Putting the 'role' back in role-playing games since 2002.
Donate to Codex
Good Old Games
  • Welcome to rpgcodex.net, a site dedicated to discussing computer based role-playing games in a free and open fashion. We're less strict than other forums, but please refer to the rules.

    "This message is awaiting moderator approval": All new users must pass through our moderation queue before they will be able to post normally. Until your account has "passed" your posts will only be visible to yourself (and moderators) until they are approved. Give us a week to get around to approving / deleting / ignoring your mundane opinion on crap before hassling us about it. Once you have passed the moderation period (think of it as a test), you will be able to post normally, just like all the other retards.

Obsidian under Microsoft

Monocause

Arcane
Joined
Aug 15, 2008
Messages
3,656
So I was wondering how Obsidian's doing under MS.

Years ago I've worked as QA, intermittently in an outsource outfit providing services to MS and directly at Microsoft. I've seen studios that appeared to thrive under Microsoft's umbrella (e.g. Playground Games), but also witnessed how fe. Lionhead sank during the disaster that was Fable Legends development.

F:L was an interesting case study, especially in hindsight, now that I've learned a fair bit about project management and participated in different projects from both tech and business perspectives. To give a few examples - it was IIRC 2012 or 2013, and Agile + Scrum were becoming hugely popular in software development. MS was a big ship, so adoption of Agile was usually patchwork and more often than not resulted in frankenstein monsters. F:L was no exception - the Lionhead guys technically worked in Agile, but MS management had a very waterfall approach and, AFAIK, expectations. I remember being baffled about how you can talk about being Agile, but also about milestones you committed to deliver in the next 2-3 quarters.

My impression was that the Lionhead guys kept alternating between trying to satisfy the actual users that they've engaged in the beta and satisfying contracts and commitments they had with Microsoft. End result was that the project went on in a snail's pace, and went massively over budget - don't remember the figures anymore but I think we were talking about tens of millions of pounds in the red. Microsoft, on its part, kept alternating between trying to satisfy the creative studios need for autonomy and creative independence (and they honestly seemed to try, and succeeded with playground) and MS need for control coming from the fact that they just kept losing money and also really wanted the title out of the door to boost recently released Xbox One's game portfolio which at that time was still rather underwhelming and outgunned by what PS4 offered.

The inflated budget didn't come only because of Lionhead's own failings too, mind you. A great example of how Microsoft blew its own money (and also of how the Agile vs Waterfall cultures clashed) was automated testing.

See - some MS engineers and execs were absolutely enamored with pushing the boundaries of test automation. This is a long story in QA in general, back then and even today people keep going around making powerpoint presentations claiming how you can achieve 90% automation coverage which reduces the need for manual QA, increases security and saves budget in the long run. Thing is, it's incredibly hard to pull off, and once you invest it's difficult to turn back - and overhead resulting from maintenance just keeps growing if the automation plan goes south. In a different company, I've seen 4 highly-paid automation developers working on solutions that tracked pixel positions on a webpage; it was an overengineered piece of crap and similar results would've been achieved by a single guy testing the website sporadically with his own pair of eyes. No coming back though, cause it'd make someone look bad.

With Legends there was a similar case. A bunch of highly-qualified and well-paid SDETs (Software Development Engineers in Test) laboured very hard on delivering a solution that technically could deliver 100% automation coverage. The idea was that the game would be developed with hooks in the engine in all the right places, so you could automate characters performing complex actions and spend 3 minutes (once!) writing scripts for the automation rather than 30 minutes doing the same thing manually every time code changes could've impacted the scenario.

In principle it was a good idea; 100% automation coverage was a joke that probably was supposed to brainwash some people in management, but it would indeed be a cool system. Problem was, the codebase was notoriously unstable, likely because of the clash between the automation guys expecting solid specs and Lionhead guys going agile (and also messy). Thing with test automation is that maintenance is key as you need it to be super reliable. Meanwhile what would happen was that we in manual QA would quickly start ignoring the automation reports completely as they were very rarely right about anything.

I still remember coming in to the office one morning and receiving an automation report saying that build #XX crashed on boot. I look up above my PC screen and see the guys working on build #XX which appears to run just fine. That was the last time I read an automation report.

Considering the likely wages of these SDETs involved, I'd estimate that the automation system alone consumed at least a million pounds, probably a fair bit more, while delivering no value whatsoever.

