Is there such a thing as too much iteration?

[Klingers]

TLDR: Blizzard dick around too much in their expansion development process changing small stuff on the fly rather than locking stuff down early on paper and developing that, and that's the main reason for all the frustrating extended content droughts.
TLDR of the TLDR: Blizzard are hopeless at sticking to a development timeframe because reasons.

This isn't a rant thread about content being cut or anything. Rather, this relates to a small part of one forum thread Celeston posted today about Arms warriors:

So, we don't often share our thoughts on upcoming changes this early, as plans can change very rapidly. Please keep in mind that none of the following is set in stone – it hasn't even been developed yet – and there’s any number of issues that could cause us to decide to take a different approach.

This isn't about Arms warriors by the way. More generally, at some point one has to ask: Why are they still theorycrafting nine months out from the last content release?

I can get behind (wanky sounding phrase here, but I don't know how to put it any differently) agile software development methodology to a point, it gives you some wiggle room to iterate on ideas and refine a design on the fly.

But when you're about five months out from your projected "final final final totally final" release date window and you're thinking it's OK to totally redesign core class mechanics from the ground up, you're seriously disconnected from your customers. This crap takes time they don't have.

Blizzard are at this point basically faffing about. Dicking around. Playing tiddlywinks with a lamp-post. Throwing monkey poo at a wall and seeing what sticks. Meanwhile a train is rushing towards them at high speed.

You need to prioritise these things and realise that you need to have a cut-off point close to release where you say "Right. This is what we have as our feature set, these are our class designs, it's all finalised, let's get everything on this list finished off and tested!"

Serious system and mechanical game changes are the kinds of things that should've been whiteboarded, designed, built, pushed out to internal testers and finalised half a year ago. It doesn't have to be finished by this point, but you should know WHAT your goals are and WHAT a bug-free version of something will look like.

Am I alone in thinking this is really the core cultural problem at Blizzard right now?

[Korgoth]

It signals to me that they lack any real vision for what they wanted to do with arms. With the removal of Overpower, Heroic Strike, and Deep Wounds it also shows they have no care for its history either.

Patch note after patch note dismantling what Warriors were, with no guidance as to what they want Warriors to become. I have not pre-ordered WoD, as I did every other expansion. This is why.

So yes there is a thing as too much iteration.

[Klingers]

It signals to me that they lack any real vision for what they wanted to do with arms. With the removal of Overpower, Heroic Strike, and Deep Wounds it also shows they have no care for its history either.

Patch note after patch note dismantling what Warriors were, with no guidance as to what they want Warriors to become. I have not pre-ordered WoD, as I did every other expansion. This is why.

So yes there is a thing as too much iteration.

Exactly. In your case that's understandable. But it's broader than just Warriors.

They needed to have a real vision in place far earlier in the development process and basically stick with it. Broad-stroke intentions for every feature, every class spec, every zone, all the non-story stuff, should have at minimum been dot points on a whiteboard 18 months ago.

Considering development on expansions starts far earlier than we know about them, chopping and changing is what makes us have to wait 12 months between expansions.

[Regalbeast]

Iteration is at the core of Agile development. So to answer your question....no. However, there does come a point when you really have to move on in the development process and start developing your cemented ideas.

As a developer myself, I completely understand why Blizzard doesn't end up producing or delivering everything that they talk about. I also understand why things get changed or introduced last minute.

If Blizzard DIDN'T iterate then the game would be a shit-fest, trust me.

[Korgoth]

Yeah but to begin development you typically have a good idea of what you want. You have your use cases, you have your requirements; and you build off that.

Too much iteration shows a lack of core understanding for the project at hand, even in Agile.

[Klingers]

Iteration is at the core of Agile development. So to answer your question....no. However, there does come a point when you really have to move on in the development process and start developing your cemented ideas.

As a developer myself, I completely understand why Blizzard doesn't end up producing or delivering everything that they talk about. I also understand why things get changed or introduced last minute.

If Blizzard DIDN'T iterate then the game would be a shit-fest, trust me.

I totally agree to a point... I'm also a developer. I've been involved in different agile projects of varying success, and yeah. As Korgoth said I think you need to go in with a clear vision of what you want, and at some point you have to cut your losses and move on.

Blizzard have gotten way too comfortable operating without limiters or clear-set guidelines.

[WanderingBreezes]

at some point one has to ask: Why are they still theorycrafting nine months out from the last content release?

It's fairly obvious there was an intense delay back in January to February. They went from saying "Back on Earth, we're gearing up for the beta! Remember to sign up!" to a long draught of new information aside from mulling "this, nevermind, that, nevermind" from Celestalon and then four months later "Alpha time!"

