Smash the ability scores

The "Smash" maneuver from the Rules Cyclopedia has always intrigued me, not only because it's one of those small idiosyncrasies of that book absent from other versions of D&D, but also because it's one of the rare cases outside roll-under situations where the attribute is used in full, point by point, rather than just a modifier.

It goes like this:

Smash
This is a Fighter Combat Option maneuver, first available at 9th level to fighters and mystics, and at other experience point totals to demihumans (see their experience tables). With this hand-to-hand maneuver, the character automatically loses initiative and takes a - 5 penalty to the attack roll (he still gets his Strength and magic adjustments to his attack roll). 
If attack hits, the character adds his Strength bonus, magic bonuses, and his entire Strength score to his weapon's normal damage. 
For example, a Strength 17 fighter ( + 2 to attack and damage) using a sword +2 ( + 2 to attack, 1d8 + 2 damage) would perform a smash this way: He rolls to hit with a net penalty of -1 ( + 2 + 2-5). If he hits, he rolls ld8 + 21 (17+ 2+ 2) for damage!

The smash maneuver is also a solution to various combat situations, but I'll set that aside to focus on the matter of ability checks.

A persistent problem in D&D is that there are few mechanics that interact directly with the full attribute score instead of just the modifier. 

The obvious answer to this dilemma is, well, ability checks. The problem is that good examples of them are nearly absent from published D&D modules. Most are simply Dexterity tests used as if they were saving throws — which, in my view, only adds confusion. The example in Moldvay is climbing a rope, which has the odd consequence of making the thief better at climbing sheer walls than ropes. 

AD&D offers some inspiration for broader uses: though it never says so explicitly, the attribute tables include chances to learn spells and resurrection survival odds — both of which could reasonably be framed as ability checks. Strength checks have a few obvious uses — for example, open doors, which unfortunately is treated under a different mechanics, with similar results.

None of this helps much with the harder problem: finding situations that naturally call for a Wisdom or Charisma check.

Using ability checks with skills, non-weapon proficiencies, etc. seems to be a good solution. In a game like AD&D, where abilities average 12.2, it might be as simple as giving a −10 penalty to anyone untrained (minimum 1), and adding level if trained. 

So a thief with Dex 14 might start with a skill of 5, so 25% chance of success (14+1−10), reaching 90% by level 14. A warrior with the same Dex might have only 4 for all his career - and the GM might decide certain tasks are impossible for the untrained.

It doesn't matter whether you prefer roll-high or roll-low. You simply add this to a 1d20 and try to meet or beat 20 (gaining an extra 5% chance in the process, which is a good tradeoff in my opinion). So our thief would go from 30% to 95% over the course of his career.

[I'm more inclined to go roll low, but since the math is the same I'll probably ask around to see what people prefer.]

That's probably what I'll go with, because it's a lot simpler than the idea that got me writing this in the first place.

I was thinking of keeping the modifier as the standard for skills, but allowing certain situations to let you add the entire ability score instead. What would be the equivalent of Smash for other abilities? Anything done slowly could fit... maybe it could be the old-school equivalent of "taking 10":

Taking Your Time

When the character is not in a rush and is not being threatened or distracted, they may add their entire ability score — rather than just the modifier — to the check.

This brings to mind Siegfried using his prodigious strength to forge a sword despite little blacksmithing training. And it still leaves some chance of success and failure.

Another idea: allow a natural 20 to trigger a "roll again, adding the entire ability score" — letting PCs accomplish nearly impossible tasks if they're talented enough.

I may develop these further down the line, but for now I think I've found my next skill system.

BTW, I'm working on my "Old School Minimalist" again, and this time I think I'll go all the way and publish a 20-30 page PDF, after I share the whole thing here.

More d100 D&D

I've been playing with this idea again.

Basically, you have percentile values for each ability. 

Each level gets you at least 2% better at (basically) everything, with each ability point giving you 3% as a starting rule.

A level 8 PC (example above) might have Strength 13/55% and Intelligence 10/46%.

We'd basically use that 55% for skills, attacks, and magic.

Probably thieves get 20% extra on their skills, fighters on their attacks, and mages on their spellcasting.

Limit PCs to level 10 or so, and leveling bonus will not surpass differences in abilities., although the system does still work up to level 20 and more.

Saving throws use the same number. The 30–80 range is quite perfect for that.

[Strength 10 gives you 32% on level one; with Strength 20, you'd get 80% on level 10, or 100% if you're an specialist; maybe you could do ability x4 plus level x2 for slightly more competent PCs].

Combat wouldn't be hard to do; you probably need some defense stat in lieu of AC, maybe even some dodge chance, parry, etc. One reason I like d100, BTW, is because they can produce crits on doubles, and use the "tens" digit as damage, to make combat quicker and more exciting.

HP are trickier. You can't really get the same range as old-school D&D without a completely different formula.

Spells would be a lot weaker unless you spend some of your magic for additional effects.

But at this point I'm not even sure you'd need the ability scores; the percentiles would be enough. I can hardly think of a task a PC should be able to attempt without getting better through leveling. Breaking down doors, for example, could certainly improve with practice.

"But why not just roll under every ability, like in The Black Hack, etc.?" This is doable too, but I like the small increments from level instead of raising attributes directly. And the curve from my method, going from around 30% to 80%, feels more akin to D&D’s zero-to-hero style than starting at a 40–50% chance of success.

And, well, I just like the d100.

Old school dice pools

So, I just had a fast combat/mass combat idea for D&D, probably from  Chainmail or Delta's blog:

Roll 1d6/level for fighters, half as much for clerics, 1/3 for MUs.

1 misses, 2-5 hit/miss depending on AC, 6 always hits. 