So, after this lengthy intro - would anyone know how OBS is doing under MS? Any culture clashes, technical meddling? Not asking about classic "publisher interference" but about the grittier details such as those seen in examples above.

Do we still have any OBS devs here?
 
Last edited:

Alienman

Retro-Fascist
Patron
Joined
Sep 10, 2014
Messages
17,046
Location
Mars
Codex 2016 - The Age of Grimoire Make the Codex Great Again! Grab the Codex by the pussy Codex Year of the Donut Shadorwun: Hong Kong Divinity: Original Sin 2 Steve gets a Kidney but I don't even get a tag.
Guess we will know when they are out on the street again in a couple of years. Might coincide with another Kickstarter from them ;)
 

Infinitron

I post news
Staff Member
Joined
Jan 28, 2011
Messages
97,236
Codex Year of the Donut Serpent in the Staglands Dead State Divinity: Original Sin Project: Eternity Torment: Tides of Numenera Wasteland 2 Shadorwun: Hong Kong Divinity: Original Sin 2 A Beautifully Desolate Campaign Pillars of Eternity 2: Deadfire Pathfinder: Kingmaker Pathfinder: Wrath I'm very into cock and ball torture I helped put crap in Monomyth
Re: Microsoft's past with game development, I believe the optimistic theory of the case is that they changed for the better after Phil Spencer replaced Don Mattrick as head of Xbox back in 2014.
 

Duraframe300

Arcane
Joined
Dec 21, 2010
Messages
6,395
Too early to get any relevant info. They haven't even released a game published by Microsoft yet.
 

StrongBelwas

Savant
Patron
Joined
Aug 1, 2015
Messages
482
Obsidian has released one game and announced another since the acquisition. Even Microsoft admits Outer Worlds was basically done and required nothing much from them, and Grounded started around the time Deadfire came out. The same is true of most of Microsoft's recent acquisitions, its just too early to say anything for certain. Even if they announce something this year to shill for Microsoft's new console it will probably be no more then a CGI trailer.
Of the games Microsoft has recently published, they had two games that shipped with serious issues that needed heavy patching (Sea of Thieves and State of Decay 2), one game that is just a shitshow (Crackdown 3), one game that seems mostly positively received except for a hated monetization system (Gears 5), and one real complete winner (Forza Horizion 4.)
 

RepHope

Savant
Joined
Apr 27, 2017
Messages
400
You won’t see any impacts for a while. As mentioned, TOW and Grounded both got started before the acquisition went through. There’s reasons to be optimistic and pessimistic.

Optimistically, while MS was never a home to the more hardcore RPGs the Codex likes, they were at one point a huge supporter of WRPGs. Bioware made their name with the mainstream working with MS on games like KOTOR, Jade Empire, and Mass Effect. MS helped CDPR port Witcher 2 to 360 iirc. And while in the long run they killed Lionhead, they were pretty hands off with the first three Fable entries, everything there was all Molyneaux good and bad. Since Spencer took over they’ve been pretty hands off creatively with their studios, letting them do what they want, which is why Rare isn’t stuck making Kinect games anymore. They’ve put some of their games on Steam like the Master Chief Collection and they’re overhauling their store front to be less dogshit. Their main focus is Game Pass, and they’ve been pretty open that for a service like Game Pass (basically Netflix for video games) you need a wide variety of content so I don’t expect them to force Obsidian to make looter shooters or what have you. They want to get as many genres on GP as they can because that gets more people to subscribe and stay subscribed.

Pessimistically, MS has a bad habit of committing to something and then doing a complete 180 if it doesn’t immediately work out. Say Phil Spencer leaves, then everything could completely change. The new guy might close studios, or just force the studios to chase trends like what happened under Mattrick. So everything could take a turn for the worse if upper management changes hands.

TL; DR Right now I’m not worried but that could change. Right now if Obsidian fucks up I think that will be more Feargus and Parker’s fault then MS, since those two are like incompetent Saturday morning cartoon villains.
 

S.torch

Arbiter
Joined
Jan 4, 2019
Messages
938
After Outer Worlds and Grounded the proper question to make soon would be: How is Microsoft doing it with Obsidian?
 

