At the end of the day Bedrock Games is about making RPGs that we ourselves want to play. However we are guided by a simple philosophy that design goals should be held in balance with one another.
When we first started out, Bill and I settled on three core principles to reflect this approach: playability, plausibility and personality. In my case, I like games that hit all the right buttons. So three principles that broadly encompass what I expect from a game make sense. Anything beyond this would honestly confuse me as a designer.
First let's take plausibility. All plausibility means is keeping things believable. It doesn't mean our games have to simulate reality at a granular level, just that the rules should support plausible consistency. What is plausible of course will vary from game to game (since a fantasy setting has different expectations than a military setting), something we've tried to pay attention to with each release. In general it means we strive to create mechanics that don't strain suspension of disbelief. It is also about evading blind spots or oversights in design.
A simple rule of thumb is to ask "why ?" with any mechanical consideration (something Bill excels at). It is very easy to design with a broad brush and forget about the rare cases where a mechanic breaks down. A good example of this occurred during the design of Servants of Gaius. We wanted a simple mechanic to govern the inequality of women in the setting. So I proposed a -1d10 penalty when any character interacts with someone of the opposite gender. Since men control most positions of power, I assumed it would wash out and that the penalty was an elegant way to handle the issue. I was wrong.
I didn't want to be wrong. At this point I was in love with the mechanic. It seemed so simple, so easy to remember and so all-encompassing. Bill pointed out the problem by asking a simple question "What about when a woman tries to seduce a man?" At first I resisted, but he was right. There was no denying that my simple and elegant solution would produce situations that would strain credulity. The mechanic was disruptive and had to go. This is an example of playability coming into conflict with plausibility.
But sometimes plausibility comes into conflict with the second principle: playability. This isn't just about game balance (though it includes that). Playability also includes the smoothness of the mechanics and the excitement they produce. It is about numbers and mechanics, but also about how people perceive and experience those things. At times what is the most plausible isn't the most playable. So we try to balance the two out. Another approach would be to go completely in the direction of playability or of plausibility (and some believe this is the only true option). We adhere to the belief that balance between the two is not only possible, it is desirable. Done well, you end up with pleasant mechanics that don't disrupt the believability of the setting.
Sometimes when you strive for realism, you create problems if you don't keep playability in mind. It doesn't mean every single aspect of the game needs to be balanced out perfectly. But it means we need to be aware of the impact of realism on the mechanics and take ownership of that. A good example of plausibility conflicting with playability is how we designed the poisoning rules for Servants of Gaius.
Initially there wasn't much containing the poisoning rules. You could brew poisons all day until you have a stockpile. The concern was there was nothing stopping someone from brewing too much poison. The best solution for a playability issue is a plausible one, so we went with the following: anytime you brew poison, on a failure there is a small chance of accidentally ingesting a dose yourself (this chance decreases with skill level).
The final principle is personality. To be honest we picked this label so all of our principles would start with the letter 'P'. This is essentially about flavor. In Crime Network we wanted the game to feel like a mob movie, so the text , the setting and the mechanics were designed with that in mind. When we made Horror Show, our focus was on bringing Horror Cinema to life, so we created mechanics to emulate horror movies.
Flavor can also interact with plausibility or playability in funny ways. The important thing with Flavor is making sure it is reflected in the rules. For instance, when we made Servants of Gaius we wanted the gods of Rome to be real. That meant they needed real powers and we needed mechanics for appealing to the gods. Otherwise there would have been a serious gap between the personality of the game and the playability. With plausibility things are even more critical. The flavor has to be internally consistent. We can't say on page 10 of the rule book that Caligula likes dogs but contradict this statement on page 50. The flavor content has to not only be consistent but it has to make sense too.
Flavor is important, but so are the other elements and this is what keeps our products firmly in the realm of traditional role playing games. There is nothing wrong with hyper-focused design. But we want the focus of our game to balance out with the other principles. For us this creates a more enjoyable experience for everyone at the table.