We're like the Egyptians. Building the pyramids through sheer brute force, dragging massive blocks of stone across the ground 10,000 slaves at a time. Meanwhile across the pond, the Romans have figured out this amazing thing called *wheels*. Don't be so quick to assume that if there's a better way, you'd be using it already!

Quoting: Eike

Quoting: KristianI never understood the people that are under the impression that Valve has stopped developing games. Valve employees artists, level designers, game designers and the like. What do they imagine these people do if not work on games?

Dunno.

But if you look at the list of their games, there's large gaps.

Dota 2 in 2013, fine. Never heard of Counter-Strike Online 2, Counter-Strike Nexon: Zombies, Left 4 Dead: Survivors (an expansion, I guess?) or The Lab (water testing for VR it seems). Than there's infamous Artifact in 2018. Of course, Dota e.g. needs work during its life. But I'm not overwhelmed, to be honest. What do you imagine these people were working on?

Look at that list of cancelled project..if accurate it would explain alot of things.

Quoting: Mountain Man

Quoting: TrainDoc

Quoting: Kristian

Quoting: Mountain Man

Quoting: EikeI can understand frustration there. Imagine you were a software developer working for a cool company with some of the coolest franchises out there - and you're not allowed to work on them, for years, for decades maybe...

As I understand Valve's corporate structure, it's the employees who more or less decide the direction of the company. So it's not like there are a bunch of people champing at the bit to make new games but the bosses won't let them. If enough people got together and decided they wanted to make Half-Life 3, or Portal 3, or whatever, Gabe wouldn't stop them. The real question is why there's no momentum internally to get those projects rolling.

There is the infamous Valve employee handbook.

Total PR stunt btw. No such document existed before or after and is just a recruiting tool. Valve's problems is it's rockstar developers lead the charge on everything and juniors who don't fall in line risk termination.

At least one former Valve employee has gone on record to say that the handbook is real. However, she does agree with you that there are people within the company who have managed to acquire quite a lot of internal influence which rather defeats the purpose of a flat management structure.

https://www.wired.com/2013/07/wireduk-valve-jeri-ellsworth/

The handbook is real in that exists but not much more. Appreciate you citing the sources that I was too lazy to find myself 😅.

Quoting: TrainDoc

Quoting: Mountain Man

Quoting: TrainDoc

Quoting: Kristian

Quoting: Mountain Man

Quoting: EikeI can understand frustration there. Imagine you were a software developer working for a cool company with some of the coolest franchises out there - and you're not allowed to work on them, for years, for decades maybe...

As I understand Valve's corporate structure, it's the employees who more or less decide the direction of the company. So it's not like there are a bunch of people champing at the bit to make new games but the bosses won't let them. If enough people got together and decided they wanted to make Half-Life 3, or Portal 3, or whatever, Gabe wouldn't stop them. The real question is why there's no momentum internally to get those projects rolling.

There is the infamous Valve employee handbook.

Total PR stunt btw. No such document existed before or after and is just a recruiting tool. Valve's problems is it's rockstar developers lead the charge on everything and juniors who don't fall in line risk termination.

At least one former Valve employee has gone on record to say that the handbook is real. However, she does agree with you that there are people within the company who have managed to acquire quite a lot of internal influence which rather defeats the purpose of a flat management structure.

https://www.wired.com/2013/07/wireduk-valve-jeri-ellsworth/

The handbook is real in that exists but not much more. Appreciate you citing the sources that I was too lazy to find myself 😅.

It not only exists (or at least used to exist), but according to that former employee, is actually given to new hires.

Quoting: Eike

Quoting: KristianI never understood the people that are under the impression that Valve has stopped developing games. Valve employees artists, level designers, game designers and the like. What do they imagine these people do if not work on games?

Dunno.

But if you look at the list of their games, there's large gaps.

Dota 2 in 2013, fine. Never heard of Counter-Strike Online 2, Counter-Strike Nexon: Zombies, Left 4 Dead: Survivors (an expansion, I guess?) or The Lab (water testing for VR it seems). Than there's infamous Artifact in 2018. Of course, Dota e.g. needs work during its life. But I'm not overwhelmed, to be honest. What do you imagine these people were working on?

Game stuff that never gets released, maybe? Games tend to have a lot of prototyping (and Valve is notorious for going above and beyond in this) and experimental content that gets cut and not released.

Quoting: SkyGuyWhyWe're like the Egyptians. Building the pyramids through sheer brute force, dragging massive blocks of stone across the ground 10,000 slaves at a time. Meanwhile across the pond, the Romans have figured out this amazing thing called *wheels*. Don't be so quick to assume that if there's a better way, you'd be using it already!

Ok, I'm not a historian, but I have to point out some historical errors with this analogy: the pyramids in Egypt could not have been constructed* without some sophisticated engineering know-how**, because no number of humans (as in, without tools of some kind) would be able to move stone blocks of that size through sheer brute force. (There's a limit to how many people can fit around a rock to pick it up, after all [and friction precludes dragging/pushing]…) Also it's a common misconception that the pyramids were constructed with slave labor; more modern scholarship points out that the pyramids were constructed by well-paid specialist labor, with reports of workers even striking and getting better compensation. And yet when Saladin's son decided to destroy the pyramids of Giza in AD 1196, the attempt was given up after 8 months due to the extreme difficulty involved in even deconstructing them, after accomplishing little more than a gash in the smallest of the three. Sometimes the earliest attempts to do something, though not as good as that which will come later, can already be pretty good.

*Over a thousand years before Rome was even a collection of huts on the banks of the Tiber, for reference; the great pyramids of Giza are as far removed in time from Cleopatra, as she was from us.

**Including the wheel; I can't imagine where the idea the Egyptians didn't have those is coming from—Egyptian chariots were the tanks of the Bronze Age, cutting-edge military hardware that was (rightly) feared throughout the surrounding world for centuries before Rome existed as a political entity. They even had an approximation of pi to better than 1% hundreds of years before the Greeks began theorizing about it.


Last edited by Philadelphus on 19 April 2022 at 7:36 pm UTC