A Crazy Open Letter to CD Projekt Red Re: Witcher 3. Also, a parable about Graphics Cards.
This is my point-in-time summary as to what I perceive about Witcher 3. I'm normally hesitant to spam letters over forums, but alas, this is neither a 1) Tech Support inquiry 2) A media/business/career inquiry, therefore this is my contact method. So I'm going to put my long rambling thoughts into a text wall that no one will read, because I have free time whilst waiting for a patch. I'll begin by saying that I am an immense fan of previous titles in the series. I'm quite pleased with your ability to bring this interesting work of fiction to a world-wide audience. Your company has always produced great and solid work in my experience.
Good things: As to The Witcher 3 - Wonderful game, excellent immersive story telling. Graphics are good, not exactly up to what I was expecting, but a respectable improvement over Witcher 2. World seems to be very deep and realized, relatively authentic capturing of period-inspired culture. Voice acting is well done. Combat and game mechanics resemble previous two games, but are well-realized enough that further innovation may not be required or productive. Game experience is cohesive and well thought out. Soundtrack is context appropriate and well-fitted to the game.
Problems: Game Shipped as playable for the first 8 or 9 hours. Became unplayable beyond 5 minute increments afterwards (Combat mode CTD, specifically ATI285x Sword Hit Crash). I understand a patch has been promised. I'm not going to ride you with"OMG WHERES MY PATCH FOR MAH CARD ITS BEEN FIVE DAYS OMG!!?!".
This is more of a commentary that I engage in from time to time, to convey, in Quixotic futility, certain feed backs and unsolicited opinions to developers that I see value in.
So, checking online, there are the standard range of bug reports regarding people's unwashed configurations, such as "WHY WONT THIS RUN ON MY UN-UPDATED WINDOWS 7 THAT HAS AMD DRIVERS FROM FOUR YEARS AGO OMG", or "MY GEFORCE IS DOWNCLOCKING WHEN I GO TO INVENTORY" but large cross-sections of bugs appear to be platform specific, most commonly for AMD cards, although some hiccups are reported for consoles and Nvidia cards. I have an ATI 285x, which I upgraded, comically, the week before the TW3 released (Dungeons 2 bluescreened and somehow magically shorted my power supply, which nuked my components. Object lesson: Don't build your internal test rigs with anything made by OCZ, as the overvoltage protection is worthless).
But I'm a root causation guy. So, checking development history, these GPU issues can sourced from two issues: the IHVs and CD Red.
In this case, AMD appears to have screwed up their development support. AMD's line on this is the correct assertion that Nvidia applied back-handed strategies to obtain developer integration of an Nvidia specific architecture with no qualitative value (Specifically, Nvidia Hairworks) that was dependent on architecture considerations that Nvidia knows AMD has issues handling. That said, Hairworks is pointless at this point in time. It will be immensely valuable in four years when GPU's have progressed, but currently is in the same vein as rendering realistic water flows - its a like wrapping an ice cream sandwich in a pancake wrapped in bacon. TW3 is an adrenaline fueled story game with moments of intense combat. Hairworks a "maybe nice-to-have possibly I don't know?" feature. It is not critical to the game experience that the werewolf I just frantically dodged and chopped to pieces has realistic fur. The game being stable across multiple end-user platforms outranks prettiness. I acknowledge this is as an unrealistic minority philosophy.
Nvidia correctly counters asserts that AMD are being silly tiny babymen who are making excuses for their 1) Failure to properly pipeline their software and 2) Not paying enough attention. Although their cards are of a good quality, and they have strong software suite support, I don't buy Nvidia anymore as I don't approve of their role in the industry. As stated, tis an unrealistic minority philosophy I hold.
I would personally add that AMD has once again failed to anticipate Nvidia's business model. For the fifth or sixth year in a row. As a long-time AMD customer, I have the authority to assert that AMD needs to stop complaining about the fact that Nvidia's business model is based upon ruthless dickery and either beat them at their own game or use their market share to alter the industry in order to make that model uncompetitive, similar to what Intel is trying to do.
The other half of this is a developer failure. I get it. Graphics is HARD. The architecture and mastery of API calls alone is something I roughly understood 10 years ago. I look at even a cursory summary now, and it reads like some kind of bizarro magic. Some of these things are millions of lines of code long, written by PhD wizards over a decade. Its not easily intelligible. So naturally, you need someone who is an expert on the constantly shifting and updating netherworld of graphics. Hence, IHV reps.