Monsters only need one hit per HD and we don't even need d20s. 

The idea is making combat against dozens of opponents a bit quicker.

But come to think of it, it could be pushed into an entire system. Let's see.

Duels

Two 9-HD fighters facing off would each roll 9d6 and cause an average of 3 hits per round (assuming they hit on a 5-6), so combat would be a bit quicker than, say, B/X.

Ranged combat

Both in real life and D&D, ranged combat is not usually as efficient as mêlée combat. Maybe adding 1 or 2 to AC is enough. To avoid treating an archer like a machine gun, you can rule "missed" shots are time spent in aiming, drawing, etc., and only "hits" or 1s waste arrows.

Against a single target, maybe all damage comes from a carefully aimed single arrow; so a very powerful fighter with a magic arrow could kill a dragon immediately, but this is very rare.

Weapons

Certainly there is some nuance lost here. Let's assume everyone is using a single-handed weapon. 2H weapons might add a dice, while maces may remove a point of AC, etc.

Turn Undead

Cleric rolls 2d6/level. 

Rolling 2-5 turns one HD of undead, 6 damages them. You can alter these numbers to make the cleric more or less powerful, or maybe make turn undead a spell (see below).

Spells

Casters have 2d6/level "magic dice" per day. 

When casting a fireball, it works identically as a fighter's attack, but any 6s you roll are removed from your pool until the next rest. 

(I think I got this idea from Necropraxis).

This fixes a number of fireball problems I usually have.

Same works for curing wounds.

But what about spells that deal no damage? Maybe we could still keep the "roll to cast" and "magic dice" aspects. a 4-6 counts as a success; a 1st level spell requires only one success to function, etc.

You can use several dice to cast a 1st level spell, so you can be sure it succeeds in the first try, but that way you'll also roll more 6s and spend more dice.

Skills

Let's use "hear noise" as an example. Non-thieves have 1-in-6 chances, thieves start with 2-in-6.

So let's say a normal PC rolls 1d6, but a thief adds 1d6/level. Rolling a single 6 means success, so the thief start with 30.5% chance. By level 10, he rolls 11d6, with a 85% chance of success. He won't get to 99% until level 19-20 or so, which is nice, so there is always some chance of failure.

Maybe multiple 6s mean extraordinary success, and rolling all 1s means disaster (e.g., falling from a climb or getting caught in a trap).

Backstabbing is easy; a thief simply attacks as a fighter while backstabbing, and maybe lowers the AC by one if you want them to be really deadly.

Saves

Saves can work similarly to skills. Everyone gets 1d6 plus 1d6/level. 

Notice that the progression from 30% to 85% between level 1 and 10 is quite fitting. You can give fighters, dwarves, paladins etc. an extra die or two.

You do not usually "save" against damage; treat this like an attack against AC (see below).

HP

There is no more HP, only "hits". To make things a bit softer, I'd give each PC one hit PLUS level for fighters, level/2 for clerics etc.

Maybe you could do the same for monsters so that a 1 HD monster has 2 hits and so on. 

AC

AC now is 2 (unarmored) to 6 (plate+shield). 

If you want magic armor etc. you could go even higher, but then you'd need special rules. For example, each time you roll a 6 you can roll again and add 5 to get a result from 6 to 11.

In conclusion...

Well, if you like dice pools, you can see that you might was well play old school D&D with them and a little conversion. But you'd lose some nuance in ability scores, weapons, etc. Maybe just sticking to the d20 is easier.

Still, we have some nifty systems for mass combat, and maybe skills, spells and saves, to experiment with.

AD&D 2e reaction table

The AD&D 2e reaction table is... interesting:

Compare it to the BECMI reaction table, discussed here.

The tables are different for several reasons, but the main distinction is that the AD&D 2e table requires you to check the player characters' attitude before finding out how the monster behaves, while every other D&D table I can remember goes the opposite way: first consider the die roll, then check how the monsters behave.

Of course, in practice you can always ask how the PC's react first (or ignore the rection roll altogether, etc.). But I think it would be better to rely on initiative here - if the PC's have the initiative, they can choose to show they're friendly before the monsters decide how to react, which would certainly give them some advantage in the reaction roll.

If they LOSE initiative, the monsters "react" first - but if they are uncertain, this gives the PC's another chance to make a peaceful gesture, etc.

Another interesting aspect of the 2e table is that it can result in flight. This makes some sense as the table is affected by morale modifiers. A curious idea! Should scared monsters be friendlier? It makes some sense if they are intelligent, otherwise they should just run away if they can (which is the case if PCs are hostile).

Unfortunately, the actual morale score is irrelevant here; a monster with morale 18 and other with morale 10 are both as likely to flight or be hostile. 

Curiously, since chaotic creatures have -1 to morale checks for some reason (they are probably more cowardly and less organized), they are also more likely to be friendly, which is a mistake IMO.

Overall, the 2e table is not any clearer or better than other tables, but it has several advantages we can use - and a few disadvantages I'd like to change.

It feels too friendly to "indifferent" PCs, do not contain immediate attacks, and is organized in a 4x19 grid instead of the usual 5 entries. It also seems to lack a "cautious" column that should be the default for PCs, with equal chances of friendliness and hostility.

Maybe my ideal 2d10 table would be smaller, containing a single column instead of a grid. Give the PCs a -1 if they manage to show they're friendly before you roll (e.g., if they win initiative); let he "speaker" or "leader" make any kind of Charisma "check" you feel appropriate to change this to -2 [simply including the charisma modifier feels too extreme, IMO; it would make everyone your friend]. 

If they are hostile or attack, roll with +1 to +2 (it is unlikely you need to roll after the PCs attack).

