Re: GM-controlled Bad Luck points

Simon Smith wrote:
"gleichman" <fox1_217.nospam@xxxxxxxxxxx> wrote:
"Peter Knutsen" <peter@xxxxxxxxxxxxxxxx> wrote in message
If you have any specific criticism of the roll mechanic, I'd like to hear it.

I intensely dislike variable dice resolution systems be they simple X number of dice per X skill levels (4.0 Shadowrun) or different dice per skill level (Savage Worlds).

Whatever their virtues in managing probabilities (and I don't consider them significant), their impact on the player (and even GM) is nearly uniformly negative. They hide the real chance of success and the impact of individual modifiers behind highly unintuitive and in some cases exceedingly complex math. If the system works correctly (which they often don't, and I've learned to check this myself as designers are often very wrong about the outcome of their own work), the best in-play feel you normally get is a fuzzy 'more is better'- and that IME is an highly fustrating way to play. A reaction that has been uniform in anyone who fits in with my own desired gaming style.

Now here I'm with Peter. I too like variable-numbers-of-dice mechanics. I
like competent characters to be consistently competent. 'Consistently

Yup, that's what I want. Consistently competent. And that's not something you can get in any kind of 3d6-based mechanic, because at the absolute maximum you can't go higher than a 99.5% chance of success.

(Unless you go to the extreme of giving a 100% chance of success, which is no more realistic.)

competent' is a bit of a nebulous phrase, but among other things it's going
to include fumble chances for routine tasks on the order of 1/1000 or less,

That's important to me, but even more so, I want the Fumble probability to *keep* going down as effective skill goes up. Very few roll mechanics can do this. It is very easy, with most roll mechanics, to reach a level of effective skill where higher skill does nothing to reduce the Fumble probability. You've then hit the Fumble plateau.

For instance in GURPS, no matter how skilled you are, you'll always Fumble on a roll of 18. That's a 0.5% chance of Fumbling, no matter how highly skilled you are. Sherlock Holmes will Fumble 1 out of 200 Deduction skill rolls. Doc Savage will Fumble too, in spite of being extremely skilled at a wide range of thinghs. And he'll Fumble frequently. They will both Fumble *much* more frequently than what is realisic for those characters.

Another advantage of a multiple-die roll mechanic is that opposed rolls work much better.

Unless you want to go into explicit, in-the-book success ranges (which must then be memorized by the GM and all the payers), where if you make your roll by 1-3 points you get one grades of success, but if you make your roll by 4-6 you get another grade of succes, and if you..., all you have is a binary "I made my roll, now you roll to see if you made your roll", which fails to let the very skilled character triumph not just consistently but also *quickly* over the character who has little skill.

With a multiple-dice roll mechanic, if my character is highly skilled and I roll a lot of successes, and I know my character's opponent to be moderately skilled or worse, then I can already be confident of having won, whereas in the typical "roll under"-mechanic, he may very well have a 30% or 40% or even 60% chance of also making his roll, which means that neither wins, even though my character is much more skilled than the other one.

which gibe reasonably well with real-world failure/fumble chances for
routine tasks. To generate a 1/1000 chance using dice, you need at least
3D10 or 4D6. So you can do it with a D1000. But if you do it with a D1000,
you have a skill roll range spanning the 1000 points from 0001-1000, which
is overkill in my view. 4D6 by comparison, generates a 1/1000 fluke at
either end of the probability curve, but only spans the 21 points from 4-24,
which means a one point difference between two skill rolls is large enough
to be a significant difference. Using multiple dice may have downsides in
computational complexity, but I find their variance characteristics, and the
small points span covering a wide probability range to be very attractive.
And I've got computer programs that can calculate the odds for me, so I for
one can check that my probabilities are behaving the way I want them to even

I have checked my probabilities. I even checked the Sagatafl probabilities two different ways. First I used a QBASIC program to make hundreds of thousands of rolls for each combination of number-of-dice and Roll Difficulty, and collected all the percentages in a TXT file. Later on I used mathematical formulae, provided by people on the RPG-Create list, and compared those probabilities with the programme-generated ones (obviously the match was very, very close).

The MA RPG roll mechanic has only been examined with a programme, but as it is a lot simpler to analyze, I have been able to manually verify the results against my knowledge of roll probabilities.

To repeat: The probabilities of both roll mechanics *always* behave as they should. There are *no* abnormalities.

if I use wierd dice. Although frankly multiple D6s is quite complex enough,
and even then it's really not all /that/ complicated.

