Featured Blog | This community-written post highlights the best of what the game industry has to offer. Read more like it on the Game Developer Blogs or learn how to Submit Your Own Blog Post
This is our official post-mortem for our latest game, Hero-U: Rogue to Redemption. It took us 6 years to make this game, and we learned a lot from the process. Post-mortem by our programmer, Cidney Hamilton and edited by both Lori and Corey Cole.
HERO-U: ROGUE TO REDEMPTION POST MORTEM
Written by: Cidney Hamiltion, Programmer
I joined the Hero-U: Rogue to Redemption dev team at the beginning of 2015. By then, the first Kickstarter was two years in the past, and the second was in the works. The Coles anticipated a 2016 release date. In reality, the game would not start alpha testing until 2017. We shipped the game in July 2018, and it’s currently available on Steam, GOG, Humble Bundle, and the Mac App Store.
Hero-U began as the Famous Adventurer’s School for Heroes. The Famous Adventurer is the guy who wrote the Famous Adventurer’s Correspondence School pamphlets that were included as “feelies” in the original Quest for Glory games by Corey and Lori Cole. Quest for Glory, of course, is licensed to Activision, but the world was based on the Coles’ tabletop campaign. In 2008, the Coles created an online roleplaying game around teaching everyone how to be a hero in everyday life. The website included a Test that sorted would-be heroes into classes (Rogue, Wizard, Paladin, and Warrior), and in-character lessons from the various instructors. Paladins got rewarded for doing good deeds; Wizards got rewarded for doing something clever; and Rogues got rewarded for getting away with questionable (though still heroic) deeds.
The School for Heroes eventually became too big to manage. Moderating this website and roleplaying the teachers became an enormous project! In-universe, it closed and relocated to Sardonia to become a University and accept new applicants. Hero University would be a single-player video game that simulated the experience of learning to become a hero. The player character (in the first game, an aspiring thief named Shawn O’Connor) who goes to class, fights monsters in the school dungeons, and eventually succeeds or fails in becoming a hero.
The Coles had kicked around the idea of building a text adventure, but decided that they needed graphics. Other old-school adventure game developers were taking to Kickstarter to launch their own studios; this was how Broken Age, Moebius, and Leisure Suit Larry Reloaded were funded. Fulfilling this project, though, required building a team.
I had been a software engineer in Boston. I’d originally started in a hybrid designer/developer role, with assurances that I had flexible hours and could work around my as-yet-undiagnosed genetic disorder.
As time went on I had less flexibility, and quit my job in favor of short term freelance projects. I’d been working on a short game in Unity when I saw that the Coles were looking for additional programmers. I reached out, interviewed, and joined the remote team in January 2015.
At first, I was struggling to get up to speed with where the game was. Corey estimated that 90% of the art and writing was finished, but only 10% of the programming. In reality, there was much more left to do. Many of the art assets were not final, especially the 3D environments. The writing backlog was long enough that Coles ended up hiring another writer (Josh Mandel, of King’s Quest and Freddy Pharkas fame) to help out.
Only the first scene of the game was programmed; the rest were just grey-boxed. Lori had written scripts for most of the game, but they were totally untested and untestable. The progress thus far had largely been on the art side. Core systems (including combat and advancing the calendar on a game that took place over a 50-day semester) had yet to be implemented.
We were focused on getting a playable demo for the Kickstarter, but there was clearly much more than a year of development time left to have a finished game. The scope had also increased. The original Hero-U pitch had been a 2D puzzle/RPG, which would have been relatively easy to program. But the backers, by and large, wanted something that was more like Quest for Glory. The Coles also wanted to do something more ambitious, but it had been 20 years since they last managed their own studio. Working with a remote, part-time team is also very different than being in an office every day and fighting externally imposed deadlines.
The second Kickstarter made its goal, but didn’t vastly exceed it. Stretch goals (such as voice acting, some boss fights, and a companion NPC) had to be scrapped.
The major challenge of Hero-U was that the game was not ever going to be Quest for Glory. The Coles were new to producing a game; we didn’t have the resources and structure of Sierra. Corey based the budget on the early low-budget Quest for Glory games, which used 16-color art and relied on Sierra’s development tools. Fans wanted a game like Quest for Glory IV, which had a $1.5 million budget in 2015 dollars. The Kickstarter provided 25% of that.
I had backed Jane Jensen’s Moebius. The game was released on time, with a $300,000 budget, but it suffered due to the inevitable comparisons to Gabriel Knight. Hero-U was even more ambitious; it needed to be an adventure game and a roleplaying game, have the same humor and heart as the Quest for Glory games, and allow nonlinearity and choices in how to play your character. We could have shipped sooner by cutting features; the Coles chose not to do this. If they had, it would not have come close to the expectations of Quest for Glory.
Simply put, there’s nothing on the market that’s quite like Hero-U. It’s not another Quest for Glory at all; Quest for Infamy and Heroine’s Quest come much closer to the Quest for Glory style and mechanics. It’s a fantasy school simulator; not another rags-to-riches save-the-world RPG. While you can explore dungeons and fight in combat, you never leave the school grounds, there’s no epic villain to defeat, and you don’t save the world. At least not in this game.
The player spends most of their time going to class and learning how to be a hero, both in theory and in practice. It’s been compared to the Sims and Harry Potter; but there’s really nothing like this.
We had promised some of our Kickstarter backers alpha and beta test privileges. The alpha test begun well before the game’s dungeons were finished, but it allowed us to get feedback from our supporters early enough to make major design changes (though unfortunately, our art budget was exhausted at this point!). The feedback was also great for team morale and productivity.