2–3. Friendly

4–6. Positive

7–10. Curious

11–13. Indifferent

14–16. Suspicious

17–18. Negative

19–20. Hostile (fight or flight - morale check)

As you can see, I added morale only to the last entry. But you can also use it whenever intelligent NPCs feel threatened or unable to escape, to see if they negotiate or surrender.

This is not much better than the original 2d6 table. Except that 2d10 allows you some extra room to give +2 and -2 modifiers. Maybe a simpler version would be better:

2–4. Friendly 

5–7. Positive, indifferent

8–12. Cautious curiosity  

13–16. Negative, suspicious, aggressive

17–20. Hostile (fight or flight - morale check if needed)

But then again, I've written about this before... more than once! 

So I'll leave this as a small post about 2e reaction, and point you to some older posts about reaction rolls in general:

https://methodsetmadness.blogspot.com/2023/05/reaction-rolls-in-practice-osr.html

https://methodsetmadness.blogspot.com/2022/10/a-d20-reaction-check-osr.html

Stretching the d20

The d20 is granular enough for me.

Each number in the d20 represents a 5% chance.

I really don't feel the need to distinguish "stealth 46%" from "stealth 48%", for example.

I could even use smaller dice, but I play D&Dish games and I like the d20.

However, there are a few situations in which the d20 is not enough: when you want to assign chances that are extremely high (over 95%) or low (below 5%) instead of saying "automatic success/failure".

Now, you might say you don't really need that; some situations are just impossible.

But D&D/AD&D has many such cases - thief skills (and system shock) going to 99%, 1% of low-Strength folks bending bars, and optional rules to allow someone with THAC0 20 to hit negative AC.

In other words, the d20 is only insufficient in extreme cases; I'd be happy to use it in every other circumstance (which is about 90% of the time).

Like Moldvay says, "there is always a chance". Before this section, he mentions another example: "Looking down into the chasm, your character can estimate that he has a 98% chance of dying, no saving throw, if he jumps."

I can imagine many other circumstances where 1% chances would be better than either 0 or 5%. 

For example, if you want fumbles in your game, it is ludicrous that you fail ridiculously 5% of the time, especially if you're a experienced fighter. 

Even for spell mishaps, 5% chance is just too much. A powerful mage casts several spells a day and shouldn't be dealing with fumbles every other day. 

I also love critical hits; 5% of the time seems fine for having "maximum damage", for example, but I'd love having the occasional "super crit" with double or triple damage (or maybe permanent damage, etc.)

There are several ways to "stretch" the d20 at the edges if you want more than 20 possibilities. I've discussed a d100 conversion in the last post, now I'll present an alternative:

Whenever you roll a natural 20, you can try again with a +10 bonus, picking the best result.

Conversely, a natural 1 forces you to roll again with a -10 penalty and pick the worst result.

This way, a d20 can easily generate results form -10 to 30, and beyond (you'd need several 20s in a row, but you get get to 100 once every 500 billion rolls...).

This would give you a small chance to hit even -10 AC.

If using crits, you could easy say that a margin of 30, for example, will give you triple maximum damage. Awesome, but rare.

Notice this can work for ANY kind of dice. 

For example, I hate the idea that you add your Charisma bonus to a reaction roll, making an "immediate attack" impossible (although this is not how I use the table). 

But you could re-roll a natural 1 or 6 with a -3/+3 bonus, which will allow an immediate attack by anyone, if rarely.

This also allows you to assign bonuses and penalties to enemies' reactions without taking some possibilities out of the picture.

This is similar to "exploding dice", but for me it has the advantage of not taking any result off the table. Rolling a 21 is possible, but less likely than rolling 20. Rolling a 22 is even less likely, and so on.

AD&D and ability checks - from d20 to d100

I have often wondered if people playing AD&D RAW use ability checks, and how often.

From a quick look at the rules, it would seem that if you don't, having Dexterity 7 and Dexterity 14 is identical. Same for Wisdom 8 and 14.

Is that part of the reason why Dexterity and Wisdom would become some of the most common saving throws in 2024 D&D? I'm not sure. 

Certainly rolling under Dexterity was used as a saving throw in some old school modules (to avoid falling into a trap, slipping, etc.).

For all other stats, however, there is some consequences to having a few extra points. The exact number are all over: Strength 18/33 gives you +1 to hit, +3 damage, +100 encumbrance, 50% chance of forcing doors and  20% of bending bars. Strength 18/53 will give you almost entirely different numbers.

[The table below if from 2e; the numbers are similar, but notice how they almost turned open doors into "roll under"].

I have a feeling that Gygax got enamored of the d100 some time between OD&D and AD&D. While AD&D uses multiple types of die, the d100 appears often, and it seems to be useful especially when the d20 is not granular enough.

I've tried streamlining theses numbers before, and maybe replacing some of them for ability checks.

One issue with using a d20 is that you lose the finer detail of chances that are lower than 5% (e.g., bend bars) of greater than 95% (e.g., system shock).

But this is not impossible to fix either.

Just revert to the d100 when (and only when) the d20 is not granular enough to give you chance of success/failure.

We could just use ability checks with a bonus/penalty; usually, usually from -4 to +4, but -10 for extreme tasks (e.g., bend bars). If something is impossible to roll on a d20 (e.g, you need to roll under 3 but you have a -4 penalty), we could give the PC an extra chance by rolling a d100 - your chances decrease by 1% instead of 5% or each point.

[Notice we sometimes say "roll under" when we really mean "roll equal or under"; for example, rolling under Dexterity 7 means you have to roll 7 or less on the d20].

