William Stiernberg

The first draft: How the final version of Rain-slick Precipice 3 differed from the original design

The first draft: How the final version of Rain-slick Precipice 3 differed from the original design

Robert Boyd of Zeboyd Games tweeted about finding the original design document for Rain-Slick 3, and I invited him to write some thoughts on what changes were made during the course of development. This guest post is an interesting look at how many details are fluid during the course of a game’s development, and why certain decisions are made. Enjoy!

Planning out a game and then actually making it according to that plan is something that rarely happens in game development. New ideas are thought up, old ideas turn out to have drawbacks that weren’t thought of, and before you know it, an entire game can end up as something drastically different than what was originally written out in the design document.

Here are some of the many changes that Penny Arcade’s On the Rain-Slick Precipice of Darkness 3 underwent during development.

Playable Characters

Since we knew that we wanted to use a multi-class system for character progression, we decided early on to keep the playable character roster small. Jerry Holkins (aka Tycho) said that he wanted the focus to be on Tycho and Gabe and didn’t want much, if any, dialogue to be coming from the other 2 main characters. Because of this, we decided on Jim the Skull (from the comics and also viewed in the previous 2 games) and Polly the Mime Parrot, both silent characters.

However, about halfway through development, Jeff Kalles (our producer) and Jerry came to us and said that they thought the game needed a good, strong female character. Their initial suggestion was to just take the Mime Parrot and make it human but I was having trouble getting the Mime class to work like I wanted to, (I was going to give it the ability to copy ally & enemy abilities but this was proving difficult to debug) so I suggested a red-haired trenchcoat-wearing gumshoe. Thus Moira came into the game.

Combat systems

The big “revolutionary” gameplay feature that Rain-Slick 3 was going to have was called the Side Effect system. Instead of using MP to cast spells, all abilities would have a Side Effect that they added to the character who used it which would be a buff of some kind with each ally and enemy able to have up to 5 Side Effects (SE) at one time (after which the newer Side Effects would start to replace the older ones).

The idea was that some of the weaker abilities would have stronger buffs, and vice versa, and the whole thing would balance itself out. Really powerful abilities (Super Abilities) would have a cooldown timer to further balance things. Even generic commands like Attack & Defend would have SEs attached to them. For example:

Gabe’s Attack SE: Speed +20%
Gabe’s Defend SE: Damage +50% the turn after defending

Notes: Using Defend twice in a row and then attacking or using an ability will give you a +100% Damage bonus to the attack or ability. But an even better move would be to defend, attack, defend, attack since the first attack would get a +50% Damage bonus and the second one would get a +100% bonus.

I loved the system and thought it had a great deal of potential for making combat interesting since you were effectively getting temporary level-ups each turn. But the challenge of making each ability interesting AND balanced turned out to be too much for me (Jeff respectfully refrained from giving me a big “I told you so”). We ended up replacing it with the “MP starts at 0 and goes up by 1 each turn” system which turned out to be a fun and interesting system in and of itself, and much easier to understand. I still think the Side Effect system could work but it would require an extensive amount of iteration, fine-tuning, and testing.


Originally, I wanted the game’s maps to be entirely path-based like the world map is now. It seemed the natural next step in our quest to speed up the pacing and remove the filler - walk down a path, activate a node, fight a battle or see the story progress, and repeat. However, once we had the game up in running like this, several people (aka everyone on the dev team EXCEPT for me) expressed concerns that this setup was too constricting and that they missed the ability to freely move about the maps.

Jerry mentioned the game felt “hemmed in” and Bill Stiernberg wanted to take the game “off the rails.” It took a lot of convincing but eventually I agreed that they were right and we ended up modifying every map in the game because of this change. It was a pain to have to redo all of the maps like this but we ended up discovering a more efficient way to make maps in the process, so it ended up being for the best. Our new map process also allowed us to upgrade our lighting system and use actual shadows and lighting effects (our previous games had shadows that the characters would just walk on top of).


Rain-Slick 3 was going to have a similar scope as our previous game, Cthulhu Saves the World. We intended to have about as many enemies as Cthulhu Saves the World (120 or so), and for the enemies not to have animations, like the old SNES Final Fantasy games. Once we had our side-view battle system initially in place, with animated player characters, Bill decided that it would be more visually interesting if we could also animate the enemies.

