Friday, April 3, 2020

Stat Squish and the Lawful Roll

People have criticized the GLOG thus:

"For a game that says that stats don't matter very much, they seem to matter a lot.  Since so many things are decided by rolling under a stat, and stats vary so much, your starting stats matter a lot.  This reduces the impact of good gameplay."

FIRST OF ALL: a lot of that comes down to how often your DM asks for stat checks, and how often the players work to bypass situations that require stat checks.  (This is all your fault, not mine.)

Secondly: Eh, that's a fair criticism.

Let's Talk Shit About Stats

This is not something that I am unaware of.  After all, this is why I've been asking for stats to be generated with 4d4 instead of 3d6--to have a lower standard deviation.

Since moving to 4d4 stats, I've learned two things.

1. People hate rolling 4d4.  That sacred cow has too many hit points, man.

2. There's still a lot of starting variation, even with 4d4.

It's pretty common for one character to have a 14 and another character have a 6.  That's a 70% chance competing against a 30% chance.

I'll admit that this is a matter of taste.  A lot of DMs and players are comfortable with that high level of variability.  A character that has one exceptionally bad stat often has other stats that make up for it.  Or it's okay to have one shitty character, because other characters in the party will compensate for them.

Sure.  Fine.

But think about this: why have that variability there in the first place?  What do you gain by it?

1. Your stats tell a story about your character.  Many players (myself included) first start to get a feel for their character during the process of rolling stats.  It's like turning over Tarot cards--each new dice roll reveals me more about my character's abilities, goals, and personality.

2. Your stats help you qualify for (and synergize with) certain classes.  Wanna be a wizard?  Well you better have good Int.

3. Your stats help your character feel distinct during gameplay.  If you are playing a burly barbarian, you expect to have a easier time jumping over a pit than an asthmatic wizard.  When you succeed on your jump, and the wizard fails, this is reinforced.  If the opposite happens, the narrative feels inconsistent, and it feels bad.

In game A, the stats range from 3 to 18.  Game B is identical, except the stats range from 7 to 14.  When you switch from Game A to Game B, how does that affect the 3 points above?

1. Minimal effect.  A Str 14 is just as exciting and descriptive as Str 18 used to be in the old system.

2. I sort of hate synergies.  Game B is preferable to Game A in this regard.

3. Game A might be ruled superior here, since the mechanics reinforce the fiction that we expect.  The barbarian outjumps the wizard 90% of the time instead of 75% of the time (percentages are speculative).  But I would argue that a more unpredictable world is potentially more exciting, since the wizard has more chances to surprise you with his jumping acumen.  And anyway, I have another solution with the Lawful Roll, below.

Anyway, here's my new proposal.

Stat Squish

First, I'm switching to roll-over for a while.  (I may switch back; I'm fickle.)

Second, ability scores are rolled with a 3d6, then divided by two (round down).  This is your bonus.  Everything is bonuses.

Throw the ability score in the trash.  We only track bonuses now.

Here's the probability breakdown.

Bonus %
+1 0.5
+2 4.2
+3 11.6
+4 21.3
+5 25.0
+6 21.3
+7 11.6
+8 4.2
+9 0.4

25% of stats will be perfectly average at +5, just like how 25% of people used to be roughly average with a 10 or 11.

A+9 is just as rare as the coveted 18 used to be.  Both are 1/216.

The DC for everything is 16.

Smoking Math

Why do this?  Because it reduces the impact of exceptional stats without affecting average stats.  Consider what your chances are of succeeding on an average stat check, for a character with average stats and a character with exceptional stats.

Old 3d6 Method
Average 10: 50% chance to succeed.
Mighty 18: 90% chance to succeed.

Squished Stat Method
Average +5: 50% chance to succeed
Mighty +9: 70% chance to succeed.

By the way, this is functionally identical to the system of ability scores and bonuses that has persisted from 3rd edition all the way into 5th.  You can wiggle around with DCs and proficiency, but it's the same beast: every 2 points you gain in a stat gives you a +5% chance of success.

Since we want a stats to have a smaller contribution, we have succeeded at our design goal.

The Lawful d10

There is one tremendous advantage of roll-over that people don't use very often.

We shrank the score into a bonus in order to reduce the contribution of the stat to a random event.  We can shrink the size of the die to reduce the contribution of randomness.  The smaller the die, the more the stat matter.

The barbarians might have Str +8 and the wizard might have Str + 4, and that might not matter much on a d20, but it matters a lot more on a d10.  So in situations when the DM wants to call for a roll that has less randomness in it (and stats are weighted more heavily), the DM should call for a Lawful Roll (as opposed to the usual roll, which is a Chaotic Roll).

