[Legendary Lives] Three games to talk about
Ron Edwards:
Hi Christoph,
I'm not sure how much of the quick summary I'm about to give is known to you, or standard knowledge for everyone reading. I apologize in advance for any inadvertent patronizing. I'm also limiting the discussion to very basic, long-standing methods of rolling dice in role-playing games. So the first thing is setting up a simple contrast.
- Flat. The chance of getting any single value is the same for every value. The most common example is rolling a single die for its facing value. Using two d10s for a d100 role is flat too.
- Curved. The chance of getting a value in or near the middle of the range is higher than getting a value at either end. The most common example is rolling a fixed number of dice (more than one) and adding up the pips for a total.
(I am leaving out all other ways to read dice, especially those for which many dice are rolled and successes are rated per die and then accumulated, as in White Wolf's mechanics or Fudge dice, and also those which use highest-value, like Sorcerer and many games since its publication, and single-die rolls subject to re-rolls as in Dying Earth. They are relevant to this discussion but have various curves of their own, especially when the target numbers vs. opposed rolls issue is brought into the mix.)
I'm not saying one is better than the other, but their distinct properties are consequential. It's not surprising that both were retained through what I think of the three "tracks" of RPG design throughout the 1980s: flat in the form of d20 for various permutations of D&D, and also in the form of d100 for BRP games; curved in the form of Tunnels & Trolls handful o'dice and The Fantasy Trip's 3d6 (roll equal to or under a target), immortalized by Champions and GURPS.
Anyway, perhaps best illustrated by Rolemaster, the flat method includes a huge pitfall awaiting the uncritical game designer. In Rolemaster, the far ends of the spectrum (close to 01, close to 00) led to rolling on further tables, the critical hits and critical misses ones. And those as well featured a variety of mild-to-extreme results, with the latter at the "ends" of the numerical spectrum. The problem is that when rolling "flat," getting a number at the far end (03 out of 100, say) is just as probable as any other. If we're talking about many rolls (to hit with a sword, say), then yeah - having 75% skill is better than having 35%. But if we're talking about a single roll, not only the chance for success but especially the chance to jump to the critical tables based on rolling a particular value, then the difference in skill is totally irrelevant - especially in terms of whether you're going to the gruesome-miss table or the awesome-success table.
Early BRP had similar problems, but it's more complicated because (i) it was really a d20 roll rather than d100 because the units were always in 5% increments and (ii) the damage was a separate roll. But this is why early RuneQuest was quickly nicknamed "LimbQuest" because so many characters lost limbs in the first combat of play, and why Rolemaster's critical hit tables are so notorious. Especially since those games featured customized, rather laborious character creation, it was pretty grim to see one's guy mutilated in ways which - although not necessarily understood in full by players - were obviously orthogonal to whatever the high numbers on one's sheet were supposed to indicate about the character. The late-80s game Justifiers was insanely prone to this problem.*
(Interesting aside: this is exactly why in the brands of D&D that I played back in 1978-1982, one's roll to hit was not particularly valued as a score to want to improve - it was the damage one tried to boost and make most reliable, favoring multiple-dice damage rolls for consistency and lots of pluses to avoid the dreaded hit-for-one-point anticlimax.)
(Interesting aside II: Unknown Armies uses percentile resolution. Its "criticals" are found in doubled digits: 11, 22, 33, and so on. Since the roll is flat, the difference between this mechanic and putting the criticals at either end is negligible in terms of probabilities, but it's easier in terms of handling time. More on UA in a minute.)
OK, so for percentile or any flat-based rolling to work, here's what comes to my mind to avoid the pitfalls.
1. Since the differences in percent between two characters with the same skill/ability will be manifest only over many rolls, not a single-roll-each comparison, if you want that difference to mean something, then the consequences of any single roll should be relatively minor, allowing for lots and lots of instances of rolling without getting, for instance, decapitated on the second one. Designing mechanically-minor but narratively-forward-moving consequences becomes the primary design challenge, and if that's too hard (as I think is indicated historically), then points 3 and 5 below should be maximized.
1'. Which is also related to what die or method you'll use in the first place, specifically how many units. The rule of thumb is that you'll need at least as many instances as (for dice) sides of the die in order to discern the probabilistic differences. So this is why d100 gets such a bad rep - you need more than a hundred sword-swings to see the higher chance for success actually manifest itself between two characters. (If this seems wrong to you based on your experiences, it's likely that you perceive the higher-skilled character's success at any point as validating the higher score, and the converse with the failures of the lesser-skilled character, cognitively diminishing the contrary results as "less important.")
