[New Game, New Designer] My first steps into the pool

[Harlequin]

Yeah.  In some ways it's easier to detect a feedback loop in playtest, then come up with a brilliant way to fix it, than it is to avoid them entirely in the original design.  Sometimes this sort of thing - the necessity of repair - can help spur better inventions than what they supersede.  So build a system, keeping these things in the corner of your mind, and set up a one-shot.

- Eric

[Thespian]

Try out one way that appeals to you, and playtest :)

Just so.  I mentioned to watch out for this without using the term positive feedback in my original post, so I totally agree you want to be aware of this.  But xenopulse hits the nail on the head... you won't know what the "right" way to deal with it is until you try a few.  However, you have many options.  I'll offer a few off the top of my head:

"Failure *is* an option" - If you want your players to feel they have something to lose, leave it as is.  Your despair gets too big and you can't go on.  End of story.  Or, conversely, you are so good that everything becomes easy.  This is a "success breeds success and failure breeds failure" model.  This might work for games with a definite end.

If you simply want something to dampen the feedback loop, you could do something like this:  task/conflict target number = difficulty (picked by GM) + players doubt/fear (which ever is appropriate) score.  Player rolls dice + appropriate skill + confidence/courage (which ever is appropriate) to beat target number.  If roll + bonus is greater than target *and* *unodified die roll* is also greater than current confidence/courage (which ever was used), then increase confidence/courage by one.  Likewise, if roll + bonus fails *and* *unodified die roll* is also greater than current doubt/fear (which ever was used), then increase doubt/fear by one.

Essentially, you get additional Hope/Despair based on succeeding or failing a roll only if the *unmodified* die roll exceedes the mood stat in question.  That way, things still get way easier and harder based on extreme Hope/Despair values (a feature) but it gets harder and harder to  keep pushing the extreme, and easier and easier to start pulling it back to the center if your task resolution goes the other way.

(numbers completely made up to illustrate point, scale appriporiately for your system)

Example: I'm on top of the world.  I'm confidence:90 Courage:90 and doubt:10 Fear:10.  I have skill DoStuff: 50

I'm going to be 90 (Hope stat) + 50 (skill) + die roll in DoStuff tasks.  The tasks are going to be GMset difficulty + 10 (Despair stat).  Lets look at tasks the GM sets difficulties of 150, 200, and 250 in cases where I roll a 5, 15, 50, 85, and 95

Code:

                       My roll
task |  5   | 15   |50    | 85   | 95
150  |  f   | F    | s    | s    |  S
200  |  f   | F    | F    | s    |  S
250  |  f   | F    | F    | F    |  F

"f" and "s" means the roll failed or succeeded respectively, but no change in Hope/Despair levels.  "F" and "S" means the roll failed or succeeded respectively, but levels changed.

Note that in these cases, a GM set dificulty of 150 is impossibile for an average person with balanced Hope/Despair levels of 50/50 with anything less than a skill of 50.  150, 200, and 250 are ridiculously high targets.
 
An interesting aspect to this mechanic is, that while low rolls fail, and high rolls succede, really low *unmodified rolls* don't increase your Despair, and high *unmodified rolls* can, if you fail in doing something really difficult.  This seems reasonable to me.

Hope that helps.

Thespian

[Remko]

Another thing you could do is create a playtest document and ask some folks over here (either by putting in a link in your sig, or by posting it here) to playtest it. I know there are enough people willing to playtest.