The Chaotic Roll (d20)
Rolled against DC: 16

The Lawful Roll (d10)
Rolled against DC: 11

Let me prove it to you.

Barbarian (Str +8) Jumps a Pit
Chaotically: 65% chance of success
Lawfully: 80% chance of success

Wizard (Str +4) Jumps a Pit
Chaotically: 45% chance of success
Lawfully: 40% chance of success

If you're good at something, the Lawful Roll makes you better at it.  If you're bad at something, the Lawful Roll makes you worse at it.

This is a useful tool for DM's to have in their toolbox, because some things are more random than others.  Rock Climbing is more random than arm wrestling, even though I would use a Strength roll for both.

As a bonus, the Lawful roll is identical to a roll-under using the bonus as the target number.  (+8 = 80% chance of success.)  This is nearly identical to rolling under the non-squished stat with a d20, so in a way, we're right back where we started.

But What About Roll-Under?

I'm not interested in discussing the merits and pratfalls of a roll-under system compared to a roll-over system.  I've lost sleep thinking about this and I guarantee I've already considered all of them.

I think I'm going to completely purge roll-under from the GLoG, and yes, it does feel like a betrayal on a fundamental level.  And if you think angst is not appropriate for a discussion about dice mechanics, you obviously don't know me very well.

Fuck, man.  Who am I?

Joesky Tax

Idiot Birds
HD Defense leather  Peck 1d10
Move horse  Int 4

Aura of Idiocy
Anyone within 5' of an idiot bird must save or fall into an idiot rage (as the spell) except even dumber.  All you can do is make "WAWB WAWB WAAAAW" noises and attempt to break/kill things with your bare hands.  If you succeed on an Int check, you retain enough awareness to use your weapons instead of your bare hands.  The effect ends as soon as you move 5' away from the bird, the bird dies, or the bird stops making its stupid noise.

Idiot birds look like obese cormorants.  They are six feet tall and smell like fruit and dog shit.  They appear in groups of 1d4+1.  You can hear them a mile away, because they never shut the fuck up.


  1. The death of roll-under brings me joy, but I will maintain a stone face and a sense of decorum. It will be a trying time for many.
    Do you mean for Lawful to be d10 or d12? You reference both.
    The main problem I always had with rolling under your stat (one that isn't affected by how squashed the stat rolls are) is that I might want to change the DC independent of the stat. A strong person is better than a weak person at lifting 100 pounds over their head, but lifting 100 pounds over your head is different from lifting your body weight over your head. This move to rolling against a target number with a stat bonus is very exciting to me, and I want to hear how this Lawful/Chaotic thing works out.

    1. It should be d10. I made tables for both, but d10 felt lawful enough than the alternatives. (d8 was way too lawful, d12 almost worked.)

      I don't like variable DCs because that road leads to DC inflation and wrestling-with-dadism.

      DC inflation is in 3.5, where the level 1 dungeons had wooden doors, the level 4 dungeons had stone doors, the level 9 dungeons had copper doors, etc. You ever played WoW? You know how the level 1 scorpions are small and red, but the level 88 scorpions are big and green, but they're all still the same shitty scorpion? Same thing.

      Wrestling-with-dadism is when the DM looks over at the barbarian's character sheet, sees the Str score, and decides on a DC that gives the barbarian an appropriate chance of breaking the door. It's no different from the DM saying "it's a tough door, you have only a 30% chance of breaking it", which is fine, but it also negates the ability score.

      So there's only one DC: 16. Maybe 20 if you really want to have a difficult DC. But that's it. No more.

    2. AC is (or, was) already 10/12/14/16, which is (or, was) a nice spread. I feel more comfortable with a small palette of DCs, to give a feeling of variability. How are you planning on doing to-hit rolls with this system?

    3. d20 + atk vs 10/12/14/16.
      Atk hasn't changed. It ranges from +1 to +4 (for non-fighters)

  2. Man roll-under seemed to perfect to me, forever and ever. It was clean, it was neat, the player could see all the maths, everything was supposed to work... but big number = good. You just can't get away from that fundamental fact. Big number = good.

    I don't quite know how I feel about the lawful roll. I figure we've currently got three options: it definitely happens, it might happen, it definitely doesn't happen. If I was tempted to use the lawful roll because I thought someone should/shouldn't succeed because of their stats, I'd be inclined to just make it so. Also I am suspicious of d10s. I could see myself using a d12 though. What about rolling 2d10+stat (or d10+d8+stat) over the normal DC 16?

    1. 1d12+stat totally works. It's less random than a d20 roll.

      2d10+stat totally works. It's also less random than a d20 roll.

      It's a matter of taste, I think. How much weight do you want to give randomness vs stats? Some people will want more deterministic games. Some people will roll d100 + stat against a DC of 55.