For example, you'd need Strength 11 to even try to bend bars (similarly to AD&D). This requires rolling a 1, which means 5% chance. Strength 10 could reduce that to 4%, and Strength 6 to 1%. Strength 5 makes it impossible. Strength 19 gives you a 45% chance.

[I'd probably get rid of percentile Strength, BTW].

Same reasoning for system shock: say you roll with a +4 bonus. Constitution 3 gives you a 35% chance (exactly like AD&D), since you have to roll 7 or less. Constitution 15 gives you 95%. Since there is always a small chance of failure, Constitution 16 will give you 96%, Constitution 17 will give you 97%, up to 99% if you have 19.

This also works for thief skills, which follow a similar progression (move quickly to 95% then slow down). Say you need to roll under thief level +3 to hide or move quietly; this gives you 20% chance on level one, 95% on level 15, up to 99% on level 19. Of course, you could use ability checks instead (with a -10 penalty, for these are exceptional tasks, but adding thief level).

Another use for this: hitting negative AC. If your THAC0 is 20, you have 5% chance to hit AC 0. In AD&D, you ALSO have 5% chance to hit AC -2, but... wouldn't it be smoother if your just apply the negative AC to that 5% chance? So, AC -2 gives you 3% chance of success instead of 5%. AC -4 gives you 1%, AC -5 is impossible to hit. Easy!]

Anyway, I've been thinking about these concept of "stretching the d20" for years. Now that I think of it, it probably deserves a post of its own. But this "d20 to d100" stretching is enough for AD&D, I think.

Make your own D&D skill system!

The various D&D skill/ability systems vary in how much they rely on ability scores and level; some abilities (e.g., forage in B/X) are unrelated to both things.

We could discuss endlessly about which way is better. I like a mix of both. Should a 15th level fighter be equal or better at foraging or knocking down doors than a 1st-level fighter?

Well, you can decide for yourself. Here is how.

Give a rating, from 0 to 5, to assess how important you think ability scores are. Do the same for levels - the sum should be 5 most of the time. For example, if you think ability scores are much more important than levels, you can rate abilities 4 and levels 1.

Simply multiply your ability score and your level for the number you chose, sum it up, and you have your percentile of exceeding.

For example, if you have Strength 12, Level 5, your chance to knock down a door would be 53% (12x4+5x1).

Note that you could choose in a case-by-case basis, e.g.:

- Open doors is mostly strength, but pick locks requires more levels/skills.

- Picking locks could be Dexterity improved by only 1% per level for all classes except thieves, that get 5% per level. Same for forage and rangers.

- For tasks that are too easy or too difficult, just double or halve the percentages.

This requires some adjudicating and math, but it is overall a decent solution because:

- It is quite instinctive and easy to grasp.

- Every ability point and every level matters.

- Heroes get better at EVERYTHING, if only slightly.

- It replaces thieve's skills quite well and also gives a clear answer to "what if you're not a thief"?

- Adds no complexity to the character sheet.

I'm tempted to say each level gets you at least 2% better at (basically) everything, with each ability point giving you 3% as a starting rule.

(Even better, for some old school flavor, each score could be paired with a percentile; Figgen, pictured above, has Strength 13/55%, but we'd probably need to write down some skill percentages. Or just allow "trained" PCs to "flip" the dice, so they can count a 73 as a 37, for example, which gives the thief spectacular odds, or use some kind of "advantage" system, or just add 25% chance if trained, etc).

I'm also tempted to create a whole system out of this, with fighters improving their attacks 5% per level, plus percentile magic and saving throws... but I probably won't.

So, anyway, if you needed a new skill system (we have a few dozens, and we only needed one...), you've got it.

Additional reading:

https://methodsetmadness.blogspot.com/2023/10/you-cant-solve-bx-skills-and-checks.html

https://methodsetmadness.blogspot.com/2024/10/skills-another-thing-that-old-d-got.html

Skills - another thing that old D&D got (mostly) right?

When I was still interested in contemporary D&D, I noticed that there are only a few skills that are not necessarily equivalent to ability scores. 

- Nature.

- Arcana.

- History.

- Medicine.

- Sleight of Hand.

- Perception? (mentioned here).

Forget Sleight of Hand for now; an agile PC picking pockets or opening locks is a strong archetype, despite the fact these are completely different skills in reality.

An indeed, in AD&D a thief with high Dex gets some bonuses to both picking pockets and locks.

Nature would include things like foraging, hunting, orienting, and tracking. In old school D&D, the first three are just X-in-chances, unrelated to ability scores (or class, level, etc.). AD&D adds tracking to rangers, but not much else.

Perception would include things like hear noise and finding traps. Modern D&D ties this to Wisdom, but there is no reason to think a wise cleric is more perceptive than a quick-thinking warrior or a sly thief.

The AD&D thief gets no Dex (or Wis) bonus to hear noise, but curiously gets a Dex bonus to find traps... A mistake, IMO.

Then there is Arcana and Medicine. These are not skills in B/X or AD&D - they are just things the mage and cleric are supposed to do (although using spells instead of skills).

Finally, there is no skill for lore - players discover that by themselves, not characters.

I don't have much of a conclusion here. Except that, maybe, some skills work nicely with ability scores, while others could simply IGNORE ability scores. 

In a modern game, having two types of skills would look strange - in contemporary D&D, for example, basically ALL d20 rolls include some ability score (attacks, saves, checks/skills). I'm not sure I'd do it myself. But it is something to consider, as it seems to work quite well in old-school D&D.