Monocause

Arcane
Joined
Aug 15, 2008
Messages
3,656
Pessimistically, MS has a bad habit of committing to something and then doing a complete 180 if it doesn’t immediately work out. Say Phil Spencer leaves, then everything could completely change. The new guy might close studios, or just force the studios to chase trends like what happened under Mattrick. So everything could take a turn for the worse if upper management changes hands.

TL; DR Right now I’m not worried but that could change. Right now if Obsidian fucks up I think that will be more Feargus and Parker’s fault then MS, since those two are like incompetent Saturday morning cartoon villains.

Re: Microsoft U-Turns, it vastly depends, quite possibly on the top-level manager in charge of a specific part of the business. Many people don't remember nowadays, but the reason Xbox One is called "One" was AFAIK tied to this ridiculous concept that Xbox will become your "One platform for home entertainment" The idea basically was Xbox will not only provide games, but also movie streaming, news, all kinds of home entertainment so the product will penetrate the market typically not reachable for gaming consoles.

Of course, pretty much everyone I talked to about this at the time felt the same thing - that it's stupidly unlikely that anyone buys an Xbox for any other reason than gaming. Still, for a year or two after the console got released MS supported fe. video content teams who set up themed weeks, produced lots of copy, tried to make the non-gaming bit attractive. Then the entire staff taking care of that got laid off at some point, and lots of Xbone features got axed. Probably the right call.

Conversely, Fable: Legends example above shows how MS could do the opposite. That project should've been axed probably like a year earlier, maybe more. When I touched the title for the first time my impression was that it feels competent enough, but there was not enough depth to support a playerbase - for an F2P title, for example, you need a solid progression system and IIRC Legends had none of that. Years of development and millions of pounds produced something that was probably enjoyable for an hour, and after said hour you felt you've seen everything it has to offer.

Warframe was already out at the time and I remember wondering if the Lionhead guys ever took a good look at that title to see what makes it work.
 
Last edited:

Monocause

Arcane
Joined
Aug 15, 2008
Messages
3,656
Re: Microsoft's past with game development, I believe the optimistic theory of the case is that they changed for the better after Phil Spencer replaced Don Mattrick as head of Xbox back in 2014.

Oh, I'm sure that's true. I remember Spencer being liked slightly too much and given too much praise, but the way Mattrick handled Xbone's launch made it clear he's a proper dumb-as-bricks bellend in the context of the industry he was trying to swim in - which is both surprising and not-so-surprising if you look at his resume. He grew mostly in EA and joined Zynga after leaving MS. It still baffles me that anyone in their right mind considers hiring ex-EA execs a good idea.

Since Spencer's took over, MS started making the right moves in trying to leverage their PC audience (to which they had access all the bloody time with Windows being the clear choice for most consumers), not only with stuff like the GamePass or porting Forza games to W10 but also with revitalising old Microsoft franchises such as AoE. I also remember there was a big push at some point to port lots of X360 titles to One, which also felt a correct move to boost One's catalogue and help persuade the customers to ditch their old beloved consoles.

Generally looking at Microsoft Games Studios now makes it seem like MS is considering it a proper long-time investment rather than the 3rd-class part of the business that existed to support the 2nd-class part of the business which I felt Xbox was at the time I was around. During Mattrick's time it felt like "cut the costs" and "gaming sucks, focus elsewhere" was always the name of the game around Xbox, and while I obviously don't know enough to be able to attribute it to Mattrick's decisions directly, things surely did seem to change after a few years with Spencer in charge.

It's all blurry but I think the consensus was that Spencer is no godsend, but at least he *gets* games, understands the market and his customers (or has the capacity to listen to a team that does), while Mattrick clearly didn't. Case in point: a fair bit of early Xbone titles had to waste lots of dev time and resources to support useless platform features such as voice commands or gesture controls via Kinect - even if it didn't support gameplay at all. That's exactly the kind of shtick that comes as a directive from the top management responsible for strategy, hence probably Mattrick.
 
Last edited:

Monocause

Arcane
Joined
Aug 15, 2008
Messages
3,656
In principle it was a good idea
Why? It sounds pretty bad to me and it is no surprise that it failed.

Well, the scope was overly ambitious, but a more lightweight and manageable system like this could seriously shave off hundreds, if not thousands, of hours of QA time. Think about it this way - let's say you're testing the first Fallout. Currently the dev team is heavily into level design and maps keep changing all the time. You want to ensure the Boneyard level loads properly and no major issues are present there.