As an aside, I do reckon most game designers try to use different die
mechanics just to differentiate their games from everyone else's. And the
majority of them do *** up the probabilities. But the different die

Yes, that's why I always point out that the probabilities of *my* roll mechanics are not broken.

mechanics of different games do undeniably affect the tone of those games. I
/like/ the tone engendered by the Star Wars D6 system. I also like Feng

The Star Wars d6 roll mechanic is bad because the rolls are slow to read. One design goal of both the Sagatafl and the MA RPG roll mechanic was to make the roll very quick to execute and "read". Not only don't you have to add up dice, ever, nor do you need to memorize any kinds of range values ("made my roll by X-Y, made my roll by Z-W..."). There are also no re-rolls. You make one roll, and that tells you the outcome.

Shui's radically different die mechanic. EarthDawn wouldn't be EarthDawn to
me without its frankly rather screwy resolution system. Same for DC Heroes
and Golden Heroes. The RPG universe would be a far poorer place if
everything used the same dice mechanic, even if it was provably the Perfect
Dice Mechanic for All Possible RPGs. I say keep the oddball dice mechanics,
warts and all.

I disagree. Being different just to be different is a big mistake.

All I want is a good roll mechanic. Something with decent probability behaviour. Something that lets a good character actually *be* good.

I don't see much of a problem. The BL points of all incidents are player-knowable, so if a player is concerned, he can keep track of the GM's BL point expenditure. If he notices something amiss, such as the GM not spending his full quota on one particular PC, the track-keeping player can talk to the GM about it, and if the problem then persists the player can confront the GM with an accusation of cheating.

I'm sorry, but all this (including the previous unquoted parts) doesn't really alter my reaction. The mechanic's design impacts me as "GM gets to screw the players" at whim, limited only by minor resource management. It's not something I'd accept.

That may be in part due to my own gaming style. There the ability of anyone to alter a single event could alter the final outocome of the entire battle. There instead of 'bad luck' being favor, it can be decisive. The GM could in fact decide the make the players lose, and that is something I build my games not to do.

I am less unhappy about the bad luck points given that I now know their use
is more precisely quantified than I had originally thought. Doesn't mean I

What made you assume that in a 3-player campaign, the GM gets a single pool of 18 Bad Luck points that he can distribute among the PCs as he sees fit?

How does that assumption make sense, given the known existence of point-purchaseable character-specific traits such as "Less Bad Luck" and "No Bad Luck", and even "More Bad Luck"?

The only way to combine those two is to assume that the effect of a trait such as "Less Bad Luck" simply is that the character contributes less Bad Luck to the party.


Another possible misunderstanding is that 6 Bad Luck points per PC pr week equals 6 incidents per week pr PC. It doesn't.

It is pretty clear from my original post that an incident typically costs something like 3 Bad Luck points, which makes for something like 2 incidents per PC per week. Given that a lot of adventuring can take place in a single week of game world time, that's a small amount of Bad Luck spread very thinly.

Assuming a standard party of 3 PCs, it'll be slightly less than one incident per day.

(I've decided to dump the extra 6 Bad Luck points per month. If unspent Bad Luck points carry over, there's no need to have this complication.)

like them, merely that I dislike them marginally less. One question I ask
myself is, would I use them for NPCs? Answer is no, because it's at best a

They're intended to be given to major NPCs. Minor NPCs are rather like "mooks" in Feng Shui, except without its broken or unrealistic mechanics.

PCs (and major NPCs) can also have the Jinx Foe ability, which lets them inflict extra Bad Luck on their enemies. This will be using the same BL point costs as the regular Bad Luck mechanic.

Major NPCs also won't get all their Bad Luck points in engagements with the PCs. Major NPCs are assumed to lead full lives of their own, in which they engage in adventuring-worthy conflicts, which helps them burn away some of their Bad Luck. The guideline to GMs would probably be that half of a major NPC's Bad Luck should be available for confrontrations between the major NPC and the PCs, with the other half of his or her Bad Luck having been spent in situations when the PCs aren't around.

waste of time (if you use them for something harmless), at worst a rather
blatant fudge in the PCs favour, if they are used for something critical.
But if you don't use the bad luck mechanic for NPCs, you either have to use
some other mechanism to represent it (which would probably be less precise
and could still vary from 'harmless' to 'critical' in a potentially
unpredictable manner), or you are treating NPCs differently from PCs,
because NPCs never get random bad luck like the PCs do. That will really bug

Bad Luck points aren't random.

