Dragon Quest II – 1987 Interview

Dragon Quest II – 1987 Interview

In this 1987 interview Yuji Horii, creator of the Dragon Quest series, talks about the challenge of balancing the game difficulty, Famicom memory limitations, and an intriguing alternative ending that foreshadows the more cinematic direction modern RPGs would follow. This interview was found at the GSLA, a Japanese website that preserves game developer interviews from older, now-defunct print sources.

Delays and Difficulty Balancing

We started planning Dragon Quest II about 7 months after Dragon Quest was released. It took a little longer than we had planned at first, but by the start of November we had the main programming mostly completed. However! As you know, DQII wasn’t released until January 26th of this year. Technically we were able to release it by the end of last year, so why did it get extended a whole nother month?

The answer is simple: the programming was done, but when we actually played it we found it far, far too difficult. At that point, the game’s balance still wasn’t right, and fixing it delayed us for a month.

To explain the problem more concretely, the balance between your character’s strength and the monsters’ strength wasn’t right. Of course we couldn’t just haphazardly modify everything; instead, Nakamura created a fight simulation program and we went through each monster one by one, adjusting the player level data and monster data individually.

Yuji Horii

But even after those adjustments, when we went back to the game and tried it the balance sitll was off. Why? Two reasons. The first is that we adjusted the monsters’ strength in the simulator one by one, with an eye for making each individual fight tense and challenging. But when we tried it in the game, consecutive fights made it way too difficult. The second reason is that we had underestimated the difficulty of battles with multiple monsters. As a result our first pass at DQII was a game you could hardly play. You’d die right away… you couldn’t even make it to the Prince of Cannock!

Because of this, we delayed the release for a month and spent that time balancing the game. In the beginning of the game, when the player was still alone with no companions, he would die constantly. To solve this problem we limited the number of monsters that could appear in a battle on the first continent to only 3. That made it a lot better, but it was still too hard. The problem is that the Prince of Midenhall1 is a fighter. No matter how much you level him he’ll never learn Heal, so he always has to use medical herbs and stay at inns. In other words, the further he travels from town, the harder it gets.

To address that we moved the location of Cannock Castle2 to be closer. Cannock Castle was originally located where the Lake Cave is now. But this was too far from Leftwynne Village3 where there was an inn. So we changed the Lake Cave dungeon to be a place you’d go after getting the Prince of Cannock, and we moved Cannock Castle much closer.

(By the way, in the Dragon Quest II LP record, the back of the jacket shows a castle surrounded by a lake. This was the original Cannock Castle. We finished the jacket design for the LP fairly early, so it ended up having that first version pictured. Interested parties can check it out at their local record store!)

Original location of Cannock Castle.

But this meant that the lake in the far east where Cannock Castle had been was now a rather pointless area. We’d spent all that time working on the landscape so it seemed like a waste! So we put the Lake Cave dungeon where Cannock Castle was, rearranged the distribution of monsters on the map, and moved the location of certain items to retain the game balance.

We then did further adjustments on the monster data and player levels, as well as the price and strength of weapons/armor.

It goes without saying that finding the right balance is key to a good RPG. And to adjust that balance there’s no other way but to actually playtest the game, make some small adjustments, test it again, make more adjustments, and so on. Practically speaking this meant that all of our staff, as well as many playtesters working part time, were all playing through the game during this period. Their feedback (as well as their complaints) were then all conveyed directly to me.

“I can’t make any money. Lower the prices please.”

“Battles take too long. Can’t you make them quicker?”

“Can we adjust the speed at which you level up? The transition from level 4 to level 5 is especially rough.”

“No, I think the leveling is fine as it is now, but I’m still dying too much.”

…and so on and so forth. I’d listen to their comments and then play it myself, adjusting the data for parts I thought were off. As soon as I finished I’d send a fax to Chunsoft, where Nakamura would swap the old data for my updated data and make a new prototype version from that. Then I’d distribute that back to all the playtester and the process would begin again.