AD&D ability tests, streamlined

 As I've mentioned in my Hyperborea review (check it pout!), I really liked how the game tries to streamline AD&D ability checks: 

This is a simpler version of the (much more complex) original, which included percentile Strength (see OSRIC, for example):

TRENGTH	BONUS TO HIT	BONUS TO DAMAGE	ENCUMBRANCE ADJUSTMENT (IN LBS)	MINOR TESTS, E.G. FORCING DOORS (CHANCE ON D6)	MAJOR TESTS, E.G. BENDING BARS AND LIFTING PORTCULLIS (CHANCE ON D%)
3	-3	-1	-35	1	0
4-5	-2	-1	-25	1	0
6-7	-1	0	-15	1	0
8-9	0	0	0	1-2	1
10-11	0	0	0	1-2	2
12-13	0	0	+10	1-2	4
14-15	0	0	+20	1-2	7
16	0	+1	+35	1-3	10
17	+1	+1	+50	1-3	13
18	+1	+2	+75	1-3	16
18.01-18.50	+1	+3	+100	1-3	20
18.51-18.75	+2	+3	+125	1-4	25
18.76-18.90	+2	+4	+150	1-4	30
18.91-18.99	+2	+5	+200	1-4 (1 in 6 extraordinary success)	35
19	+3	+6	+300	1-5 (1 in 6 extraordinary success)	40

Although I like the simpler version I think it would be easier to go even simpler - while keeping vaguely similar chances.

Here is a simple formula:

- Ordinary ability checks: roll under ability.

- Extraordinary checks: roll under ability-10 if your ability is remarkable enough (13+, which is where you start getting modifiers), otherwise it is a % roll. 

Alternatively, just make a percentile roll with HALF you ability score if you want to keep things more similar to the original, or your whole ability if truly exceptional (17+).

[You could probably achieve interesting results with 2d20 for extraordinary checks: less than 1% for Strength 3, and about 38% for Strength 18. But this is YET ANOTHER system to try someday...). 

Other tables could be similarly replaced: 

- "Survive" checks (Constitution): roll under ability, you get a +4 bonus.

- Thieves' skills are extraordinary checks, but add level to your ability before rolling.

Of course, the exact numbers do not really matter. It depends on what you're trying to achieve.

This is just another example of ability checks I found interesting (I probably wrote more than a dozen in this blog already, this was probably the most recent, using 1d30).

Anyway, just another skill system for you to play with if you don't like sheets with lots of data or consulting tables during the game.

You can't solve B/X skills and checks (same goes for AD&D etc.)... and ONE MORE FIX

I've been working on a minimalist D&D for ages. 

B/X checks are a bit messy, as we discussed many times before; they could be vastly simplified.

For example: by RAW, a thief usually has better chances climbing sheer walls than climbing a rope (and uses two completely different systems), while the Halfling uses two different systems to hide (90% in the wilderness, 2-in-6 otherwise).

In total, there are at least six different systems:

- D20 roll high (attacks, saves).

- D20 roll low (ability checks).

- X-in-6 chances (foraging, getting lost, etc.).

- 2d6 roll high (reaction).

- 2d6 roll low (morale).

- Percentages.

If you consider wether the roll is affected by your abilities, level, both or none, the situation gets even more complicated:

I'm still ignoring things such as damage, surprise, initiative, 10% chance of arcane error for thieves, 90% chance of drowning, etc.

I tried to fix that more than half a dozen times in this blog.

But, ultimately, the problem becomes obvious: you cannot fix this because the fix is entirely dependent on matters of personal taste.

To put some order into it, you'd have to decide:

- Does it need any more order than that? Should the mechanics be more unified than that?

If the answer is positive, you'd have to choose:

- Do you prefer 1d6, 2d6, 1d20 or 1d100?

- Do you prefer roll high or low?

- Should your chances improve with a high ability score?

- Should your chances improve with a high level?

Assuming you can make a decision for each of these circumstances, you'd get at least thirty possible answers - none strictly better than the other, since 1d20 is no better than 1d6 or 1d100 and roll high/low each have their defenders.

[Each has its benefits, however; I find "30% chance" extremely intuitive, for example, and "rolling a natural 20" a fun thing to have in the game].

Now, if the answer is negative - i.e., if you're willing to accept the mess - things get even more complicated. 

Because - unless you accepted "because the book said so" as your only criteria, you'd have to ask yourself:

- Should you level affect your chances of foraging? 

- Should you Dex affect move silently?

- Should thieves have an easier time climbing a rope?

- Et cetera ad nauseam.

Now you don't have thirty answers - you have thirty answers for each check.

And I haven't even started talking about quantity. 

- How hard should the checks be? 

- Should there be different "DCs"?

- How much should your level affect this?

- What about your abilities? 

- Should level matter more than ability or vice-versa? 

- Etc.

A 5e (Solomonic) solution

As a small aside, 5e solves this mostly by mentioning three types of rolls (attacks, saves, ability checks), but they all function in a roughly similar fashion: roll 1d20 + ability mod + proficiency.

Proficiency (defined by level) and ability mods have similar weights. 

(When I say "Solomonic" I do not mean "wise, but "cutting things in half".)

Sometimes proficiency doesn't apply, sometimes it is halved or doubled, so in the end a table would look somewhat better than B/X but not much simpler, and still need an external DC to function, which B/X usually doesn't.

A B/X (Solomonic) solution

Since there are no right answers, we can come up with our own. Target 20 is one of my favorites, but does not cover ability checks or 1d6 tasks explicitly. 

Here is one of mine.

- Do you prefer 1d6, 2d6, 1d20 or 1d100? 1d20.

- Do you prefer roll high or low? High.

- Should your chances improve with a high ability score? Yes.

- Should your chances improve with a high level? Yes.

Target 20 already covers attacks, saves and skills; we'd need a system for non-thief skills (1d6) and ability checks.

