Gamasutra: The Art & Business of Making Gamesspacer
View All     RSS
October 23, 2014
arrowPress Releases
October 23, 2014
PR Newswire
View All

If you enjoy reading this site, you might also want to check out these UBM Tech sites:

Blizzard's Martens: 'Endless Iteration' The Secret To All Blizzard Games
Blizzard's Martens: 'Endless Iteration' The Secret To All Blizzard Games
September 18, 2009 | By Staff

September 18, 2009 | By Staff
More: Console/PC

Blizzard lead content designer Kevin Martens has told Gamasutra that the key to Blizzard games like Diablo III is simple enough: "endless iteration".

Talking in an in-depth new Gamasutra interview, conducted as the Diablo III team begin public showcases of the long-awaited PC title, over 9 years after Diablo II's debut.

When asked whether "the development time has been extended to a surprising degree", Martens made it clear that he thought this was an advantage, not a disadvantage:

"Here's the secret to Blizzard games, and this is a secret that won't help any of our competitors: endless iteration. We'll take something, we'll put it in the game.

Maybe we'll like it when we put it in, maybe we won't. We'll leave it in there for a while, we'll let it percolate. We'll play it and play it and play it, and then we'll come back. We might throw it all out, or we'll throw half of that out and redo it."

Martens believes in this constant iteration as a way to actually keep things fresh when making a game:

"It can be a long time, but it is fun to work on as well. That's the thing that keeps you going. Multiplayer always works, and the builds are always playable. We've played them constantly, and it's fun. You actually look forward to the weekly play session even though the game is still in progress. That's what keeps us going, and that's also why it takes so long. We'll do it over and over again until it's just right."

Martens was joined by lead technical artist Julian Love, a Blizzard North veteran who was straightforward and enthusiastic on the company's plans for Diablo III:

"[As far as] the real question in terms of how much we're going to take forward, we want to bring back all of the stuff that was great, that was fun. We certainly want to tap into what was great in the first two games and make sure all of that stuff is coming back, and pile in all the cool stuff we can to bring it over the top and make it the definitive version of the series."

The full interview with the Blizzard Diablo III duo is now available on Gamasutra, including plenty of specifics on the construction of the still release-undated game.

Related Jobs

Demiurge Studios, Inc.
Demiurge Studios, Inc. — Cambridge, Massachusetts, United States

Senior Software Engineer
CCP — Newcastle, England, United Kingdom

Senior Backend Programmer
Guerrilla Games
Guerrilla Games — Amsterdam, Netherlands

Animation System Programmer
Nexon America, Inc.
Nexon America, Inc. — El Segundo, California, United States

Localization Coordinator


Cordero W
profile image
I sort of believe this. One thing I found out about Donkey Kong Country game during production (using this game as an example) was that the game underwent countless and countless playthroughs while they tweaked things along the way. I am a strong believer in refining a game with testing and hate the fact that a lot of games are rushed due to pressure from publishers. I am curious of how rushed developers are, and have yet to experience this due to not yet being out on the field, but if it's as bad as it's said in many articles, then it certainly explains the blandness you see in many games. To think there are games out there that could have been better only if the developers had a little more time to implement their features. Having that luxury to make a game and bring it out whenever you want must be good. Especially with the budget to make it good. Now I'm not saying a budget defines a good game, but it can still affect the game's quality in the end, as long as it's done right.

Mike Smith
profile image
This is what separates great games between those that could have been great but just fell short.


Nintendo has it, Blizzard has it, Valve has it, even some indies have it (2DBoy's World of Goo, and Hidden Path's Defense Grid for example).

Sean Currie
profile image
The question is: Why don't more companies do it? I suppose it's an issue of cost (iteration generally = longer development time = more money spent on the title) but the "secret" seems so self-evident.

John Enricco
profile image
Wow, what a fluff piece.

Good luck in trying to get a publisher to give you the funds to be able to do "endless iteration." They are only a few companies that can do that in the world without going under.

IMO, Blizzard, Valve and others got lucky. They took very big risks in the dawn of the Golden Age of games about 5-7 years ago, had mega-hits and squirreled away capital so they can do this kind of development. The only other pub that comes close is Ubisoft and that's because they put out many games that give them capital but are nowhere close to critically successful (DS games, Barbie, etc.) which then then fund the AAA titles.