Best guess I've heard is from a friend of mine who pointed out that they changed filesystems around that time and maybe it broke something that took so long to undo that they had to start over with the alpha.

[Klingers]

It's fairly obvious there was an intense delay back in January to February. They went from saying "Back on Earth, we're gearing up for the beta! Remember to sign up!" to a long draught of new information aside from mulling "this, nevermind, that, nevermind" from Celestalon and then four months later "Alpha time!"

Best guess I've heard is from a friend of mine who pointed out that they changed filesystems around that time and maybe it broke something that took so long to undo that they had to start over with the alpha.

That sounds plausible, but you'd hope they had backups.

[Voidism]

Yeah, I've thought about it. Especially with regards to Arms like you say. Feels gutted without any real reason. Same with Blood DK, removing/changing things in a very odd manner. Like two developers with different visions both trying to do something at the same time.

Why change things that works? Arms and Blood are far from the two specs needing those kind of changes.

[clevin]

Even in Agile, you don't continue to iterate on major systems late in the process unless you find something seriously wrong or you get drastically new requirements (use cases that change core assumptions about your customer, etc). But then, it's been apparent for a long time that they were not locking things down. Look at their inability to tell us about flight one way or the other... they admitted when Bashiok did his post on it that they'd not decided whether flight was in or not with 6.1. I mean, this is a major thing. It requires rendering the land differently, it requires deciding to update flight paths or not, etc. And it's not new - but they couldn't make a call. I mean.. wtf?

The problem is many faceted... First, they're not seeing drastic sub losses (unless the last 3 months have) so they don't have that pressure. For whatever reason, the player base doesn't leave en masse at the end of expansions. Second, they design by consensus and at some point there needs to be someone to say "No" or to say "Yes, this is the direction." Design by committee is never a good idea and it leads to things like worrying about tank specs in PVP. I mean, I get why it would be nice... but we have dual specs and if they're going to just make tank specs slightly more tilted to survivability, who really cares? Someone needs to say "this isn't a problem we need to solve right now. Defer it." But no.... . Third, they don't have executive management cracking the whip over the delays. One reason is the sub loss issue. They anticipate some, but the fact is that there are a lot of people who stay subbed whether or not they play much and some who come back.

[Klingers]

Yeah, I've thought about it. Especially with regards to Arms like you say. Feels gutted without any real reason. Same with Blood DK, removing/changing things in a very odd manner. Like two developers with different visions both trying to do something at the same time.

Why change things that works? Arms and Blood are far from the two specs needing those kind of changes.

But while I see your problem, but in this context it's not a problem that they've gutted Blood DK's because reasons. (While I'm sure that you care, I don't want to cheapen that at all... It matters, but just not in the context of this thread)

The problem is that they didn't plan to gut Blood DKs 18 months ago, work out how they were going to gut Blood DKs, and then implement and test their blood DK gutting prior to rolling it out in the beta, like... Now. Rather they just kind of threw it out there NOW that they feel like things need to change.

It's bad organisation.

[Gyrotex]

Even in Agile, you don't continue to iterate on major systems late in the process unless you find something seriously wrong or you get drastically new requirements ....

Exactly.

The previous poster mentioning Agile either doesn't fully understand it, or only read about it in theory. No offense!

Iterations are agreed (by the team), numbered and time-boxed. As a delivery manager I'd also point out that there's almost always a delivery date. Sure, it moves sometimes, but generally even if re-assessment of requirements happened, something got de-scoped, etc, there's a new delivery date projected and the team commits to a number of iterations to deliver a product.

Obviously things get more fun when you start considering MVP and drop 1, 2, 3, whatever. No idea how exactly Blizzard is skinning their Agile cat - there are many ways.

To answer the OP:

Yes, there is theoretically "too much iteration", if I use your terms.

As a delivery manager I'd work to release an MVP, and then add more shit to it, basically. That's better than keeping a product forever in development, and effectively miss all the feedback I could be otherwise receiving.

[Klingers]

Exactly.

The previous poster mentioning Agile either doesn't fully understand it, or only read about it in theory. No offense!

Iterations are agreed (by the team), numbered and time-boxed. As a delivery manager I'd also point out that there's almost always a delivery date. Sure, it moves sometimes, but generally even if re-assessment of requirements happened, something got de-scoped, etc, there's a new delivery date projected and the team commits to a number of iterations to deliver a product.

Obviously things get more fun when you start considering MVP and drop 1, 2, 3, whatever. No idea how exactly Blizzard is skinning their Agile cat - there are many ways.