1) In an absolutely black box QA setup, you pick up the build, roll a new character, head to Boneyard while trying to avoid random encounters (which can kill you and force reloads). You reach Boneyard, look at what's happening, tick the box off. Can take anywhere between 5-10 minutes not counting grabbing and setting up the build.
2) In a less black box QA setup, there's debug features. You pick up the build, roll a character, debug your way into the boneyard map, look around, tick it off. Probably like 2 minutes.
3) You press a button in your automation solution. It does everything for you, and then does 60 cases like this in 4 minutes and tells you if/when any errors occured. You run this every single time a new build comes out, and it always takes 4 minutes. You can focus your human attention elsewhere to fe. spot that a developer that left your team a week ago swapped Gizmo's portrait for his dick pic before leaving. Or that the combat shotgun makes squealing sounds when fired because someone messed up the SFX names.

If done right automation really can save lots of time, frustration and money - just like it does everywhere. If done wrong, it's an overengineered money sink. My approach usually is to advocate planning automation very carefully and on a need basis. You look at a project and identify the most repetitive tasks that eat up a lot of time, automate that. Keep going until you build a suite. This is not always possible, but it's hard to go wrong with this approach; if the Legends guys had this approach, it would deliver much more value with a fraction of the cost - and we'd at least be able to rely on it to say if the build works or not.

EDIT: Actually, if the MS automation worked as intended, considerign how long the development lasted it would shave off tens of thousands of hours of QA time. Convert that into money and efficiency boosts across the board and you can see why the prospect is so enticing. Even if you valuate an hour of QA at ten bucks for simplicity (which is a gross understatement because wages alone eat more and also often what you pay for QA is what you don't pay for extra dev time which is much more expensive), we're still talking millions of dollars in potential savings.
 
Last edited:

Turuko

Learned
Joined
Jun 12, 2019
Messages
352
Location
Verbobonc
As much as the Mattrick turnover, Nadella seems more keen on investing in the gaming division, especially with a subscription service like Gamepass
Meanwhile Ballmer bought the Clippers and he's making faces sidecourt

First tangible difference for Obsidian should be that the marketing is gonna be handled by the MS division, unified for all their studios
 

fantadomat

Arcane
Edgy Vatnik Wumao
Joined
Jun 2, 2017
Messages
37,087
Location
Bulgaria
In principle it was a good idea
Why? It sounds pretty bad to me and it is no surprise that it failed.

Well, the scope was overly ambitious, but a more lightweight and manageable system like this could seriously shave off hundreds, if not thousands, of hours of QA time. Think about it this way - let's say you're testing the first Fallout. Currently the dev team is heavily into level design and maps keep changing all the time. You want to ensure the Boneyard level loads properly and no major issues are present there.

1) In an absolutely black box QA setup, you pick up the build, roll a new character, head to Boneyard while trying to avoid random encounters (which can kill you and force reloads). You reach Boneyard, look at what's happening, tick the box off. Can take anywhere between 5-10 minutes not counting grabbing and setting up the build.
2) In a less black box QA setup, there's debug features. You pick up the build, roll a character, debug your way into the boneyard map, look around, tick it off. Probably like 2 minutes.
3) You press a button in your automation solution. It does everything for you, and then does 60 cases like this in 4 minutes and tells you if/when any errors occured. You run this every single time a new build comes out, and it always takes 4 minutes. You can focus your human attention elsewhere to fe. spot that a developer that left your team a week ago swapped Gizmo's portrait for his dick pic before leaving. Or that the combat shotgun makes squealing sounds when fired because someone messed up the SFX names.

If done right automation really can save lots of time, frustration and money - just like it does everywhere. If done wrong, it's an overengineered money sink. My approach usually is to advocate planning automation very carefully and on a need basis. You look at a project and identify the most repetitive tasks that eat up a lot of time, automate that. Keep going until you build a suite. This is not always possible, but it's hard to go wrong with this approach; if the Legends guys had this approach, it would deliver much more value with a fraction of the cost - and we'd at least be able to rely on it to say if the build works or not.