2. The characters' differences in ability should be expressed in terms of how many units of the spectrum are different from character to character, and this should be more nuanced than merely success vs. failure, or also, in the case of BRP, more consequential than the difference between 5% vs. 4% for a fumble. For instance, in Legendary Lives, each gradation of one's score (1 to 20) is not itself a probability; it's a line of the resolution table which expands and contracts all the various ranges of the standardized qualitative results within the 1-100 range. The link in my first post takes you to the game, so you can check this out yourself and I won't try to describe it fully here.
3. The roll should be modifiable through immediate play, whether just before or just after, to affect the chances. Legendary Lives does this with column shifts, due to played circumstances and also set-up augmenting rolls, and Unknown Armies does it too if I recall correctly (flip-flopping, but when/how you get to flip-flop, I can't remember). Any number of bonus-y techniques and associated fictional content are available for this purpose; my point is that all rolls of (for instance) to-hit are not narratively equal, so sometimes, saying, "Well, I missed this time, but in the long run, my 75% will show up as a high level of skill," is not enough. When this roll is more important, different logic applies in terms of personal-enjoyment. A lot of mid-80s play utilized a "burn experience for a roll bonus" house-rule for just this reason. Unknown Armies has its flip-flopping, for instance.
4. Deeply reconsider utilizing opposed rolls, and if you do, then have a really operational, tested-for-fun reason. Legendary Lives was one of the first role-playing games for which the GM never rolls (you know, I keep thinking I remember one before the early 1990s, but then I forget again - can anyone help with that? Or am I remembering only almost-RPGs, like Fighting Fantasy?).
5. Failure needs to be interesting. In both Legendary Lives and Halmabrea (which also uses d100), the instructions for the GM are helpful regarding (i) a failed roll may not mean actual failure of the intended action, but some delay or problematic elements along with the success; and (ii) catastrophic failures should be very distinct plot-moving moments, not merely graphic depictions of the character's ineptitude. Legendary Lives also has the skills improve with either maximal success or maximal failure, so that the latter feeds into the reward mechanics just as much as the former. A little more subtly, the ordering of successes and failures can also impact the character, as with the direction your character slides on the Madness Meters in Unknown Armies (am I remembering that right? something like that, anyway).
Anyway, that's what I came up with for now. Christoph, is that helpful or interesting? Too basic or obvious maybe?
Best, Ron
* I bring this up to point out (i) I have in fact played Justifiers by the book, for which I expect some relatives were released early from Purgatory; and (ii) I GMed a very satisfying Justifiers game at Forge Midwest, using Levi Kornelsen's The Exchange. That'll be a thread soon.
Lance D. Allen:
Ron,
Your last post is deeply important to me, as Mage Blade uses a single d20 roll -vs- target number (roll under, though I doubt that matters much) but I'm having a very hard time grokking everything you just said.
Most specifically, what I don't get is how on a single roll a difference between 35% and 75% is irrelevant. If you need to roll between 1 and 75 as opposed to 1 and 35, it seems like this difference, even on a single roll, is pretty important. I get that it is statistically identical to roll a 99 as it is to roll a 1, but it's not the specific number that matter so much as which side of a dividing line it's on. There are 75 possible chances to succeed in the former case, and only 35 possible chances in the latter. Please explain to me what I'm not understanding?
Also, I've heard your opposition to opposed rolls before, but I've never understood it, either. Mage Blade also makes extensive use of this.
If it's an important datapoint, in most, if not all, instances, the difference between the target number and the roll is important. This is true even in the case of opposed rolls.
As this isn't talking about Legendary Lives precisely, I'm perfectly fine with taking this to another thread if you feel that's appropriate. I have been expired to gank me a copy of Legendary Lives to see what about the lifepath system is so spectacular, though.
Ron Edwards:
Hi Lance,
This is a fine topic to stay within this thread, no big deal.
The difference between 35% and 75% using a flat technique is indeed enormous ... given a lot of instances to compare. The trouble is that it's practically impossible for the mind to not think in terms of lots of instances; we're hard-wired to do it even when thinking about "this one time." Even saying "the chance" implies a whole bunch of instances.