More money, more time, better titles. Simple as that.

This article almost seems to gloat..

Simon Ludgate
profile image
I don't think it just comes down to time and money. It's also a different corporate philosophy. In many companies where I've worked, feedback only goes down. People at the bottom (junior positions, QA, etc.) are not permitted to comment on the game. You might be a junior artist and think you have an idea about how the controls might work better, but no one wants to hear it.

This article suggests that, at Blizzard, everyone has a voice. Whether or not this is true I cannot say, but it does highlight the other key part of iteration: it's not enough to simply iterate. Each iteration has to be productive. You have to be open to feedback. It does you no good to run multiple iterations if no one is listening to the results of the iterations.

Tyler Shogren
profile image
That's certainly half the picture...

Iteration is meaningless without traction. Traction is elusive. Meaning, you can iterate, but you won't necessarily get anywhere.

You need smart and curious and restless individuals composing your entire team for this to work. Some smart, others curious, others restless is no substitute.

Jeffrey Crenshaw
profile image
It is easier said than done; having eleven million people paying you money every month is a ludicrous safety net that very few companies will ever enjoy. However, I think that emotions -- fear and apathy -- hold us back as well as the mathematics of the economy. Fear is understandable when a company's livelihood is at stake, but I see apathy as a larger problem, stemming from the conflict of interests of having many game development companies be publicly owned. Polish and iteration are long-term strategies that increase the cost of your game, likely beyond how much the ROI of that title will be increased. However, increased quality is an investment in your company image, which has long term benefits beyond the next quarter. But what do you think Joe Wallstreet, who trades during the day and hasn't played a video game since pong, wants you to work on? Polishing a game for an extra year so the fans can truly enjoy it, or jumping on a movie license and rushing something out under the guarantee that the lowest common denominator will blindly jump on it because they recognize it?

I don't see how publicly traded companies, whose owners have full legal rights over the direction of the development and legal sanctity (lawsuit potential if profits drop in a way that can be proven as "mismanagement") without any obligation to quality beyond the bare minimum to fill their wallets is anything but an enormous shortsighted blunder holding back the artform. Whatever good has come out of the industry lately (and I have become more and more impressed over the past few years), we need to give more power and more faith to the artists that craft the works if we are going to see the medium flourish. But, at least Blizzard seems to be in a good, agile position. Honestly, I can't say they don't deserve it, and it is an inspiration, perhaps a guiding light to show us all what we should strive for.

Kevin Reese
profile image
Ya everyone pretty much beat me to it --- Blizzard's 'secret' is they have enough cash reserves to polish a game until it is as glittering as a diamond.

Marc Merrill
profile image
They didn't have those reserves when they made Warcraft II, Starcraft, Diablo, etc.

The secret sauce is deeper than just having a lot of money.

shawna olwen
profile image
I think that this article is spot on, and rings with truth not just for the game industry but for any creative endeavor that values quality and building something within in a community that meets the high expectations of all involved. The ability to cast a critical eye on the project at hand is a skill that is not born from the project that starts with a budget that is just barely big enough to make the first pass of development. And so many houses are working with this mindset that I think it forms an averaged opinion on what the game industry accepts as a quality game.

The the blizzard iteration model has more of a mindset that I have seen in film and I firmly believe that the iterations directly contributes to a better product. For me revisiting ideas to see if they are working is not just surface polish, it is conscious development. Empowering the creative management to not accept art that is 80% in the vision, and doing surgery on any part of game play that is not working at any point in the cycle is part of a strong commitment to building a better game. I believe that this attitude starts with upper management and their commitment to support the team with the resources and schedule to build this vision.

Then when it is playable... the objective process/fun starts all over again with everyone taking out their objective eyeball out of their pocket, make some realistic calls on what you can change based on budget and schedule, and the imposable shifts to a wish list for v2. Kudos to Blizzard and others that put these iterative passes into their plan from day one. I think more studios should make a commitment to release games that they are proud of on every level. There is nothing wrong with the saying "it was the best game that we could make given our resources and schedule" you should just know what size of bite you can chew... and remember chewing 10 times makes it easier to digest and having eleven million people paying you money every month means you CAN chew 20x.