They're like the Sorceress lodge in the Witcher series, actually. They're necessary, but what they do is complicated enough to seem like some kind of creepy magic, yet is usually just disguised science, but is also vital enough that you have to use their services. But once you see how they do it, or more importantly, why they do it sometimes, you walk away feeling like your objectives are subordinate to some convoluted, yet ultimately childishly pointless scheme. This forces developers to make hard design choices that pits their desire to make a great game with a desire to actually release that game, which results in them using IHV resources, which sometimes compromise their final product. I've seen in interviews that this developer has given, that they care what effects those choices end up putting on gamers sometimes
The fact that you care about poor choices you might've made, however, is not a substitution for the fact that you chose poorly. That said, developers do not always realize that if you own the weight of whatcha done did, and whatcha learned from it, gamers will respect you, so long as you tell them. Loudly. They'll give you crap regardless of what you do, because they are ultimately customers, but you can earn their respect. Acknowledging the problem is good. You appear to have done that. Its a bit early for you to do a "Lessons Learned" podcast, but I'd encourage you to put that on your calendar for the next two weeks. Once you solve the major problems.
But I get it. You're struggling to live up to the promises you made. You showed people that it would looked absolutely wicked sick in 2013, and now you don't want to let them down. You had to dumb it down for the consoles. So an IHV rep emails you and says "Hey, we're really interested in making sure this works on our platform, why don't we show you how to integrate hairworks and some of our other interesting APIs?" basically offering to make your game look as epic as you told people it would be. They offered to fix your bike for free. No Strings attached.
You're probably aware already that it ain't that simple. But in case you aren't, or if you just want to hear that someone understands what you've gone through, we will now have "Mr Insignus's Colorful Storytime", in which you, like so many AAA developers before you, enter what appears to be a lost episode of the Twilight Zone.
You find yourself driving down a busy highway at night. There is a trio of ten year olds in the back seat, and a passenger next to you. Two of the children hate each other, and the other doesn't understand why it is in the car in the first place. The fighting pair won't stop poking each other, but you're deathly afraid that if you pull over, you won't get there on time.
One of the pair insists that you should stop for Pizza, but refuses to tell you what toppings are acceptable. When you buy the pizza, he comes back and says "Oh, I told you those toppings are fine, but I'm not going to eat them, because I'm allergic. This doesn't have peanuts or tessellation in it, does it?."
The other insists that you stop for Ice Cream, but only if its chocolate, with mint, and a scoop of peanut butter swirl, and only if neither of the other two gets any ice cream. IceCreamLad distinctly told you 2 hours ago that he hates peanut butter, but he now wants it anyway.
You know that if you buy Pizza, PizzaKid will shut up, but IceCreamLad will keep hitting him. You know that if you buy Ice Cream, Ice CreamLad will shut up completely, and Pizzakid will only whine for the first ten minutes.
In all of these situations, the third boy is still clueless as to the fact he is in a car. He's texting his friend about how to inflate price points on mini-cogitators. You ask him if he wants anything, and he says "Oh nah its fine. Just as long as its on an integrated platform and is open source! It needs to work on my phone and my laptop and my tablet! Call my friend Jeff, he'll read back postings from the support database." Touched lad, that one.
Through all of this, your passenger, who has lent you his car because his license is suspended, is constantly badgering you with commentary about why people can't raise their kids properly, changes his mind as to which child is right or wrong every five minutes, constantly demands that you just pull the damn car over, and then complains about why you're going to be late when you do. He also has an array of nonsensical comments about why you are doing full stops at stop signs, while at the same time yelling that the police car you just passed is totally coming after you now.
As the supposed driver, you solve for Z to get through Atlanta and past Long's Peak in under 15 days, which means buying Ice Cream. In your heart of hearts, you really, really just wish they'd all grow the hell up so you can at least make it past Denver.
But you can't make them.... or can you?
This isn't a developer problem, as it is secretly a thinly veiled parable framing what is really an ethical problem, and not a business problem. There is a question you should be asking:
Who do you really want to be driving? The passengers, despite how annoying they are sometimes in their insistence that you drive the way they want you to? The children, even though they are flippantly, innocently unaware of their status as patently unreasonable ****s? Yourself, perhaps, because you want to get there on time without breaking your wallet on gas money? The ethical solution to this parable was to not to pull the car over, but to not buy them treats, either. Keep on driving. Talk to the passenger, sometimes you'll agree, sometimes you'll need to tell him to stow it. You might've gotten there fifteen minutes later than intended, but you'll have gotten there in one piece.