  3. I'm kind of... excited? Don't plan to move from roll-under at the moment, but this is an opportunity to see if we were using roll-under because it worked or because of Arnold's founder effect.

    1. I mean, roll-under is great. I might come back around to it.

  4. For lawful vs chaotic dice, wouldn't it be better to do 3d6 vs d20 rather than d10 vs d20? Then you have the same range, but a normal vs uniform distribution, which (admittedly somewhat pedantically, but also not) to me seems more dl appropriate for those terms, and has a similar effect. Beside you're more likely to roll towards the middle of the distribution, a character with a higher stat will be more likely to succeed and a character with a lower stat more likely to fail. I haven't looked at the exact outcome distributions so maybe it works out in a way that for some reason you don't like?

    1. Not same range lol but more or less the same range

    2. 3d6 is totally viable, and mostly achieves the same goal. It's just more addition.

      I liked the d10 because it seeemed simpler, but now rereading it, maybe 3d6 against the same DC is simpler? There's definitely less addition.

    3. Hmm I still prefer the logic of uniform vs normal distributions and how that affects the "feel" of a chaotic vs lawful roll, but i ran the probabilities and how they scale and I guess the d10 with DC 11 vs 3d6 with DC 16 are pretty similar.

  5. So checks are rolled against a DC of 11 or 16. This is lovely and elegant. *kisses fingers*

    Now, do these bonuses directly affect damage? Spells known? I suppose there’s an entire list of common sense answers but I’m wondering where exactly the bonus fits in on the sheet when it comes to non-d20 rolls and class features we’re used to bonuses modifying.

    1. I still preach against high synergy. Stay bonuses don't affect anything except for stat checks. The fewer derived stats the better.

      If you have Str +7 or higher, you can use a larger damage die when you wield a 2H weapon. I think that's the last vestige. Everything else is a stat check.

  6. My two cents on Roll Under:

    One of the biggest advantages for roll under systems for me was that systems that used them were much easier to explain to new players then bonuses. For context, I run a lot of one shots with people who haven't played before, so I end up introducing a lot of people to new systems.

    Going 'to do something hard, roll and get less then this stat' is easier to explain then 'you have this stat and you ALSO have the stat BONUS and these are DIFFERENT THINGS' was periodically annoying and a large part of why I stopped running 5e for people who hadn't played RPGs before.

    I feel like a lot of people value easy first-accessibility of RPGs pretty low (not that every game has to be 100% good to jump in, Into the Odd is my go-to for introducing people nowadays and then we usually end up shifting into other games)

    GLOG already had bonuses if I recall correctly so this I guess this is a bit of a moot point (as much as this actually HAS a point) for this specifically.

    1. My GLoG changes pretty often--I'm always playtesting new ideas.

      I'm dropping the stat. It's just going to be the bonus from here on out because you're correct, it is overly confusing for no good reason.

      And yes, "roll this number or less to succeed" is wonderfully easy to explain. I will miss that simplicity. I've also seen the opposite, where I get refugees from 5e and they are constantly struggling to get behind the low=good mindset of roll-under. So part of the reason for this change is just to acknowledge the current ecosystem of TTRPGs.

  7. Not a GLOG player, but check in from time to time as I’ve always liked roll under systems, and I find the GLOG has lots of ideas that can be applied in other places. For Flashing Blades, a roll under system, I used 3d6/2, round down, +6 to generate stats, giving a 7-15 range that made character possibilities significantly less extreme, but still varied. I too had thought of just using it as a modifier and converting to Talislanta style mechanics, which is consistently ‘roll high’ with d20+modifiers, but it never quite seemed right. I don’t know why: presentation and ‘look/feel’ are interesting in how they can make something more or less acceptable to different people. I think I went with 2d6/2 instead for my modifiers, and played with lowest 2 of 3d6 and highest 2 of 3d6 as well once I discovered Over the Edge in the middish 90s. None of my players then were interested in such experiments, however.

    Interestingly, one of the versions of Tal also used d10+modifier for anything a character was unskilled at. I always liked that because it meant aptitude, as represented by the stat bonus, suddenly meant a lot more as you’ve pointed out.

    1. Huh. I should probably get around to reading Talislanta. That's really cool.

  8. So, how do you calculate Save?

    1. Not to mention, how do you test to improve stats?
      (The more I think about the first question, the more I think the narrower probability spread might negate the need for a save stat altogether, but still curious to hear your response.)

    2. The base Save is 5 + CHA Bonus, which, for average characters, works out to be 5+0=5.

      So just make Save identical to CHA + any additional bonuses from class or race.

    3. Save is just your Cha bonus. It no longer increases with level. When something *almost* kills you, you get a permanent bonus to that type of Save. DC for most Saves is 20, so Saves are tough.