We had an in-game bug reporter, which allowed us to find and crush 3000 backer-reported bugs during this phase. Our supporters also offered their own expertise on cultural issues, German language, music arrangements (our composer wasn’t involved in the game’s final sound design), and game balance and combat.
Our fans were especially helpful during crunchtime, when morale was low and the team just wanted to get the game released as soon as possible. Having a large fanbase to tap into was an amazing asset; we couldn’t have done it without this support.
We originally planned for a top-down, cartoony tiled 2D art look for the game. We had technical and aesthetic issues with this approach - the unrealistic style didn’t work well with our immersive, serious story. We then tried using a proscenium stage (2.5D) approach similar to Sierra and LucasArts adventure games. Unfortunately, we had even more technical difficulties working with 2D objects in a 3D space. Our character movements looked stilted and unnatural. The tiled backgrounds looked flat and dull with no lighting. We made the decision to start over using 3D technology to animate the characters and bring the backgrounds to life.
Due to the nature of our production process, in which several artist contractors left the team for other obligations and opportunities, our two remaining artists had to come up to speed -- fast -- on art processes outside their respective wheelhouses. Through physically remote but functionally close collaboration, we were able to facilitate this as a team, successfully implementing lighting, particle effects, and additional modeling in a relatively painless way.
Hero-U had a very granular time system. The clock was always ticking. Classes, dinnertime, and bedtime all took place on a set schedule. Usually the player character was forced to go to bed at 1am, but some events had him up late at night.
In concept, this was really cool. This was how the Quest for Glory clock worked; it was realistic. Most simulation games and JRPGs have a very abstract concept of time; either an action takes no time, or it takes the entire night. This allows for time optimization logic puzzles, such as maximizing all of the social links in Persona 3, but it doesn’t allow for flexibility and roleplaying.
Hero-U, in contrast, gave you several hours in the afternoon. You could spend time in the practice room, or run around the castle to wrap up an errand for the school receptionist, or grab the school supplies you need from the store in the basement, and then still make your elective class.
It turned out that this was a nightmare to code, and a nightmare to design for. The beginning of the game felt overwhelming for a lot of players; the constant sense of the clock ticking discouraged exploration. The ending was also a bit empty, and there were sections of the game that were just slow. I fixed dozens of bugs where staying up past midnight broke scripting conditionals and made the game think the plot had already advanced to the following day.
By the time the programming team was assembled, the game script had already been written, the character models were already in the game, and most of this was final. This led to a lot of issues that couldn’t cheaply be solved, because so much of the game’s budget had already gone into creating the assets we had.
The character models weren’t compatible with Unity’s Mecanim animation system, which was released in the middle of our production. While we hired an animator who was able to make a lot of improvements, cutscenes were often flat because it was difficult to animate the NPC rigs within Unity. Shawn’s weapons and armor didn’t show up in the game-world, which many playtesters reported as a bug. In fact, his model was never “generic” enough to easily accommodate costume changes, even with texture swapping.
We were able to redo much of the game art-- including the user interface, 3D environments, and in-game vignette paintings-- but all of this should have been something the designers accounted for in pre-production, not something that came up at the last minute.
Hero-U: Rogue to Redemption shipped in July of 2018, two years after the release date promised in the Kickstarter. The game appeared quietly on Steam as soon as we had a stable version of the game in beta test. All of the developers, including me, were busy fixing bugs and getting the game solvable. The designers were doing all of the production materials, including the trailer and press contacts, themselves.
This resulted in a “stealth” release, with little buildup and very little coordination. We hit Steam right during a bug with Steam’s discovery algorithms, impacting most indies released in late 2018 - Six months after launch, Hero-U was still listed as “Steam is still learning about this game,” which likely reduced player confidence. Many people are still telling us “wow, we didn’t know that Hero-U was out yet!” on social media.
At the advice of an investor, Transolar chose to release the game at a premium price point of $35 to position it as a high-end game. Most indie games on Steam have a $10, $15, or $20 price. Unfortunately, we didn’t have enough of a marketing push to be treated as a premium AAA game. Companies such as Electronic Arts and Activision spend millions of dollars in advertising to promote their AAA games, which pushes most other games into the “cheap indie game” category.
We were unable to contact Steam to get a personal representative, so were flying blind on pricing, marketing strategy, and so on. Steam is largely built around automated algorithms. In particular, reaching $100,000 in first-month sales and maintaining a similar level is critical to getting premium “front page” placement during Steam promotions. We did not come close to hitting that level.
In early 2019, we reduced the game price to $20. Sales since then have been modest, but are a steady trickle. Steam buyers are used to buying most games during sales; we have not yet participated in a sale since lowering the base price.
60% of Steam’s market is non-English speaking countries. We launched only in English, and localized versions might improve sales. However, localization is expensive due to the huge amount of text and dialogue in the game and the many puns, jokes, and pop culture references.
While we’ve struggled to learn from our mistakes, navigate the world of indie game marketing and distribution, and keep Transolar Games around, our fans keep asking us when the next Hero-U game - subtitled Wizards Way - will be out. We have postponed development on that game so far because it will be another expensive, high-risk development project.
Instead, we’re working on a much smaller game, called Summer Daze at Hero-U, with a smaller development team and a much shorter development process. This gives us room to experiment and fail rapidly, rather than make the same mistakes over the next five years. Summer Daze will use a different scripting engine, a streamlined time mechanic, and 2D art. All of these things would have shortened the development time of Hero-U; but then it wouldn’t have lived up to the expectations for “a new Quest for Glory game.”
Read more about:
Featured BlogsYou May Also Like