What you choose determines your reputation in the industry. Someone will hate you regardless of what you do. The question is whether or not you'll hate yourselves because they don't like you. Also who ultimately matters, and how you balance weights to service the customer. Parable doesn't cover that one, but its truefax.
Summary: There are multiple lines of criticisms regarding why this much anticipated AAA title from a respected dev house has had a moderately embarassing launch, at a level of roughly 3.65/10 (Nowhere near AC Unity embarassing, which will retain the 10/10 for a while yet). While my opinion is not that of an expert, I conclude that most lines of criticism have some kernel of truth of them. Industry enthusiasts might deflect that on the IHVs. Gamer enthusiast will form their camps. Most gamers however, will blame you. Some developers might view this as aggravating and unfair.
I would advise you strongly to view it as an opportunity to engender respect by owning up to it, letting your customers peak behind the curtain into what you'll do differently going forward, and work really hard on fixing it.
To be fair, most users are having smooth, fun experiences with the product. However, the depth of failures for some users does produce a negative perception in the community. Gamer enthusiasts, who monitor and track publisher information, will read up on this and be concerned. They then blog and comment through the forums about it. This spreads ill-will about your company, some of which I think may be undeserved.
Unfortunately, these tend to be the same gamers who will tweak their configs and produce GPU Edge cases, which makes them prone to experience game ending bugs. Explaining this fact to them, however, is unwise. Cross-platform baseline stability should be a goal instead.
You can't fix your customers. You probably can't fix the state of the industry vis-a-vis graphics. But you can be more attentive to cross-platform stability. You're a company known for its compelling RPGs, rich storylines, and detailed content rendering.
Lesson: The hair is unimportant. You may be able to have the game render the blood spatter of five different guys that the player just cut in half with a single sword strike, but if it drops to 2 FPS or CTD's, no one really cares outside of the blogosphere. Be mindful of people pushing you to use the latest sickest graphics that only 5,000 people can utilize just because its available and it makes them money when it advances an upgrade cycle.
Yes, you do have to progress and be competitive, graphically, but you can make strategic decision about not utilizing features without alienating vendors or diminishing your sales. We will respect you for it if you let us play.
So tell me a story, challenge me as a gamer, and have your game reward me in measure to my willingness to play by the rules.
But let me play, please.
This is my point-in-time summary as to what I perceive about Witcher 3. I'm normally hesitant to spam letters over forums, but alas, this is neither a 1) Tech Support inquiry 2) A media/business/career inquiry, therefore this is my contact method. So I'm going to put my long rambling thoughts into a text wall that no one will read, because I have free time whilst waiting for a patch. I'll begin by saying that I am an immense fan of previous titles in the series. I'm quite pleased with your ability to bring this interesting work of fiction to a world-wide audience. Your company has always produced great and solid work in my experience.
Good things: As to The Witcher 3 - Wonderful game, excellent immersive story telling. Graphics are good, not exactly up to what I was expecting, but a respectable improvement over Witcher 2. World seems to be very deep and realized, relatively authentic capturing of period-inspired culture. Voice acting is well done. Combat and game mechanics resemble previous two games, but are well-realized enough that further innovation may not be required or productive. Game experience is cohesive and well thought out. Soundtrack is context appropriate and well-fitted to the game.
Problems: Game Shipped as playable for the first 8 or 9 hours. Became unplayable beyond 5 minute increments afterwards (Combat mode CTD, specifically ATI285x Sword Hit Crash). I understand a patch has been promised. I'm not going to ride you with"OMG WHERES MY PATCH FOR MAH CARD ITS BEEN FIVE DAYS OMG!!?!".
This is more of a commentary that I engage in from time to time, to convey, in Quixotic futility, certain feed backs and unsolicited opinions to developers that I see value in.
So, checking online, there are the standard range of bug reports regarding people's unwashed configurations, such as "WHY WONT THIS RUN ON MY UN-UPDATED WINDOWS 7 THAT HAS AMD DRIVERS FROM FOUR YEARS AGO OMG", or "MY GEFORCE IS DOWNCLOCKING WHEN I GO TO INVENTORY" but large cross-sections of bugs appear to be platform specific, most commonly for AMD cards, although some hiccups are reported for consoles and Nvidia cards. I have an ATI 285x, which I upgraded, comically, the week before the TW3 released (Dungeons 2 bluescreened and somehow magically shorted my power supply, which nuked my components. Object lesson: Don't build your internal test rigs with anything made by OCZ, as the overvoltage protection is worthless).