[User Banned]
profile image
This user violated Gamasutra’s Comment Guidelines and has been banned.

Jeff Lindsey
profile image
I think in many cases it comes down to whether the product can be marketed properly and has a successful launch. There are many examples of great games (regardless of how long they took to make) that did not generate a profit, and there are hundreds of times more games that are not great but are profitable, sometimes to excess. In the eyes of investors, be it publisher or whoever, this means quality and creativity alone will not drive a product, you need marketing muscle; you need to know what is going to make that consumer pick your game over the competition. Unfortunately, that typically leads to the quick and dirty path of creating something derivative of other games, because the thinking is that the consumer can relate to it easily, i.e. "Oh this looks like Game X which I loved, but it has (bigger or cooler aspect)."

The end result is a very risk-averse industry for the console-level/larger-budget titles, which means focusing on finding the sweet spot of ROI before diminishing returns set in. I think in most situations, the publisher is conservative in approaching that sweet spot (as it's seen as better to undershoot slightly than overshoot) which results in more mediocre games, or ones that don't get the extra polish of Blizzard/Valve/etc titles.

Joshua McDonald
profile image
Funding may be the top barrier to repeated iteration, but a close second would be feature creep. The standard development process is to over-plan what your game will have, then start cutting half-programmed features so that you can hit budget and schedule (or rather, so that you don't go as far over as you otherwise would).

One of the reasons that some indie games manage to come out polished, despite having a fraction of the budget of AAA titles, is that they take the core fun part of the game and just build that up, instead of trying to cram as many features in as they can.

If Braid was developed by a standard AAA studio, it would have had driving sections, inventory, fancy cutscenes, unnecessary NPC's, and voice acting. It would have been five times the budget and been shipped with half the polish.

Peter Schloensge
profile image
It's not only iteration. It's a combination out of iteration and prototyping. So to say in software engineering termes.

Ken Nakai
profile image
The time thing also stems from the fact that this sort of thing isn't incorporated into the schedule. I've read countless post mortems where time was wasted building in this feature or that one. It isn't about time and money, it's about practicality. If you keep smashing a square peg into a round hole over and over again, you aren't necessarily learning anything unless you look and analyze what you're doing. Iteration is more structured than just build, try, build again, try again.

Even so, there definitely is a bit of secret sauce that just comes from being able to see what's working and what's not. Part of this comes from people being able to not only say "This is fun for me" but being able to say "This is fun for a lot of people".

The "endless" part is definitely the result of Blizzard's success and ability to deliver a game "when it's done." No doubt. But you can definitely incorporate iteration and get some measure of success from it without having to sacrifice 30% of your schedule or budget. Rapid prototyping is a big part of this. Do you need a fancy menu system to see if your new combat animation looks good? No. Just build a shell of a game engine so that a decent 3D model, your new combat animation, and a nice and basic environment (or even just a background) can be rendered. Hit a key and watch the animation fly. Does it work? Same thing goes for game play.

I've seen a few post mortems talk about rapid prototypes that helped lockdown game play and other features early on so that the decision could be made relatively early as to whether or not to keep a feature or drop it.

In my world right now (web application development), we use it all the time. It's especially useful when you've got something new that people just can't wrap their heads around without being able to see/touch/feel it. Once they do, it clicks and you can go through and make more changes and check again. You'll see a lot more progress in the first few iterations than you will later on so it has a distinct benefit without having to spend months and months trying to figure it out on paper or white board.

Carlos Lopez
profile image
I think that there is more to this. Blizzard doesn't release a game when "it's done", they only do when "it's good enough". Otherwise I want you to do the next:

1. Somehow get the pre 1.10 version of the next games: Starcraft, Diablo 2, Warcraft 3. Now play it, enjoy it, if you can, get some friends on a LAN and play that.

2. Update the games to their most recent version and play it again, notice all details added, the way somethings changed but work much better now.