My suggestion:

Roll 1d20, add (ability+level)/2, Target 20.

To keep thief skills similar, just use:

Roll 1d20, add (ability/2)+level, Target 20.

E.g., Dex 14, level 8 means +15 to hide if you're a thief, +11 if you are not.

+11 seems a lot for a non-thief, but in the unlikely event you have a Dex 14 MU, it gives the sneaky bastard lots of personality!

If you prefer to ignore levels for non-thieves, I think this method (1d20+ability/2) is still better than simple ability checks as it makes things a bit harder (and closer to the usual 1-in-6 chances for general tasks). Or, to add more complexity, you only get to add half level if the task is somehow related to your class. But I'm not sure this is even necessary.

One side effect I enjoy is that every point matters - Dex 10 is different from Dex 11 at least half the time.

Similar methods have been used in "Action Throws" and BFRPG optional rules ("Ability Rolls"). 

Well, TBH I might have mentioned something similar in this blog before; I'm sorry to repeat myself, this seems to be an unsolvable matter indeed... but at least we gave it a shot! 

Recommended reading:

https://methodsetmadness.blogspot.com/2023/02/class-x-skill-some-quick-thoughts.html

https://methodsetmadness.blogspot.com/2023/10/point-buy-d-osr.html

https://methodsetmadness.blogspot.com/2016/06/ability-scores-making-each-point-matter.html
https://mythlands-erce.blogspot.com/2023/10/using-ability-checks-in-bx.html
https://mythlands-erce.blogspot.com/2023/10/ability-score-improvements-have-been.html

https://mythlands-erce.blogspot.com/2023/10/how-difficulty-class-and-d20-engine.html

Point-buy D&D (OSR)

I have played point-buy systems for decades. 

In some of these systems, you choose your abilities scores, powers, perks, etc. by spending a limited amount of points - instead of rolling dice and picking a class.

Point-buy mechanics are not uncommon in D&D; even in OD&D (IIRC) you could reduce one ability score to improve another. Later on, you had proficiencies to choose. 

In modern D&D, you get to pick your ability scores (and sometimes skills) with points too.

The best thing about this is that you can customize your character however you like. 

Do you want a spell-less ranger, nature paladin, witcher, white mage, etc.? Easy to do without multiclassing rules.

In systems like Runequest or Savage Worlds, you don't even need classes; you just create the character you imagine.

The main problem, of course, is analysis paralysis.

[Another downside is that all PCs of the same class become "optimal" and "samey" if you don't introduce some randomness].

Too much choice becomes burdensome. At the very least, you'd need a few examples or templates to help players out, unless they are familiar with the system (with that said, I played such systems for decades without much issue).

Anyway, I have often wondered if old school D&D could be easily reduced to a point buy system - probably after playing Lamentations of the Flame Princess, which uses a similar system for skills.

by James West.

Ideally, we'd have a number of things to "buy" with a similar cost. 

Let's start with 3 points per level (maybe 5 on level 1). You can assign them to (no more than one point per level to the same ability/skill/etc.):

- Attack bonus.
- Saves (bonus apply to all saves).
- Ability scores improvements
- Skills (1-in-6 otherwise).
- Spells (1 point per spell).
- Spell-casting.
- Turn undead.
- Feats/features.

Some entries deserve special consideration.

Attack bonus probably requires at least 1/3 of your current level, to keep in line with old-school mages.

Saves are the same (no matter if using a single save or more than one).

Ability scores improvements enhance your abilities, no matter how you define them in the first place (3d6 in order, standard array, etc.). To balance things out, I'd probably use WotC modifiers, i.e., +1 for 12-13, +2 for 14-15, +3 for 16-17, etc.

Skills replace thief skills but also ranger stuff (forage, hunt, directions). Everyone starts with a 1-in-6 chance, it becomes 2-in-6 by spending a point etc. 6-in-6 means you roll 2d6 and only fail if both dice get 6 (which means about 97% chance of success).

I'd reduce thief skills to five or six (see below).

Spell-casting means your MU level. Spellcasting 3 means you cast as a 3rd level MU. You still have to learn the spells, and you probably need at least the same number as you spellcasting (e.g., at least 3 spells for spellcasting 3).

To Turn Undead you roll 1d6 and must beat the target's HD by 4 or more. TU 2 means you roll 1d6+2, etc. A margin of 8 or more means destruction.

Feats are various perks, including all existing features. Some of them could cost more than 1 point (e.g., multiple attacks) and they might be limited (e.g., one for every three levels) to reduce complexity. You can find many examples in Old School Feats - they'd cost 2 points each.

Let's try to create a character - say, a 6th level thief, with 20 points - that will feel similar to the original version.

- Attack bonus: +3.

- Saves: +3.

- Skills: +12 (3-in-6 for six skills).

- Feats: +2 (backstab or read languages).

This is pretty close to B/X. 

What about a fighter?

- Attack bonus: +6.

- Saves: +6.

- Ability scores: +4 CON (for the extra HP).

- 4 extra points to spend as needed - probably a feat giving him an extra attack.

Eh, not perfect. The attack bonus is closer to AD&D than B/X, but I like the AD&D progression better anyway.

Cleric:

- Attack bonus: +3.

- Saves: +3.

- TU: +6.

- Spell-casting: 4.

- Spells: 4.

Close enough (I prefer the BECMI progression for cleric spells). 

If the cleric has some taboos, he might get an extra point or two, but his budget is pretty tight.

Magic-user:

- Attack bonus: +2.

- Saves: +2.

- Spell-casting: 6.

- Spells: 6.

- 4 extra points to spend as needed.

I like it. Maybe the MU can get some Lore skill, etc. 