But I'm a root causation guy. So, checking development history, these GPU issues can sourced from two issues: the IHVs and CD Red.
In this case, AMD appears to have screwed up their development support. AMD's line on this is the correct assertion that Nvidia applied back-handed strategies to obtain developer integration of an Nvidia specific architecture with no qualitative value (Specifically, Nvidia Hairworks) that was dependent on architecture considerations that Nvidia knows AMD has issues handling. That said, Hairworks is pointless at this point in time. It will be immensely valuable in four years when GPU's have progressed, but currently is in the same vein as rendering realistic water flows - its a like wrapping an ice cream sandwich in a pancake wrapped in bacon. TW3 is an adrenaline fueled story game with moments of intense combat. Hairworks a "maybe nice-to-have possibly I don't know?" feature. It is not critical to the game experience that the werewolf I just frantically dodged and chopped to pieces has realistic fur. The game being stable across multiple end-user platforms outranks prettiness. I acknowledge this is as an unrealistic minority philosophy.
Nvidia correctly counters asserts that AMD are being silly tiny babymen who are making excuses for their 1) Failure to properly pipeline their software and 2) Not paying enough attention. Although their cards are of a good quality, and they have strong software suite support, I don't buy Nvidia anymore as I don't approve of their role in the industry. As stated, tis an unrealistic minority philosophy I hold.
I would personally add that AMD has once again failed to anticipate Nvidia's business model. For the fifth or sixth year in a row. As a long-time AMD customer, I have the authority to assert that AMD needs to stop complaining about the fact that Nvidia's business model is based upon ruthless dickery and either beat them at their own game or use their market share to alter the industry in order to make that model uncompetitive, similar to what Intel is trying to do.
The other half of this is a developer failure. I get it. Graphics is HARD. The architecture and mastery of API calls alone is something I roughly understood 10 years ago. I look at even a cursory summary now, and it reads like some kind of bizarro magic. Some of these things are millions of lines of code long, written by PhD wizards over a decade. Its not easily intelligible. So naturally, you need someone who is an expert on the constantly shifting and updating netherworld of graphics. Hence, IHV reps.
They're like the Sorceress lodge in the Witcher series, actually. They're necessary, but what they do is complicated enough to seem like some kind of creepy magic, yet is usually just disguised science, but is also vital enough that you have to use their services. But once you see how they do it, or more importantly, why they do it sometimes, you walk away feeling like your objectives are subordinate to some convoluted, yet ultimately childishly pointless scheme. This forces developers to make hard design choices that pits their desire to make a great game with a desire to actually release that game, which results in them using IHV resources, which sometimes compromise their final product. I've seen in interviews that this developer has given, that they care what effects those choices end up putting on gamers sometimes
The fact that you care about poor choices you might've made, however, is not a substitution for the fact that you chose poorly. That said, developers do not always realize that if you own the weight of whatcha done did, and whatcha learned from it, gamers will respect you, so long as you tell them. Loudly. They'll give you crap regardless of what you do, because they are ultimately customers, but you can earn their respect. Acknowledging the problem is good. You appear to have done that. Its a bit early for you to do a "Lessons Learned" podcast, but I'd encourage you to put that on your calendar for the next two weeks. Once you solve the major problems.
But I get it. You're struggling to live up to the promises you made. You showed people that it would looked absolutely wicked sick in 2013, and now you don't want to let them down. You had to dumb it down for the consoles. So an IHV rep emails you and says "Hey, we're really interested in making sure this works on our platform, why don't we show you how to integrate hairworks and some of our other interesting APIs?" basically offering to make your game look as epic as you told people it would be. They offered to fix your bike for free. No Strings attached.
Yes, that was a Full Throttle reference. One forum post a month, must, per rules, reference a classic pre-2000's game
You're probably aware already that it ain't that simple. But in case you aren't, or if you just want to hear that someone understands what you've gone through, we will now have "Mr Insignus's Colorful Storytime", in which you, like so many AAA developers before you, enter what appears to be a lost episode of the Twilight Zone.
You find yourself driving down a busy highway at night. There is a trio of ten year olds in the back seat, and a passenger next to you. Two of the children hate each other, and the other doesn't understand why it is in the car in the first place. The fighting pair won't stop poking each other, but you're deathly afraid that if you pull over, you won't get there on time.
One of the pair insists that you should stop for Pizza, but refuses to tell you what toppings are acceptable. When you buy the pizza, he comes back and says "Oh, I told you those toppings are fine, but I'm not going to eat them, because I'm allergic. This doesn't have peanuts or tessellation in it, does it?."