3. Read the patch notes, notice the other big changes.

4. Google glitches and abuses in this games that are non-existant now.

Blizzard releases patches for games years after their release. They have even released patches that generate content and change fundamental aspects of how certain dynamics work. This isn't just some balancing, tweaking or bug/glitch fixing. It goes beyond that.

I'd guess that a game at blizzard has the next phases:

1. Experimentation/prototyping. Game development begins, as the concept gains more strenght more resources are given to them. You could see them as prototypes for games, filter out the less promising ones until you get one who's basic core is good enough.

2. When the prototype has advanced to a point where you can create screenshots that look pretty sick and are able to run a full game without mayor problems, release screenshots/press release.

3. Work on the game, focusing on removing/aleviating the fustrating aspects, grab the fun stuff and make it useful within the game as well.

4. When you constantly always have fun with the game and have little fustration release closed beta.

5. With more players more fustrating aspects, or cool details should appear. You should by now have a pretty clear idea of what is what makes the game "fun" (you should have from playing it some time ago).

6. Keep pulling in more players to discover more issues. New players generally aren't as tolerant as older players who've seen worse.

7. When you can't find anything fustrating with the game, and can't seem to find any issue with it: publish

8. Stare in horror as millions buy your game, crash it, abuse it, complain about all kinds of issues that are fustrating and not fun at all and start making mods that focus on one aspect of the game while ignoring everything else (dota? tower defenses?)

9. Keep working as you did with the beta, eliminating vexing points, focusing on fun stuff, etc.

Now the problem is that fixing any issue might be the wrong solution and bring it's own problems, further polish, or removing if it's not that fun, is what you keep doing. I still have to see a Blizzard game that came out half as good as it was a year after the expansion.

Wasin Thonkaew
profile image
That's one of the ways we should take a look to.


Mickey Mullasan
profile image
Blizzard has created a big niche out of the nerd culture of the 80s and 90s. This is like trying to figure out why Coca Cola sells so well. There's no secret to their success other than their ability for consistency among their products to have brand subservience. For anyone trying to understand the real reasons behind their market dominance, you'd first have to not enjoy their products. Otherwise you're just paying lip service to your own adolescent fantasies of how great Blizzard is and not trying to look constructively at why their Lemmings Effect works so well.

Mark Morrison
profile image
@ G-Crump

You might have some of the most interesting insight here in my opinion, except you hide behind an alias. Why don't you come out from behind the curtain and help make an impression/impact in this forum ;)

Richard Cody
profile image
Who would invest in Blizzard without knowing their deal? They make tons of money at whatever pace it takes. It's a pretty sure fire deal, as an investor you're money is safe there. They have a monstrous following. Anyways, I'd imagine that's the magic that Blizzard has in keeping the higher ups to give them more time. What other company is known to that degree? Not Valve, Infinity Ward, or Naughty Dog (though they are all very popular. In the mainstream no one has any idea which Rockstar Games made which game.

You've pretty much got Bungie and Blizzard as companies who have a massive devoted following and will continuously iterate on their games. Valve could be part of this group but it takes more of a PC type to know them (I'd argue Blizzard's name carries regardless).

Basically, you could ruin that company at this point by forcing a few undercooked games out. Or you could leave them to their own business, as they're doing. and get the rewards.

Mario Campos
profile image
Hello - I don't normally comment but I wanted to give my opinion on why Blizzard has been able to iterate even though they have almost always been Publisher-funded by companies like Vivendi. The key thing isn't that Mike Morhaim is a fantastic negotiator, though he may be. It was because Blizzard was such a small part of Vivendi. For Blizzard to miss a deadline it really was no big deal, because it was less than 5% of overall revenues for that quarter. Vivendi would rather them iterate and make more money than to release early to make an impact on their stock, as Blizzard is such a small percentage of their net worth.

theo geurts
profile image
Hellgate London was a prime example of a game that was rushed. Diablo 2 was not a bad game. But it was not the game it was ment to be. Sure it was alot of fun. But it did not have the impact as D1 had.

However D2 did the thing as it was supposed todo and it was backed up by the publishers due to succes in the past.

Mike Lopez
profile image
Sure Blizzard has the insane luxury of silos of cash now to do anything they want but back in Warcraft II/III and Diablo I/II days that was not the case.

