5. Halting Development Right Before Master Up

Iwata - Usually when making a game you’re putting your own blood, sweat and tears into writing the setting, deciding on the actions, and that kind of thing - but this time around all of the data is stored on Google’s servers, so from the beginning the idea was that you’d just take the data that worked for you and apply that to the game, right?

Soya - Right. That’s why we figured we’d be done in no time.

Nishimura - Originally it was supposed to be a fairly short project.

Iwata - That’s what I heard as well.

Nishimura - It was supposed to be easily achievable, so we thought it would be short, anyhow.

Iwata - Although you say “easily achievable,” it took three whole years.

Nishimura - Yes. We started working in June of 2007, and after about six months, near the end of the year, we were getting to the point where the end was in sight, and that’s when we began to see what parts just didn’t work as a game.

Iwata - That’s one of the nightmare scenarios for development - just as you’re starting to see the light at the end of the tunnel, you realize it actually doesn’t work as a game.

Nishimura - Indeed… it really was a nightmare. For instance, first a keyword like “car” would display, and then if a related word like “driver” or “wheel” or “gasoline” were to show up next they would match perfectly, and that’s what your player would pick. But words just wouldn’t show up as expected.

Iwata - Words totally unrelated to “car” would show up, then.

Nishimura - Right. “Fire” or “swimming” - words that were totally unrelated would show up.

Iwata - “Car” with “swimming!” (laughs)

Nishimura - At that time we had about 4,000 words, and those words would automatically be picked up randomly. So at that point when we had testers play the game they said, “It’s like playing the lottery, you have to leave it all up to luck.”

Iwata - It’s rough to have to hear that this kind of game is like “playing the lottery.”

Nishimura - Yeah. Even if the ideas for the new modes were incredibly interesting, when the words we expected didn’t show up, we finally realized that it just wasn’t fun. At that point we decided to group each of those 4,000 words by characteristic. For instance, grouping “car” and “gasoline” together.

Iwata - Did you do that all manually?

Soya - Basically, yes. We had 3-40 different staff members make those decisions.

Iwata - The point of Google’s search engine is that everything is supposed to be automated, so originally this software was supposed to follow that way of thinking, but in the end you had no choice but to rely on man power to finish it up.

Soya - Yes. It was like we were little ants up against this absolutely massive statue.

Yui - But that grouping ended up being subjective, so in the end it still felt awkward. And then, as luck would have it, we happened to stumble upon a programmer who specialised in Japanese processing systems, and he proposed that we use co-occurrence from Japanese text conversion.

Iwata - Co-occurrence?

Yui - Written as “together” and “occurring” in Japanese.

Iwata - Ah, “co-occurrence” (※6) - I see. Two words that occur together, in other words, you used a kind of algorithm that would group words that would have a high chance of being searched for together.

※6 “Co-occurrence”＝ Indicates the high relate-ability of two words given their chances of the two of them appearing together.

Yui - Right. In the Japanese processing world it’s apparently a very common term. We used that program to rearrange all of the data we had previously included. And that’s how we got rid of the awkwardness.

Iwata - Indeed, having played this software myself I almost never felt that any word combination was totally out of left field. By the way, how many words are in the game now?

Soya - Over 10,000.

Iwata - Having tried all kinds of things from automating that process to grouping everything together manually, about a year ago an incredibly “dark event” occurred that you’d rather not remember, didn’t it, Nishimura-san? Shall we start discussing that episode?

Nishimura - Sure. That was… about a year ago from now, around January of 2009. We had already entered debug, and had finally gotten to a point where we’d be done soon.

Iwata - And that’s when you experienced a complete reversal of fortune.

Nishimura - ….yes.

Iwata - Let’s hear it!

Nishimura - O.. Oh… You want me to start off (laughs). Um… Originally the name of the game … we never found anything that just clicked…

Iwata - Right. In my tenure as CEO no other title has had its name rejected so many times.

Nishimura - Right. It really was a battle. The thing is, as we were going back and forth about this, we started to think that maybe the reason the title was so hard to pick was because the content of the game wasn’t understandable enough. Furthermore, even having added more modes, we had heard from people that the single player still wasn’t fun, and that “AND-search is hard to comprehend,” so the conversation went all the way to “Is it really okay for us to put this game out as it is?”

Iwata - Right. And at that point development was actually put on hold. Nishimura-san, how did you feel when that decision was made?

Nishimura - That… I don’t have much recollection.

Iwata - What?! You don’t remember?

Nishimura - Yeah, it really did a number on me. And also it pretty much overlapped with a period where I was really busy with the Game Seminar.

Iwata - Just as the game you were handling development of was being put on hold, the students from the Game Seminar were entering their final stages of development and they were requiring your support.

Nishimura - Yes.

Iwata - When Nishimura-san explained that development had been put on hold, what did that sounds like to you at Shift? Here you were in the final push to complete the game, it must have felt like, “What are these guys saying?!”

Yui - Yes. To be honest, I’ve been in this industry ten years and put out seven titles, but it was the first time I had to deal with something like this.

Iwata - It’s the kind of thing that really shouldn’t happen.

Soya - But it was the first time that I had seen it first hand - that Nintendo will put the stop on development for games that they don’t feel are ready, even right before master-up. Even having heard about it before, it was quite a surprise.