Skill Rolls and how they work

[swang]

While analyzing my latest run through, I have come to the realization that our basic assumption, that skill rolls = random(1-dominant trait)+ skill just isn't correct.

 

First of all, I think a 0 result on the random is possible.

Second, the dominant trait isn't always used (it is used probably 99% of the time, but there were a couple of rolls that I know were not (via inspection by mod tools)

Third, we need to be able to account for "bonus" and "chance of success" buffs.

 

My guess is that it's a several step process.

 

1) both the skill value and any applicable bonuses are subtracted from the threshold (target number). If this number is <= 0, stop, and the roll auto succeeds (and shows as green.)

2) the new number (which is the actual target), is now converted as a percentage of 2*the attribute being used.

3) any applicable chance of success changes are applied to the target percent. (this final value is used to determine blue, black, red or purple.

4) a random percentage is generated and compared to the target percent, if it is greater than the target percent, the roll succeeds.

 

For example, a skill of 4, an attribute of 5, vs a threshold of 10, with a general CoS% bonus of 5%, and a bonus to social actions (which this example is) of 1.

 

1) 10-5-1 = 6

2) 6 = 60% of 2*5

3) 60% - 5% = 55%

4) if the randomly generated % is > 55, then the roll succeeds.

 

Several things led to my revised formula.

1) I've never seen a roll succeed when the threshold is exactly skill+2*attribute (I've done this on a roll when the attribute is 1, and reloaded at least 10 times.)

2) there are bonuses both as a raw # and a percent. The old formula didn't deal with the percent bonuses very well

3) Glory, as a modifier, just didn't seem to make too much sense. this way, glory can be easily done by a -glory% in step 3 (or maybe glory/2 after the reduction of glory effectiveness.)

 

Anyways, if anyone has more ideas, I'd welcome discussion.

[Exfides]

Anyways, if anyone has more ideas, I'd welcome discussion.

 

Well since you asked...

 

I think the first thing you should keep in mind while you research skill roles is that during Adventures the roll uses the best skill level in your clique (official confirmation here). That brings up an interesting question, what happens if a clique member has much better skill but your character has a far higher attribute for a given role? This is an interesting wrinkle because there are some buffing spells, like Cleans and Remake can send your attributes skyward with just a few castings. While ramifications such as these don't directly effect the skill roll equation per say, they can complicate efforts to "reverse engineer" it.

 

Your new outlook on the is interesting and there might be a lot of merit in it, but I have some basic issues with what prompted you to make this revision.

1) I've never seen a roll succeed when the threshold is exactly skill+2*attribute (I've done this on a roll when the attribute is 1, and reloaded at least 10 times.)

When dealing with a pseudo-random number generator a test of only ten trials is of doubtful use. Even using a hypothetical true random number generator you can still get streaks that would cause this behavior while still using the old formula.

2) there are bonuses both as a raw # and a percent. The old formula didn't deal with the percent bonuses very well

The only change you really need to make to the old formula is to include "+<%bonus>*<difficulty number>"

3) Glory, as a modifier, just didn't seem to make too much sense. this way, glory can be easily done by a -glory% in step 3 (or maybe glory/2 after the reduction of glory effectiveness.)

I freely admit that I don't know much about how Glory factors into skill rolls, but I also think anyone player-base has a full understanding of the mysteries of Glory. I just bring this up because while Glory could be, and probably is handled differently, than the old formula. However this alone doesn't mean the formula should be greatly revised rather than just tweaked.

 

Please understand, I'm not just trying to shoot you down. I am just pointing out that since the old player-derived formula seems to work accurately in the vast majority of cases, I would look at small adjustments rather than a complete overhaul. However, if nothing else your theories are worth serious consideration.

[Reianor]

While analyzing my latest run through, I have come to the realization that our basic assumption, that skill rolls = random(1-dominant trait)+ skill just isn't correct.

 

First of all, I think a 0 result on the random is possible.

Second, the dominant trait isn't always used (it is used probably 99% of the time, but there were a couple of rolls that I know were not (via inspection by mod tools)

Third, we need to be able to account for "bonus" and "chance of success" buffs.

 

My guess is that it's a several step process.

 

1) both the skill value and any applicable bonuses are subtracted from the threshold (target number). If this number is <= 0, stop, and the roll auto succeeds (and shows as green.)

2) the new number (which is the actual target), is now converted as a percentage of 2*the attribute being used.

3) any applicable chance of success changes are applied to the target percent. (this final value is used to determine blue, black, red or purple.

4) a random percentage is generated and compared to the target percent, if it is greater than the target percent, the roll succeeds.

 

For example, a skill of 4, an attribute of 5, vs a threshold of 10, with a general CoS% bonus of 5%, and a bonus to social actions (which this example is) of 1.

 

1) 10-5-1 = 6

2) 6 = 60% of 2*5

3) 60% - 5% = 55%

4) if the randomly generated % is > 55, then the roll succeeds.

 

Several things led to my revised formula.

1) I've never seen a roll succeed when the threshold is exactly skill+2*attribute (I've done this on a roll when the attribute is 1, and reloaded at least 10 times.)

2) there are bonuses both as a raw # and a percent. The old formula didn't deal with the percent bonuses very well

3) Glory, as a modifier, just didn't seem to make too much sense. this way, glory can be easily done by a -glory% in step 3 (or maybe glory/2 after the reduction of glory effectiveness.)

 

Anyways, if anyone has more ideas, I'd welcome discussion.

 

I hope none minds if I use some necromancy on this thread. From a few bits and pieces I've read here and there I've come to pretty much the same conclusion, only to have it negated by the Sphinx.

 

The sphinx mentions a roll of Insight/wit against a DC of 13.

I had my wit at 10. I was trying to find a reliable way to pass that roll. Among my spells was Asking the Universe which promises a 3 points bonus on all my actions and abilities. Matching wits with the sphinx is an ability.

 

Well, 10 Wits +3 bonus vs DC 13 was NOT an automatic success. Neither in colour nor in practice.

 

Either one of the mentioned abilities does not work as described, or the formula doesn't work as we suspect.

 

My insight was either 2 or 3 at the time of this occurrence. We'll assume it was 2, just to be safe.

I Don't have any chance of failure modifiers, but I do have a +1% chance of success.

 

I also suspect that the stress bonus is limited by the rolling attribute (or maybe even by dominating attribute).

Can't bring the numbers on this one, but I remember burning 8 stress points on a particularly difficult action and have it fail on me (about 3 consecutive reloads+retries, then I stopped retrying), even though the skill + stress was equal to DC.