I agree with Joshua McDonald and Peter Schloensge above: anyone can afford to do endless iteration as long as they control scope and focus on prototyping early to find the fun and gameplay refinement prior to entering production and so I am surprised that more do not plan for this. A team can't build great level content that leverages the fun of core gameplay until that gameplay is fully realized and fun during prototyping/Pre-production. Likewise the team needs to focus on a smaller number of features and execute them to a higher level of polish. If you find yourselves cutting features and/or levels as you near the end of Production that is a sure sign you have taken on too much.

A better production formula would be to shoot for feature complete 75-80%of the way through Production and Content Complete 95% through Production ensuring that the team has ample polish and iteration time that does not get filled with actual feature development and initial level implementation (some teams are even doing this up into Beta). To do this a team should focus on a scalable design (feature set and level scope) and assume the bottom portion of that design (say 30% or so) will fall off and not be executed. The mind set needs to be that anything close to the floating cut line may potentially get cut and thus no features more than a few spots down the list should begin implementation until the features above are all implemented, iterated, and usable. The other key thing is to NEVER enter full Production until the core gameplay has been realized and is verified to be fun by people outside the team.

Mike Lopez
profile image
The scalable design should also be prioritized for bang-for-the buck in terms of both gameplay impact and ability to best deliver on the product vision.

Mike Lopez
profile image
@ Gomez: I would have to completely disagree. At best maybe half of the industry (the wiser half) tries to make games that way...and 99% still fail miserably. This is due to:

1) not setting and achieving key fun, gameplay, tech and art gates during Pre-Production,

2) building levels without the core gameplay both fun and in place to leverage from the start of Production,

3) scope control/kitchen sink feature mentality from the team or the publisher, and

4) assuming feature completion and content completion states of development will come together smoothly, simultaneously and on time at the end of production when in reality those both nearly always run over and eat up any remaining scheduled polish time.

This article is talking about endless iteration and that iteration precisely needs to start from day one and continue on, not just be tacked onto the end. And the core gameplay systems and mechanics cannot get into the endless iteration cycle until fully implemented so that best be the bulk of the gating goals a team sets for Design before moving out of Pre-Production. Polish time at the end of production should rarely be used on gameplay iteration but instead needs to be dedicated to tuning, balancing, difficulty progression and player usability/accessibility testing and adjustment. 6 months at the end will help tuning/balancing but not the true endless iteration he is talking about since the bulk of the content will have already been built and large changes there costly.

As for who determines the fun, teams consistently have blinders on while working on a long-term project and they utilize perfect information about how the pieces fit together when "playing" their own game - a luxury no consumer will have. If you want to humbly bring the team back into the consumer mind set and verify your assumptions on fun nothing beats a series of Usability and Playability Testing sessions starting from early in Production and continuing through final. The team can start with employees from throughout the company and then later in production bring in external consumers. There is simply no excuse for a team to be surprised at negative consumer reactions to the experience or controls post launch yet it continues to happen all too frequently throughout the industry.

Even without a mega-budget or endless schedule a team can achieve a strong Pre-Production and use the core gameplay achieved out of that as the foundation for endless iteration and solid level production. Developers need to agree upon what gates are set for Pre-Production with the Publisher and both should be willing to schedule ample time for that stage (even 40-50% of the project) and to revisit the scope once the Pre-Production gates are passed. The better the quality and predictability of the projects the better the reputation and negotiating position the team will build for the future.

Mike Lopez
profile image
And by the way I know a lot about old school endless iteration having designed and iterated on many courses, systems and mechanics for 9 versions of Road Rash. Even those would have benefited significantly from the above processes of getting the core in place and fun before diving into course production.

Benjamin Solheim
profile image
Well I'm going to show my age here, but the early blizzard games were fun but who remembers lost vikings? I have to wonder in being forced to play through a section hundreds of time, might not be what made the game great. Not because it was polished to death but because the little things that came of it, like finding a way to make repetitive tasks fun. Like you have to click on these troops and have them confirm it, so they say funny things, getting sillier with each click. I mean stop poking me? I think every time they said it I poked them again. Or clicked through all the units to find out what they would say.