some types of player, and while I personally don't mind PCs and NPCs being
treated /differently/, different treatment is only acceptable if it's as far
as possible equally /fair/ to both sides. The mechanic as written is still a
game-breaker for me. I also think it's ridiculously complicated for what it
is attempting to achieve. You dole out bad luck points, then have to keep
track of how many you've spent on each character, try to spend them fairly,
replenish them according to the phases of the moon, blah blah.
>
Why not just chuck the whole system and have a Bad Luck Dice that is rolled
by the GM, and then the next character to roll the same as the Bad Luck Dice
Total suffers a piece of bad luck? You can tweak the probabilities as

Competely random, and it also fails to take into account the fact that Bad Luck is unevenly distributed in the world's population.

desired. For example, if you wanted 10% of die rolls to exhibit bad luck,
roll a D10; and the next die roll total whose last digit matches whatever

There are no roll totals in Modern Action RPG.

was rolled on the D10 gets it. If you have a character with the Unlucky
Trait, roll two Bad Luck Dice, and he gets it whenever he matches either die
roll. This suggestion or something like it saves a boatload of tedious
bookkeeping and is inherently fair, because the trigger for the bad luck is
two random numbers happening to match. That should hit all players and NPCs
roughly equally, and at a frequency rate of your choosing. Once every 50
rolls, say. There are even fair ways to trigger a bad luck event when
characters take actions that don't require skill rolls. And you could say,
set a timer, and after the bad luck has been used up, the next bad luck roll
is made after a secretly determined amount of time, all ready to hit the
next sucker.

It still isn't clear to me how you can have Bad Luck incidents for activities that don't require dice rolls.

Also, you're failing to simulate the fact that different firearms have different jamming probabilties.


More generally, the Bad Luck mechanic serves a number of purposes.

1. I already have the "Jinx Foe" Luck trait, which gives characters that have it some Jinx points that their players can use to inflict Bad Luck on their enemies. Also introducing Bad Luck points is just an extension of this mechanic (and a reason for me to quantify what Jinx Foe points can do).

2. I want players to be able to select (to fill in their Weakness & Flaws quota) some kind of Bad Luck trait, to reflect that this character has more Bad Luck than others, or (as seen in some other RPGs, including GURPS and Hero System) that he has Bad Luck where others don't.

3. I want *** to happen throughout the campaign, in a very orderly and distributed fashion, and in very small amounts (roughly 2 incidents per PC per week), including in activities that don't require dice rolls.

4. I want firearms to be able to jam. The reasons for this have been expained in my reply to Brian.


Now, #1 isn't much of a reason for also having Bad Luck points, perhaps.

I could just have mechanics for how Jinx points can be used (what they can do, and how many Jinx points each effect costs). One argument for having Bad Luck points as well is that if there were only Jinx points, then *** would only happen when there's a Jinxer around, and that is something that perceptive and intelligent characters are capable of noticing and subsequently thinking about. With the current rules, having a Jinxer around just means that *more* *** happens than if the Jinxer isn't around, thus making the "Jinx Foe" Luck trait much less blatant.

Reason #2 seems to suggest that it is okay to only give Bad Luck specifically to those PCs (and major NPCs) who have the "Bad Luck" Special Flaw, so that everybody else gets zero weekly points of Bad Luck. The problem with this is that Bad Luck then becomes a special case mechanic, with a high risk that the GM will forget to use it (up until the point where he remembers it, quite possibly several weeks into the campaign, or even a couple of months, which means he'll have a huge bunch of of stored-up Bad Luck points that he'll need to un-load on the PC (most likely there'll just be one) who has the "Bad Luck" Special Flaw.

Furthermore, if Bad Luck points are only given to those characters who have the "Bad Luck" Special Flaw, desire #3 (to have *** be able to happen to everybody) and desire #4 (to specifically have firearm jams be able to happen to everybody, while simulating the fact that some firearms are more likely to jam than others) are not met.

Reason #3 is easy to misunderstand, in the sense that one gets the wrong impression about how much *** I want to see happening. Because it isn't a lot. Just a little. A couple of incidents per week per PC is fine. It'll make for a richer world. Also, some GMs have a vague sense that "*** happening" is fun, but without mechanical guidance they are at great risk of not distributing it properly, in the sense that every PC gets roughly the same amount, seen over time, and in the sense that it is integrated with the mechanics (so that a brilliant chef doesn't cook up something that gives the entire party food poisoning, for instance) rather than being something that is outside the mechanics, or overrides the mechanics.

Reason #4, again, requires *** to be able to happen to every major character. If firearm jams can happen only to those characters who specifically have some kind of "Bad Luck" trait, then everybody else can just run cheerfully around with badly maintained WW2 submachine guns, or WW1 sniper rifles. Any cop will, as soon as he or she becomes clued-in to the physics of the game world, ditch the revolver in favour of a high-ammo-capacity pistol.

--
Peter Knutsen
sagatafl.org
.