The fact that the MU becomes stronger at lower levels and weaker at higher levels (fewer spells) is a fortunate consequence of this system.

Notice that this assumes no classes - so, same XP for everybody. Also, same HP - unless you get a CON bonus, etc.

This means the MU needs some nerfing.

The fighter loses some unique skills (wearing all weapons and armor - unless you want to count those as feats or, conversely, taboos), but gains more points to spend on abilities and features.

The cleric will be fine. It is powerful enough already.

The thief will probably need to specialize in three or four skills (and maybe they can be reduced to find, notice, climb, tinker, and stealth; with other skills such as lore, healing and nature added to other classes).

Overall... I think this is doable. 

This allows you to create a paladin or ranger (just add a nature skill) with relative ease, and to customize your own character/class.

"Racial" features can be bought by level 1, with humans getting extra points to spend.

Come to think of it, this could be the basis for an updated version of Old School Feats someday. Or an entirely new thing...

EDIT: Just found out this is a thing in AD&D 2e. The 2e DMG has a class-creation system, which is not quite I'm doing here, but  Skills and Powers contains an actual point buy system that is much harder, but still very cool. This from that books description:

About Point-Based Characters. The idea of point-based RPG characters dates back to at least Melee (1977), the predecessor to The Fantasy Trip (1980). It was popularized by Champions (1981) and has since become a mainstay of the roleplaying industry.

However, even in 1995, the idea still hadn't been officially incorporated into AD&D, which instead focused on random rolls to generate characteristics, linked with rigid class and level structures that didn't give players any room for variance in their characters. The closest that AD&D came to point-based characters was in Unearthed Arcana (1985), which offered some alternative methods for rolling lots of characteristic dice to try and generate a specific character class that the player was seeking. AD&D second edition (1989) similarly provided some methods to let players add extra dice to certain characteristics during character generation.

Skills & Powers dramatically changed this by offering a point-buy system that let players not only purchase characteristic points and proficiencies, but also allowed them to choose which class abilities that they wanted to buy. It allowed considerable variation, and thus players could have characters with "out-of-class" weapons, or even a Conan-esque fighter who could both fight and move silently. Skills & Powers even included traits (advantages) and disadvantages - two notable elements of point-based character systems that help to add detail and depth to characters.

Contains affiliate links. By purchasing stuff through affiliate links you're helping to support this blog.

Target 30

I must have proposed a dozen different systems for D&D skill/tasks already. 

B/X, one of the simplest forms of D&D available, has at least three or four - often for the same kind of test (e.g., hiding or climbing), sometimes even for the same character (e.g., the Halfling). Finding the perfect method is elusive, since each has its pros and cons. Usually, I favor methods that:

- Take levels AND ability scores into account.
- Use "roll high", with a natural 20 being the best result.
- Fit well with other rolls, such as combat, saving throws and thief skills.

My favorite so far is "fractional skills", which I refined to use in Dark Fantasy Basic, and still using in my current campaign. Xd6 roll under is also a favorite because the results match well with thief skills.

But I think there is another option. I have been a fan or Target 20 for a long time, and I have recently tried be67 with my friend Jens, which might have given me this idea:

Roll 1d20 + ability score + level (if appropriate), with a target of 30.

For example, a 4th level thief with Dex 14 rolls 1d20+14+4 to open locks, succeeding if he rolls 12 or more (i.e., 45% of the time). By level 14th, he is succeeding 95% of the time.

This system works especially well for me for a couple of reasons. First, my PCs have ability scores that average about 11 by first level*. Second, they have a few opportunities of raising abilities scores with feats. This brings us closer to an ideal Target 18 - i.e., close to 1-in-6 chances for untrained PCs (as B/X intended), but adjusted to their ability scores. Third, it makes each point matter for ability scores.

* I'm using something slightly more benevolent than "3d6 in order" to begin with (average 10.5 for each ability score). If you're using 4d6 drop lowest, this works even better to create heroic PCs. An average ability score of 12.2 is very close to 1-in-6 chances. And it would work perfectly for Knave too, as the abilities average more than 11.

To smooth things out, there is ALWAYS a chance of success or failure. If you roll a natural 20 and still fail, you can try again with a +10 bonus. Conversely, if you roll a natural 1 and would still succeed (e.g., a level 14 thief with 16 Dexterity), you must roll again with a -10 penalty or fail.

This solution seems decent for climbing, jumping, hiding, tracking, finding, etc. Use 25 for intermediate tasks or "minor tests" (e.g., open doors), 20 for most simple tasks (almost equivalent to "d20 roll under"), 35 for the nearly impossible (e.g. "bend bars" or the thief lesser skills such as hear noise and read languages).

The DC for surviving Resurrection/System shock is 15. Very close to AD&D.

The DC to understand spells is 20. Close to AD&D again.

I'm not sure I would use this for spellcasting and combat, however; the difference from B/X would be too great. Which is a downside compared to Dark Fantasy Basic, that uses a single system for everything.

Curiously, I think something similar to this could work well even for 5e D&D. Add 10 to the usual DCs (use 22 to 24 as default), reduce the proficiency bonus (maybe 1/4 level, apply to all saves and skills, half level+2 if you have expertise). If you are not familiar with 5e, this would require a longer explanation, which I'll avoid here since I'm not playing 5e anymore. But it could easily be the basis for a minimalist version of 5e.

Well, I like it. Maybe I should use it if I write something AD&Dish in the future...

Reaction rolls in practice (OSR)

Here is a small procedure for reaction rolls.

1. Goals and disposition. Decide (or roll for) goals and disposition, in any order.

2. Role-playing and actions. Ask what the PCs say and do.

3. Roll to persuade and reactions. Roll to see how the NPCs react, if it isn't obvious.

Let's analyze each step.