So by the time World of Warcraft rolled around they had games that had been making money years after they released. So they had time, money and a reputation to draw the talent they did, to work on the getting all the pieces right the first time. They took idea from the big mmos at the time, and applied their lighthearted flavor to them. So your brutal dead plenty from EQ and crazy item grind from AC become something that was more carrot than stick. From the numbers of younger players I think they either lucked out or have some impressive foresight with the timing to get younger players just missed earlier mmos. But the simply step of asking if it is fun added to people who liked playing games, resulted in a fun game not one that should have been fun, based on some formula. Wow is a Juggernaut because half the reason people log into a mmo is to hang out with your friends and getting fifty people to do the same thing is much harder than just one.

John Manley
profile image
This article is really amusing, and sure, as someone said above, it does sound like a bit of gloatation =D but that's all part of being on the winning team, right?

To be frank, I can barely remember Diablo 1, and I never got in to 2. So I thought it was interesting to bring up talk of 3, eight years later. 3's sure have been lucky, at least when I think about Fallout 3 and how long I'd drooled over it. So my hopes for Diablo 3 are high. That said...

I really do think this is one of the more interesting philosophical arguments caused by this article. You have a few separate sides looking at it differently. Likely the truth lies somewhere in between. They have the volume of income to take their time, but that volume also gives them a Wal-Mart-esque ability to hold the bargaining chip when it comes to setting release dates and such. So while we speak of ideals, if any other developer were to say 'Look, we're holding it back till its ready' they'd probably be dropped, save a select few with important IP.

I'm pretty sure for most, as it is for me, feel this argument somehow stems from how we all feel scorned, as gamers, over the lack of quality in titles, which we got used to when we were younger. Sure, everyone WANTS to make a great game, but why does it never get done like Blizzard says it should be? Do we really believe every other developer sits back and says 'Let's only iterate 10 times on this, or whatever fits into 10 weeks'?

This article and discussion, to me, points at people trying to find the root cause for the lack of stellar games in the last 5 years, on a volume basis. Why do 99% of games fail, in the end, when all developers should be as adamant as Blizzard, and why does Blizzard believe it won't help their competition any to know their "secret"?

That's the real challenge, I believe. Figuring out, not how the winners win, but how the losers lost, despite knowing better.

Mike Lopez
profile image
@ Gomez: Yeah, it's probably just me...and the 40 or so teams I have worked with...and 20 or so other senior/exec producers I know at multiple publishers...and nearly all the teams they have worked with in their career, even those with significant experience.

Kudos to you that you have found perfection but other than a handful of top developers that you cited the rest of the industry is not as lucky or insightful apparently. It must be amazing to be so great.

My point about blinders was not to replace the team decision but only to base them on real consumer impressions and not only those of the hard-core players on the team with perfect information about the systems/mechanics/missions. Microsoft uses Usability/Playability testing very effectively and they begin that testing probably a year or so earlier than most teams in the industry so that process is something all teams can benefit from, not just those you deem smart.

Also Cerny or Miyamoto would have a team prototype gameplay for years sometimes but other than the Blizzard-tiered teams of this world nobody has that luxury. The next best thing is to direct a pre-production gated with specific discipline and playable goals and not just randomly determined with x number of months. You say you know this. I get that. Yet very few teams fully embrace this and most are hihgly resistant to process change. Even great blockbuster games like Bioshock suffered very expensive content throwaway because the gameplay experience only came together at what would have normally been Alpha and after the bulk of the content had already been built (in that case they had the luxury of redoing everything for almost an extra year). This happens all the time in this least outside of the Blizzard-tiered few and apparently your own golden walls.

Since you know all and have the perfect team this info is aimed at those less insightful/successful. I concede that you are better than the rest of us (surely why you use the pseudonym).

Randle Reece
profile image
Everyone seems to have missed the fact that the first thing Blizzard released after being acquired by Activision was Wrath of the Lich King, which was rushed to make the holidays, included last-minute changes that weren't even beta tested in a desperate attempt to address glaring problems, led to stagnation in the US/Europe/Oceania subscriber base, and was unapprovable in China.