The other insists that you stop for Ice Cream, but only if its chocolate, with mint, and a scoop of peanut butter swirl, and only if neither of the other two gets any ice cream. IceCreamLad distinctly told you 2 hours ago that he hates peanut butter, but he now wants it anyway.
You know that if you buy Pizza, PizzaKid will shut up, but IceCreamLad will keep hitting him. You know that if you buy Ice Cream, Ice CreamLad will shut up completely, and Pizzakid will only whine for the first ten minutes.
In all of these situations, the third boy is still clueless as to the fact he is in a car. He's texting his friend about how to inflate price points on mini-cogitators. You ask him if he wants anything, and he says "Oh nah its fine. Just as long as its on an integrated platform and is open source! It needs to work on my phone and my laptop and my tablet! Call my friend Jeff, he'll read back postings from the support database." Touched lad, that one.
Through all of this, your passenger, who has lent you his car because his license is suspended, is constantly badgering you with commentary about why people can't raise their kids properly, changes his mind as to which child is right or wrong every five minutes, constantly demands that you just pull the damn car over, and then complains about why you're going to be late when you do. He also has an array of nonsensical comments about why you are doing full stops at stop signs, while at the same time yelling that the police car you just passed is totally coming after you now.
As the supposed driver, you solve for Z to get through Atlanta and past Long's Peak in under 15 days, which means buying Ice Cream. In your heart of hearts, you really, really just wish they'd all grow the hell up so you can at least make it past Denver.
But you can't make them.... or can you?
This isn't a developer problem, as it is secretly a thinly veiled parable framing what is really an ethical problem, and not a business problem. There is a question you should be asking:
Who do you really want to be driving? The passengers, despite how annoying they are sometimes in their insistence that you drive the way they want you to? The children, even though they are flippantly, innocently unaware of their status as patently unreasonable ****s? Yourself, perhaps, because you want to get there on time without breaking your wallet on gas money? The ethical solution to this parable was to not to pull the car over, but to not buy them treats, either. Keep on driving. Talk to the passenger, sometimes you'll agree, sometimes you'll need to tell him to stow it. You might've gotten there fifteen minutes later than intended, but you'll have gotten there in one piece.
What you choose determines your reputation in the industry. Someone will hate you regardless of what you do. The question is whether or not you'll hate yourselves because they don't like you. Also who ultimately matters, and how you balance weights to service the customer. Parable doesn't cover that one, but its truefax.
Summary: There are multiple lines of criticisms regarding why this much anticipated AAA title from a respected dev house has had a moderately embarassing launch, at a level of roughly 3.65/10 (Nowhere near AC Unity embarassing, which will retain the 10/10 for a while yet). While my opinion is not that of an expert, I conclude that most lines of criticism have some kernel of truth of them. Industry enthusiasts might deflect that on the IHVs. Gamer enthusiast will form their camps. Most gamers however, will blame you. Some developers might view this as aggravating and unfair.
I would advise you strongly to view it as an opportunity to engender respect by owning up to it, letting your customers peak behind the curtain into what you'll do differently going forward, and work really hard on fixing it.
To be fair, most users are having smooth, fun experiences with the product. However, the depth of failures for some users does produce a negative perception in the community. Gamer enthusiasts, who monitor and track publisher information, will read up on this and be concerned. They then blog and comment through the forums about it. This spreads ill-will about your company, some of which I think may be undeserved.
Unfortunately, these tend to be the same gamers who will tweak their configs and produce GPU Edge cases, which makes them prone to experience game ending bugs. Explaining this fact to them, however, is unwise. Cross-platform baseline stability should be a goal instead.
You can't fix your customers. You probably can't fix the state of the industry vis-a-vis graphics. But you can be more attentive to cross-platform stability. You're a company known for its compelling RPGs, rich storylines, and detailed content rendering.
Lesson: The hair is unimportant. You may be able to have the game render the blood spatter of five different guys that the player just cut in half with a single sword strike, but if it drops to 2 FPS or CTD's, no one really cares outside of the blogosphere. Be mindful of people pushing you to use the latest sickest graphics that only 5,000 people can utilize just because its available and it makes them money when it advances an upgrade cycle.
Yes, you do have to progress and be competitive, graphically, but you can make strategic decision about not utilizing features without alienating vendors or diminishing your sales. We will respect you for it if you let us play.
So tell me a story, challenge me as a gamer, and have your game reward me in measure to my willingness to play by the rules.
But let me play, please.