To answer the OP:

Yes, there is theoretically "too much iteration", if I use your terms.

As a delivery manager I'd work to release an MVP, and then add more shit to it, basically. That's better than keeping a product forever in development, and effectively miss all the feedback I could be otherwise receiving.

If you can at least try and map things out into sprints, you've got your time-boxing right there. I don't claim insight into Blizzard's inner workings but based on their interactions with customers I don't think they've been that disciplined.

What do you class as a viable minimum though? This comes back to what you're willing to cut. Viable in WoD's case really is a subjective thing for Blizzard based on what they think is an acceptable feature-set for release. Dot-points they can put on a game box.

I think really this is how I'd do it: Firstly: Get the world built. Get the models finished and the back-end tech nailed down. Liaise with the quest and content teams to get that stuff underway early.

Secondary priority would be broad-stroke class changes with a fairly concrete cut-off. Prioritise afterthe stat squish. No classes were fundamentally broken going into WoD (this is only my opinion).

Crap like move tweeking, new secondary stats, funk new itemisation, I'd leave myself a contingency plan for not getting that stuff done and make it bottom priority.

Overall, like you said: Time-boxing. For the wanky, fiddly stuff like new stats and class tweeks, I'd seriously be putting serious caveats on any changes I hadn't at least conceptualized 18 months ago.

[rda]

TLDR: Blizzard dick around too much in their expansion development process changing small stuff on the fly rather than locking stuff down early on paper and developing that, and that's the main reason for all the frustrating extended content droughts.
TLDR of the TLDR: Blizzard are hopeless at sticking to a development timeframe because reasons.

This isn't a rant thread about content being cut or anything. Rather, this relates to a small part of one forum thread Celeston posted today about Arms warriors:

So, we don't often share our thoughts on upcoming changes this early, as plans can change very rapidly. Please keep in mind that none of the following is set in stone – it hasn't even been developed yet – and there’s any number of issues that could cause us to decide to take a different approach.

This isn't about Arms warriors by the way. More generally, at some point one has to ask: Why are they still theorycrafting nine months out from the last content release?

I agree wholeheartedly.

It seems that the devs do what gives them more pleasure. Someone likes tinkering with numbers on spells and abilities, so they are doing that, spending months on top of months doing a small change, doing another small change, reverting the first change, then partly redoing it again with a twist, etc, all the while having discussions in meeting rooms, feeling important. All in the name of quality, of course.

And for what? Is the gameplay of, say, a shadow priest today really that much better than the gameplay of the same spec during WotLK? Yes, there were changes, and yes, they were to the better, perhaps, but much better yet alone revolutionary better it is not. Absolutely not. If you were in WotLK, and you'd both knew how your spec plays in MoP and how much effort went into constant revamps (on top of revamps on top of revamps) that lead to that state (which the devs already say is bad and are revamping again for WoD), you would have screamed "stop it! don't do these revamps, you are going to redo 99% of what you are doing anyway, so do something more productive instead". Yet this is exactly where we are now - they are spending a lot of time on things that we KNOW they are going to redo all over again very soon. This is madness.

This is mismanagement, plain and simple. The persons that Celestalon, for example, is reporting to, seems to be sleeping (or dead).

Also, it's really nice to be reading things like this 9 months after Blizzcon: "none of the following is set in stone – it hasn't even been developed yet". Every time I read something like this, I'd like to ask: what the ef did you do??

[tankbug]

No, what is the problem here exactly? I like how they have the balls to radically change a class at any stage of development. This is not an enterprise product with strong requirements to predictability and deadlines. They have to keep the game fresh and fun, and in a super complex game with a huge player base, they sometimes need to let of their initial design ideas if they see that it doesn't work that well in an alpha/beta phase.

[rda]

No, what is the problem here exactly? I like how they have the balls to radically change a class at any stage of development.

The problem is that a lot of dev time - which we pay for with sub money - is being wasted for basically nothing.

That's not having balls, that's being reckless and stupid.

[zorkuus]

Changing things for the sake of change seems to be their mindset atm.

[Deleted]

you can't have all set in stone in SW development (it may look wrong for people without appropriate knowlage, but this aproach is perfectly normal)

[Klingers]

you can't have all set in stone in SW development (it may look easy for people without appropriate knowlage, but this aproach is perfectly normal)

Quetesh, you're completely right. That doesn't mean though that you can't also use a broad roadmap, organise, time budget and follow a schedule.

[cadb]

No.
Modern cars are an iteration of the horsecart design, which was an iteration on the first wheelbarrows. It just slowly gets better. Sometimes it gets worse, sure, but those designs don't survive for long.