So after overhauling all of the maps in the game to fit the new system/design, Bill went after his enemy sprites and took the time to animate an idle, attack, damage, and sometimes special animation sequence for all of the 100+ enemies in the game. Battles became much more visually interesting as a result. On the downside, with the upgrades in visuals for enemies and maps along with the increase in overall game size, Rain-Slick 3 ended up taking much longer to make than we originally planned, about a year as opposed to 6 months.


Several classes ended up being cut for various reasons. Some of these included the Deep Crow Rider, the Mobster, and the Tarotmancer.

The Deep Crow Rider was going to be a combination of the Dragoon from the Final Fantasy series and a summoner. In Rider mode, you would gain bonuses from the Deep Crow and you could use the iconic Jump command (although in this case, it would be Fly). However, you could also choose to summon the Deep Crow separately, at which point you’d lose those abilities and instead the Deep Crow would attack on its own each turn. We decided that the class didn’t really fill a needed niche and visually it would have been hard to make it look as awesome as we wanted with our current engine, so we eliminated it.

The Mobster was full of abilities that relied on luck to succeed (similar to many of Ronfar’s abilities from Lunar: Eternal Blue). In the end, we decided that having a class entirely based on luck was a bad idea so it got removed.

The Tarotmancer was another mage-oriented class with a wide variety of abilities. The catch with the Tarotmancer was that only one random ability would be selectable each turn (the Tarot deck gives you a different card/ability). It was less reliant on chance than the Mobster since you could see what your ability was before deciding to use it but like the Mobster, but we decided we didn’t want to have chance factor too much into the game and so we cancelled it.

Other classes that were on the possibilities list that didn’t really get fleshed out or got removed because they were too similar to other classes we were already using included the Pugilist (brawler), Hippie (Healer), Inventor (like the gadget-user in Earthbound), Hooligan (Thief), Slumlord (Poison + Summons), Politician (Debuff), Pyro (Fire attacks and counter-attacks), Psychoic (Mage), Idiot (Joke class), Ragtimer (Bard), Casanova (Healer/Buffs), Urlchemist (don’t remember), and Surfer (what were we thinking?).

Story Changes (Warning: MASSIVE SPOILERS

My original plan was to have the party fight Dr. Blood in the Temple to Man, and then when Tycho is about to die, he thinks he gets thrown back in time. What actually happens is that his soul is trapped in an illusion by Yog Modaigh and to escape it, he has to fight Yog Modaigh in the Periphery to break out. Jerry didn’t like this idea and it’s his story so instead we stuck the Yog Modaigh stuff in as a side-quest that you had to do in order to get the true ending.

We were also going to have another dungeon after where the game ends now (and Bill actually drew all the art for it) but Jerry thought that would be better used early on in Rain-Slick 4 so we’re saving it for later.

Cut for Time

It’s tempting to keep adding more and more cool stuff to your game but eventually, you have to decide that enough is enough and that you have to actually finish and release the game. Some of the stuff we ended up cutting for time reasons include:

A section where you play as Dr. Blood & The Leviathan.

A much larger Periphery. As the player found different periphery maps, they would unlock new parts of the Periphery, each of which had someone in a Parenthetical for various reasons.

Flashback scenes between Tycho and his father.

An interactive Atari 2600-esque battle (the book in the Attic).

Additional special battle conditions like enemies getting reinforcements and battles where your classes & levels are set for you (puzzle battles).

Zelda/Lufia 2-style puzzles in the dungeons.


There’s a reason why so many developers are fans of rapid iteration - games frequently change drastically during development and the faster you can figure out what works and what doesn’t, the better. And once a game is further along, playtesting (both internally among the developers and externally with a selected group of fans) is a great way to fine-tune and polish a game concept. Rain-Slick 3 received much more playtesting than our previous games and it greatly benefited from the refining process. Many of the changes were minor things that people generally don’t think about when they think about game development. For example, the “Press a button to begin the battle” message was added into the game after seeing several people wait for the battle to start automatically.

In the end, we’re very pleased with the released version of Rain-Slick 3 and learned a lot about game development in the process. And not every idea that was discarded in Rain-Slick 3 is necessarily gone for good - no doubt, many of the features we cut for Rain-Slick 3 will end up finding their way into future games of ours in one form or another.