Design approachs.

[Guest 85555]

If you make one skill work mechanically differently from the other 12 in the game, because of some vague notion of "realism" trumping your consistency, you may or may not have made a good decision. If you did, it was luck. On the other hand, if you "denormalize" that skill from consistency because you realize that it being different is going to add certain benefits to the play experience, and that these benefits are surely worth that bit of inconsistency, and worth more than other such inconsistency that you could have introduced--then you've got a good decision, and it has nothing to do with luck.

I agree with you that playability is almost always important; I do think realism can also be very important if it is one of your design goals. One thing I've found is the more consistent a system is, the fewer design options I have for each part of the game. And this narrowing can make it harder to achieve the specific results you want (which is often very important if realism is a key consideration). With our game, we decided a light weight consistent game would still yield believable results because people weren't thinking of the system. But if we wanted to make a realistic game for fans of rules heavy or rules medium games that got down to the granular details of things like fire guns or operating heavy machinery, I think dropping the consistency would be really important.

But to your point, we did exactly what you said in your example. One of our crime skills works differently than all the others, it can be used actively to kill people (whereas the remaining crime skills are more like knowledges). The mechanic for it largely remained the same,but it had different uses and the Target NUmber for the roll was set differently. In this case, the reason wasn't realism, but to better emulate the mob movie genre.

 

[Beginning of the End]

Linear design has some good points, the biggest one being easier to implement.

IME, this isn't true. It's far easier using stuff like the Three Clue Rule and node-based scenario design to design and run non-linear adventures than it is to keep a railroad on track.

 

[triqui]

But with railroading, you get sure what the PC will be doing, where, and when. With sandbox design (which in my opinion is the best way, but the hardest to achieve), the risk that players decide to do something unexpected is higher. For example, imagine D&D. In a linear, Dungeon Crawling-style adventure, you can predict easily what the players will do. Even in what order. You can make sure they find the "oil of protection against petrification" before they fight the Medusa, or whatever. On a sandbox design, where you have prepared the Big Bad Swamp, the Evil Tower, The Fierce Jungle and the Dark Sewers in the Thief City, the players could say "let's see what's beyond the river".

Depending of the nature of the game, this could happen less. For example, in your game, character driven actions are naturally controlled by the nature of the game: they are counterterrorist during a terrorism threat. That mean they won't go to do fancy stuff, most of the time. Other games, however, might have a harder time. Probably worthwile anyways, but it's harder.

I found 5x5 method to be a nice compromise between railroading and free sandbox.

So, what are those Rule of Three Clues and node-based techinques about