If we're talking about how well a character swings his or her sword, rated in percentage terms, then I dunno what to tell you, except that the human mind snaps so hard into thinking about "over time," "in general," "overall," and basically "a single instance in the context of many," that we practically don't even have the language to talk about this one time as an isolated event.
All I can do is talk about the numbers. Craps is the perfect game for this point - you can't play Craps with just one die. It relies heavily on the minor but definite bell curve of possible results on a single roll. You can get a 7 by rolling 6+1, 5+2, 4+3, 3+4, 2+5, and 1+6. By contrast, you can only get a 2 by rolling 1+1. So within that single roll, the chance for a 7 is higher. That's what makes a curved method. It matters to that one roll. That's why you can bet on Craps as it stands, but no one, even in the throes of the various gamblers' fallacies, would bet on Craps played with a hypothetical 11-sided die numbered from 2 to 12.
I understand what you're saying about success being 1-75 rather than 75 all by itself. I'm saying that when you say, "the chance," or refer to more than one number possibly being rolled, i.e. the range, you're automatically talking about "over many instances." So everything you're saying is right - over many instances.
I get the impression you think I'm saying flat methods are bad. They're absolutely not. They simply have characteristic pitfalls if they're designed without understanding these points, at least intuitively. I even provided a number of ways in which historical RPG design has made flat methods work well - Legendary Lives is a good example of thoughtful design, and there are others. I should point out as well that in Call of Cthulhu, the most important roll, the Sanity Check, is a flat method like all other resolution rolls in that game ... and it works well because no single roll smashes the character down from max to nothing. Losing Sanity and not losing it, in CofC, is a matter of repeated rolls - which matches to my point #1 for making such systems work.
If you design Mage Blade with a good flat-method resolution rolling method, then that's excellent. I look forward to it. (I also agree with you that over vs. under is not an issue.)
I agree that using the roll for further information - degree of effect, not merely success/failure - adds benefit and makes immediate use of the existing range. That would be another whole variable to talk about too, 'cause people have done it in so many ways (with or without a second roll, for example, and if with, whether the first roll influences the second). I restricted my points in the above post almost entirely to raw success vs. failure, and the associated detail of extreme versions of either. I did not talk about degrees of effect at all.
And simply to close the door on further confusions, I could go into further detail about various hassles with the curved resolution methods too, with examples of crappy and good applications. Choosing a curved method does not automatically make the rolling mechanic good. I'm talking about flat methods in this thread because that's what Legendary Lives uses.
And since you brought it up, I don't know where you get the idea that I disapprove of opposed rolls. Sorcerer relies only on opposed rolls, as does It Was a Mutual Decision and (somewhat more slowly) S/Lay w/Me. Trollbabe and Elfs don't. I've written about how combining target numbers and opposed rolls has led to needless woes in many game designs, but that's a nuance, not a criticism of opposed rolls in and of themselves. And I've even seen some versions of that which have ended up working, as long as they have some supportive mechanisms along the lines of my #1-4 points, Hero Wars, for instance.
My advice was not to avoid using opposed rolls. It was to use them with concrete and coherent reasons for why in this particular game, they (and their associated mechanics) are fun.
Best, Ron
Callan S.:
I don't think there's anything wrong with a single roll that smashes someone down from max to nothing - you just have to decide as designer if you want a game that could very well end in two minutes flat (or even less).
One would need to look at maintaining regular social dynamics though - you can't have one guy basically lose the game alone, then sit there, locked to his seat by social manners after being killed in the second minute, while everyone else plays for hours. One way out is if someone gets their head lopped off or goes entirely insane in one roll, it's game over for everyone. Again, if a two minute game session is okay for the design, then this is okay.
Though that's an interesting thing to note - most traditional game sessions end when everyone decides the session ends, rather than a mechanic saying 'this session is over'. There's alot of 'play as long as we want to' attitude out there with roleplayers, rather than 'play until the game says it stops' (though really they are the same - you can just start up another game when one session of it ends, so it's kinda a pointless attitude).
Also come to think of it, alot of boardgames and cardgames have people sit out - but it's usually not for a disproportionate time - people might drop out three quarters of a game in in something like the card game 'lunch money', not in the first minute. Also these games usually don't go for hours (though monopoly does).
Ron Edwards:
Moderating: Callan, that post reads like free-association and doesn't contribute to the discussion.
This thread is about the Legendary Lives games I've played and various issues that crop up from that. I'd like to stick with them.
Best, Ron
Navigation
[0] Message Index
[#] Next page
[*] Previous page