Naturally, game balancing isn’t something you get right after one pass. We repeated this process over and over. At first it went in 4-5 day increments, and soon after 2-3, and then we were getting new updated prototypes daily. At that point we had to make sure every prototyped version was clearly labeled with a date and time: “December 8th, 4 A.M. version.” Without that you’d have no way of knowing what the newest version was.

Adjusting the order of spells learned… changing the strength of monsters… lowering the prices at the church… in this way, the month went by in the blink of an eye. We had a time limit after all. I wouldn’t say we found the 100% ideal balance, but I think we came 90% of the way. If we had more time I think we could have made it perfect, but I didn’t want to make the children wait any longer.

By mid-December we had finished the final version. But at that moment I collapsed from a stomach ulcer. Ah, it was really horrible.

Memory Problems

The memory limitations we had with DQII caused us no shortage of strife.

The first unfortunate victim were the large kamishibai-esque pictures we planned to use for cut scenes and such. We put them in at first, but as the graphics memory got larger and larger in the end we had to remove them and abandon the idea. I imagine few people have noticed this, but we stealthily printed one of those images in the story section of the DQII instruction manual. Of course on the famicom it would have actually been in color.

The missing cut-scene mentioned above, from the Dragon Quest II manual.

And we also wanted to use a full-size picture for a scenario we had planned involving the sunken treasure. Originally, if you looked out from the top of lighthouse in the direction of Rhone, you would see a picture of the Rhone’s landscape, and in the sea just beyond there would be a single sparkling point of light. That was where the sunken treasure was supposed to be, and there was a scenario where you’d talk to people and find out more info about it. But when the picture was too big for the memory, we had to change the scenario.

One of the sillier ideas we abandoned was the “abunai mizugi” (dangerous swimsuit). This was supposed to be a piece of armor as strong as the Mink Coat, and incredibly expensive. If you bought it for the Princess, her character sprite was actually supposed to change and show her wearing it.4

Alternate Endings

We also wanted to do more for the ending. After you arrive at the throne and are seated as King, we wanted various people from the game to come one by one and pay their respects. The final scene would have been a glamorous courtroom dance.

Another idea we had for the ending was this: in the final battle, the Prince of Cannock would sacrifice himself to defeat Hargon. Your quest would be complete, but the Prince of Cannock would not return. Wishing him peace in the next life, you and the Princess of Moonbrooke would return to Middenhall. The people there would be praising your great exploits, and a grand ceremony would begin…

But then!

A single girl would rush up to you, yelling “You killled my brother!” And you realize only too late the dagger in her hand, as she plunges it deep into your breast. This, of course, was the younger sister of the Prince of Cannock. The people are shocked. Life slowly fades from your body…

But that ending was far too sad, so in the end we decided against it. And yet I still really want to make something with that kind of atmosphere if I ever have the chance. I want to make something beautiful and sad, a tearful RPG (though when I’ll have that chance, who knows…). A game that while you’re playing it, tears well up in your eyes without your even being aware. I think that would make for a good game.

Those were the main things we left out due to memory limitations. There were also a number of items we had to remove… so if you find an empty treasure chest, that’s why. But perhaps it will be fun if use your rich imagination and think “what was supposed to be in this chest?”

By which I mean to say, now that Dragon Quest II is released, its out of our hands as designers. All we have done is prepare the world; the tales and mysteries you experience there are up to you. I only hope you’ll be left with joyful memories. Parupunte~!5

A cool handmade diorama of the final battle with Hargon.

If you've enjoyed reading this interview and would like to be able to vote each month on what I translate, please consider supporting me on Patreon! I can't do it without your help!


  1. Actually called the Prince of Lorasia in Japanese and the localized GBC version.

  2. Actually called Samaltria Castle in Japanese, but I’ll use the English names for easier reading.

  3. Liliza Village in the Japanese version.

  4. In a slightly modified form, this was included in the MSX version. If you talk to a certain King when the Princess has nothing equipped, he will give her the abunai mizugi and a picture will appear on screen.

  5. The Japanese spells all have different names; Parupunte is for the spell Chance.

Leave a comment

Your email address will not be published. Required fields are marked *