EDIT: Actually, if the MS automation worked as intended, considerign how long the development lasted it would shave off tens of thousands of hours of QA time. Convert that into money and efficiency boosts across the board and you can see why the prospect is so enticing. Even if you valuate an hour of QA at ten bucks for simplicity (which is a gross understatement because wages alone eat more and also often what you pay for QA is what you don't pay for extra dev time which is much more expensive), we're still talking millions of dollars in potential savings.
I fundamentally disagree,it is one of those things that shouldn't be automated. There is a lot more to checking the level design than seeing if there is a pixel glitch in it,many levels are scraped and more do evolve during such QA acts. Making that automated will take away big part of the game's....soul. As you said,with the debug mode it could take a few minutes,and it is always a good thing to have your devs play the game he is making. I have no idea why people would like such things,games are not conveyor based industry......at least not good games.
 

Monocause

Arcane
Joined
Aug 15, 2008
Messages
3,656
I fundamentally disagree,it is one of those things that shouldn't be automated. There is a lot more to checking the level design than seeing if there is a pixel glitch in it,many levels are scraped and more do evolve during such QA acts. Making that automated will take away big part of the game's....soul. As you said,with the debug mode it could take a few minutes,and it is always a good thing to have your devs play the game he is making. I have no idea why people would like such things,games are not conveyor based industry......at least not good games.

The kind of QA you're talking about makes sense quite late in the project, and is a different kind of testing. I totally agree that there's many elements of QA that absolutely require a keen human eye, and also a general love for games. But one of the core challenges of QA in any software project is not looking at the "feel" of the product, but verifying that this or that core functionality is intact, or that the software doesn't fall apart (e.g. crash, freeze) when you poke a stick at it here and there.

These tests are absolutely critical, yet usually are quite mundane, time consuming and boring. Prime target for automation so that you, as a human being, can focus on stuff that actually requires a human brain's attention - for a QA early on it's the development processes and practices, establishing quality standards, making sure everyone's on the same page.

Judging the 'feel', or 'playtesting' the product is a very tricky thing, by the way. When you provide feedback to developers about the feel, you're entering the fun, fun domain of opinion. As they say, opinions are like asses - everyone's got one but you don't necessarily want or should see them all. I've seen hundreds of cases where QA opinion on the games or software were a treasure, and hundreds when they were absolute garbage. More often than not, liberally sharing your opinion on stuff is sidetracking the project rather than helping it - as a former QA professional I tried to hold my opinion back unless I had data to back it up, or I was specifically asked for it, or I had a really good working relationship with the creatives or the devs. Otherwise I knew I'd end up creating noise that someone would have to spend precious time dealing with for little gain.
 
Last edited:

fantadomat

Arcane
Edgy Vatnik Wumao
Joined
Jun 2, 2017
Messages
37,087
Location
Bulgaria
I fundamentally disagree,it is one of those things that shouldn't be automated. There is a lot more to checking the level design than seeing if there is a pixel glitch in it,many levels are scraped and more do evolve during such QA acts. Making that automated will take away big part of the game's....soul. As you said,with the debug mode it could take a few minutes,and it is always a good thing to have your devs play the game he is making. I have no idea why people would like such things,games are not conveyor based industry......at least not good games.

The kind of QA you're talking about makes sense quite late in the project, and is a different kind of testing. I totally agree that there's many elements of QA that absolutely require a keen human eye, and also a general love for games. But one of the core challenges of QA in any software project is not looking at the "feel" of the product, but verifying that this or that core functionality is intact, or that the software doesn't fall apart (e.g. crash, freeze) when you poke a stick at it here and there.

These tests are absolutely critical, yet usually are quite mundane, time consuming and boring. Prime target for automation so that you, as a human being, can focus on stuff that actually requires a human brain's attention - for a QA early on it's the development processes and practices, establishing quality standards, making sure everyone's on the same page.

Judging the 'feel', or 'playtesting' the product is a very tricky thing, by the way. When you provide feedback to developers about the feel, you're entering the fun, fun domain of opinion. As they say, opinions are like asses - everyone's got one but you don't necessarily want or should see them all. I've seen hundreds of cases where QA opinion on the games or software were a treasure, and hundreds when they were absolute garbage. More often than not, liberally sharing your opinion on stuff is sidetracking the project rather than helping it - as a former QA professional I tried to hold my opinion back unless I had data to back it up, or I was specifically asked for it, or I had a really good working relationship with the creatives or the devs. Otherwise I knew I'd end up creating noise that someone would have to spend precious time dealing with for little gain.
Ahhh so you are talking about a glorified spell/conflict checker.
 

As an Amazon Associate, rpgcodex.net earns from qualifying purchases.
Back
Top Bottom