    4. I'm still trying to decide on the best way to improve stats, but I'm thinking that whenever you level up, pick a stat and roll d6+d4. If you get over your current stat, it increases by 1 point.

      Playing around with: if you fail that check, you can reattempt it, but with a random stat (instead of a chosen stat).

  9. I think Roll Under was great for stuff like stat tests. D20 vs the number on your sheet. Easy! But it was awful for to-hit rolls, when it had to be modified based on the combatants' armor and attack bonus.

    1. I have to disagree, and I think this comment made me realize why I'm a little miffed by these changes. There's a very elegant solution to the problem of roll-under combat: you have to roll under your attack and over your enemy's defense, which would now be from 0-10, not 10-20 (and is easy to translate, just subtract 10 from the old defense). That's how Whitehack does it and it works out really well. It also allows you to have players roll all the dice in combat, since they can roll under their defense and over their enemy's attack to see if they're hit.

      A lot of the issues with roll-under systems stem, at least from my perspective, from a desire to do everything the way things are traditionally done. We roll 3D6 for stats and the max is 18 - but what if we rolled 2D6+3 and the max was 15? Or 3D6/2 as above, but plus 5? Suddenly the problem roll-under systems have with high stats being too strong is not that big of an issue, since you are at best only 1.5x better than average, a number that's pretty close to the math in this new system. (And the average is actually 10 and not 10.5, which warms my nerdy heart.)

      I know people never agree on what OSR means but I've always preferred "Old School Renaissance" to the other terms because of the meaning it carries. The Renaissance wasn't about going back and doing things they way they were done in the past - European scholars had already been drawing heavily on Classical and Islamic scholars for centuries. It was about surpassing the old thinkers, picking up the torch and going forward with new ideas. This is a very poetic way about talking about what type of roll to use when playing elfgames, but that's why I really love the GLOG and systems like it. They're not afraid to try new things. (Which, to be fair, is still what this is - I prefer roll under, but I do like the idea of lawful vs chaotic rolls!)

  10. So, as a novice rules hacker, I've had plenty of time to think about this as of late, and I have to ask: What is the purpose of ability checks in the GLOG, as you see it?

    D&D prior to 3E either didn't have or barely used ability checks, and most modern OSR systems focus everything on the saving throw. (Into the Odd basically reduced ability scores to the three save stats from 3E, Knave basically replaces the five original save stats with the six ability scores, etc.)

    GLOG's general philosophy toward ability checks seems to mirror saving throws in many ways; instructions are to only call for them in risky situations. But saving throws are a separate mechanic, as you detail in another post, with DC 20 instead of 16.

    So, when should the GM call for an ability check? If there is no danger, don't call for one; if there is danger, call for a saving throw. The Authority forbid you call for an Intelligence, Wisdom, or Charisma check to remember, perceive, or persuade, which are the primary functions of those ability checks in post-3E D&D.

    I hope I am not coming across as overly critical here: I like ability checks too. Part of the reason I'm asking all this is to find some justification to keep them in my own game!

    1. All good points.

      Prior to 3E, there was more emphasis on defining abilities by class rather than by stats. A fighter would have a better chance to kick down a door because he was a fighter, not because he was stronger. So they differentiated characters along a slightly different axis, but they still gave them disparate abilities.

      The GLOG uses ability checks for things that the average character would have a 50% chance of success at. Nearly everyone will have a success chance in the 40-60% range, and only a rare outlier will have a success chance of 30% or 70%. That seems to be about where I want it, for the average check.

      Stats are nice because they're flexible (you can theoretically apply them to any sort of attempt) and they don't change too much over a character's career (unlike attack rolls or saves in other games).

      I call for ability checks whenever I need to resolve that ~50% successful action. There are lots of other methods. DM Fiat ("That seems like it would have a 2-in-6 chance of working to me"). Class-based ability ("Kicking down a door is a fightery action, and you're a level 4 fighter, so you have a 4-in-6 chance of success"). As a skill ("You have 4 ranks in door kicking").

      Out of all of those, the ability score check seems the most preferrable.

      As to the distinction between Saves and regular checks--there really isn't one. I was never a fan of the whole active/reactive divide. I do intend for Saves to be difficult, though, since they are *saves*--a last, undeserved chance to avoid a bad outcome. I thought about removing the word entirely but kept it because (a) it's cool to say save vs death, and (b) I wanted a form of diagetic improvement, where people who nearly died of poison once would be more resistant to dying from it again.

      The decision to use Dex, Con, and Cha as the default saves was partially done to balance the utility of the different ability scores (especially Cha).

  11. I love it, but is the strength bonus applied to damage? I feel a d6+8 is an awful lot for ennemies with 15 HP...