I’ve always liked the idea of roll-under systems, especially percentile systems – there’s something about a d100, whether it’s two different-colored d10s or the legendary Zocchihedron , which, once set into motion, will not stop until it leaves the gaming table. Unfortunately, I’ve never been quite satisfied with existing roll-under systems. The virtue of a roll-under system (as many people have pointed out before me) is that it is usually very easy to tell what a given character’s chances of success are: if her skill in Basketweaving is 57%, she’ll have a 57% chance of succeeding on an unmodified Basketweaving roll (don’t ask me what that would entail, but seeing some of the old Chaosium games skill lists, “Basketweaving” isn’t too far off the mark). If it’s a d10 based system, then a character with a skill of 8 in Driving would succeed 8 in 10 times on an unmodified Driving roll. 8 in 10 and 57% are easy to conceptualize. Heck, even 2 in 6 is pretty easy to conceptualize, if you want to go with d6’s. The main failing is how they handle opposed tests. Typically, at least in the old days, you had the “attack” roll, and then the “defense” roll. Only if the attacker succeeded and the defender failed did anything happen (as I recall, this was very much a “combat” focused thing). This led to whiff-fests a lot of the time: if the attacker had a low attack %, most of his attacks would be misses to begin with, and if the defender had a high defense %, then even successful attacks failed most of the time (being blocked). There were a few attempts to implement levels of success based upon rolling under 1/5th or 1/10th of your skill %, but this was generally awkward, and happened rarely enough that it usually didn’t help much (in my opinion).

Of course, I wrote a d100-based roll-under system (well over a year ago, I’m quite surprised I never did a column about it – maybe next month), but I never felt that it really did anything better than the Tarafore system does it (see Column 35: For What It’s Worth for my thoughts on redundant systems). Now, I have seen the beginnings of a pretty amazing d20-based roll-under system in this thread from the Design and Development forum. TRM has really come up with something good; namely, using thresholds to determine levels of success while allowing the GM to hide target numbers from players when necessary. Honestly, I wish I’d have thought of it myself.

That Said, I Did Think of Something Myself

Rather than using a thresholds-based system, I’m working on a blackjack-style system that uses margins of success. MOS require a little more math, but not an odious amount. For those of you who aren’t familiar with “blackjack” roll-under systems, it’s exactly what it sounds like; you’re trying to get the highest possible roll without going over a certain number (in this case, your Trait rating). Traits available to PC’s run from 1 to 9, with 5 being average (in the sense that average means “meh, not my specialty, but I don’t particularly stink at it,” not “this is the mean score that player characters will be expected to have in this trait”), 6 being Good, 7 being Very Good, 8 being Outstanding, and 9 being Legendary (yes, I am quite attached to the Average/Good/Very Good/Outstanding/Legendary scale. Thanks for noticing :). All checks use d10, which means there’s always a chance of failure.

Even though I envision this game as primarily focusing on conflicts, single (unopposed) checks will happen over the course of the game. These are handled simply by rolling a d10 and comparing it to the relevant trait. If the task is especially easy, a bonus to the trait may be given (+1 to +3, usually). If the task is especially difficult, a penalty may apply. It’s easy to conceptualize how likely you are to succeed at a given task: take the final Trait total (after modifiers) you’ll be rolling against, and that’s your chance in 10. If you prefer percentages, you can multiply it by 10% to get the percentage.

However, this isn’t a game of single checks, but of extended conflicts (which I mentioned in my third column). Progress is charted on a Conflict Track, which starts at 0, and goes from -10 (NPC or situation wins) to +10 (PC wins). Each “round,” the two sides roll their Trait (plus any modifications). Each turn, the winner’s Margin of Success pushes it toward his or her side. When it reaches the end (+10 or -10), the conflict is over. There are also thresholds along the way that have additional effects. Usually (and this will vary by situation), the disadvantaged character suffers a penalty of -1 to all Traits at 3, and 7, and cannot choose to break off the conflict (taking his lumps and giving up) if the threshold of 5 has been passed.

If a character’s roll is a success, then the number on the die stands as their Success Level. If the character’s roll is a failure, then their Success Level is Zero. The Margin of Success is the difference between the winner and loser’s Success Levels. The Margin of Success is added to the winner’s “Track,” taking things closer to a win.

For example, in combat, Barry Boxer, the PC is fighting Bud Bruiser, an NPC. Barry has a Fighting Trait of 8 and Bud has a Fighting Trait of 7. If Barry rolls a 6, and Bud rolls a 7, the both have succeeded, and Barry has a Success Level of 6, and Bud has a Success Level of 7. Bud’s Margin of Success is 1: his Success Level of 7 minus Barry’s Success Level of 6. This MOS of 1 is added to Bud’s side of the Conflict Track, pushing it to -1 (NPC = Negative. PC = Positive. Just a little alliteration).

The next turn Barry rolls a 5, but Bud rolls horribly, getting a 1. Both rolls were still successes, being equal to or less than their respective Fighting Traits. Barry would still have had a Success Level of 5, but Bud would only have a Success Level of 1. This would give Barry the Margin of Success of 4 (His SL of 5 – Bud’s SL of 1). This moves the Conflict Track by 4 toward Barry’s side (PC=Positive). Since it’s currently at -1 (Bud’s narrow victory from last turn), moving it four toward Barry puts it at +3. This activates the Threshold of 3, which will penalize Bud by 1 in all future rounds until the Threshold is un-passed.

The following round, Barry had rolled a 3, not so good. But Bud rolled a7. Barry succeeded, but Bud failed (His Fighting Trait was 7, but now it’s down to 6, due to the -1 penalty he picked up last turn). Barry’s Success Level is 3, but Bud’s is 0 (he failed, so his Success Level is Zero). That gives Barry a Margin of Success of 3. This moves the Conflict Track by +3, to +6, pushing past the Threshold of 5. This means that not only will Bud still be penalized by -1 next turn, but he won’t even be able to turn tail and run unless he either gets back below the +5 Threshold or Barry lets him go.

Final Thoughts

One thing I don’t like about this system is that conflicts lack any real element of surprise. It’s impossible for one side to win in one fell swoop. It may be helpful to make the threshold for defeat equal to a character’s Trait, instead of 10. That way, with an extremely lucky roll, a character with an equal or greater Trait could defeat an enemy in one round. In the example above, Barry has Fighting at 8, and Bud has Fighting at 7. This means the sliding scale goes from -8 (Barry’s Trait) at which point the NPC, Bud wins, to 7 (Bud’s Trait), at which point the PC, Barry, wins. Barry could, with a very lucky roll, get enough Margin of Success to take Bud out in one round. Bud, however, could not do the same to Barry.

I think the sliding scale may make the game more interesting, and place more of an emphasis on relative ability (since it affects both the target number to roll against and the threshold for victory). I’m also concerned with how group interactions, equipment, situational modifiers, and tactics will play into things. Those things, however, will be addressed in my next column.