1. Goals and disposition

Goals could be obvious. A beast is looking for food, or protecting if found in lair. Bandits are looking for easy money, not a bloody fight. Etc. But what if you find a chimera in the woods? Well, Teratogenicon has suggestions of goals for most monster types. Here is one that could work for supernatural, chaotic "monsters" in general.

d6 Goals

1 Hate. I want to destroy all natural creatures.

2 Cruelty. I like to play with my prey.

3 Survival. Everyone thinks I’m a menace. I’ll destroy them before they destroy me!

4 Predator. I kill because I need to eat.

5 Collection. I enjoy a specific type of object (gold, shiny things, iron) or flesh (hearts, eyes, fingers).

6 Fear. I must protect myself from anything that comes near!

Disposition is friendly/hostile. Roll 2d6 as usual. Charisma does NOT affect this roll.

Decide in any order. You can roll for disposition and then decide if the wolf is hostile because it is hungry or territorial, or find a goal (hunger) and then roll disposition (a starved wolf might be aggressive or weak and docile enough to beg for food).

2. Role-playing and actions

Ask what the PCs say and do. This part is role-playing and doesn't require a roll. Do the PCs offer the beast some food? Or prepare their weapons when they see a misshapen humanoid? Etc.

This is the time to consider languages, and maybe also alignment, class, race, etc.

The NPCs reaction will sometimes be obvious - for example, bandits that outnumber the PCs seeing weapons drawn, etc. If not obvious, proceed to the next step.

3. Roll to persuade and reactions

If you are unsure if the bandits will accept a bribe to simply walk away, or if the wolf is docile enough to accept meat from humans, make a roll. You can use 2d6, but since you already have a disposition, a good roll will mean it improves by one or two steps (from hostile to uncertain, for example), and a bad roll means the opposite. Charisma applies here.

Since my current game (Dark Fantasy Basic) uses d20 skills, I just make a persuasion check (DC 15) to change disposition.

Friendly bandits (step 1), when offered a small bribe (step 2) and a bad roll (step 3) will ask for more, not attack immediately. But if they are neutral, a small offer and a bad roll will indicate they require A LOT more - maybe ALL OF IT, while a terrible roll means immediate attack ("you try to fool us? Now die!").

And that`s it for the procedure. 

But let me ramble on a bit. After I show you the 2d6 table I'm using (also Teratogenicon):

Why I´m writing this

I have just realized I had not been not using reaction rules in my current sandbox campaign, defaulting instead to "the NPCs will do the predictable thing according to their motivations and your respond to the PCs according to role-playing".

It works fine. But random encounters become a bit predictable and stale. For example, the PCs encountered some wolves, and I thought the wolves wouldn't approach a big group of humans for no reason. The PCs did nothing, so they walked away. Which was okay - but if I had rolled aggressive or friendly wolves, I could have rolled with it and have a more memorable encounter.

Likewise, when they met a group of clerics when looking for a temple, I made the clerics immediately friendly due to circumstances. But again, what if they were suspicious of the PCs motives? Could be interesting, too. 

I have started thinking and talking online about this... And then I remembered I wrote about this before. But I thought it was worth another post.

Initiative, surprise and single-roll reaction

While I like the procedure outline above, it takes three steps instead of a single 2d6 roll. The results are more nuanced and varied, but maybe you prefer a quicker method. 

If that's the case, just use the traditional 2d6. Do not add charisma unless the PCs win initiative (or surprise etc.) and try to talk to their foes. Add circumstantial modifiers as needed (-4 to +4; e.g., NPC is a guard and PCs are invading, the weapons are drawn, NPC has a different language, etc.). 

Rolling 1 on both dice means an immediate hostility regardless, and double 6s mean a positive attitude, within reason.

Even if you don't want the quicker method, using initiative in this way gives PCs meaningful choices. Do they attack now, or do they concede initiative to a potential enemy?

Alignment

In the absence of explicit goals for most monsters, you can use alignment as an indication. Chaotic monsters are more likely to attack when outnumber, cheat, steal, hide, break promises, etc., while lawful ones might be more honorable even when hostile ("I am bound to defend this bridge with my life, scum! Retreat or perish!")

As I mentioned before, alignment would be more useful if more specific.

Morale and courage

As seen on the table above, "hostile" doesn't mean "attack" necessarily. A weak foe might simply run.

Morale can be used to see how the NPCs respond to a credible risk or threat. You can add Charisma to the roll if you want, or require a Charisma check to trigger the morale roll when in doubt (as always, with circumstantial modifiers).

Recommended reading:

https://methodsetmadness.blogspot.com/2016/12/how-to-run-npcs-in-combat-days-of.html

https://methodsetmadness.blogspot.com/2017/09/social-skills-role-playing-versus-roll.html

https://methodsetmadness.blogspot.com/2021/05/monster-statblocks-how-good-is-ad.html

https://methodsetmadness.blogspot.com/2022/10/a-d20-reaction-check-osr.html

https://goblinpunch.blogspot.com/2023/03/how-to-handle-parley-as-osr-dm.html

Class x skill - some quick thoughts

It is a discussion almost as old as RPGs: is it better to have distinct classes (fighter, mage, thief, etc.) or different skills (combat, magic, stealth, nature, etc.) that everyone can access?

There are also hybrid approaches - my Dark Fantasy Basic, for example, uses classes as a "shortcut" for certain skills and feats. Elthos has classes that allow you to become better with certain skills. D&D 5e has skills that basically anyone can pick with the right feat, but certain classes get more/better skills.

It is a matter of taste, of course, but each method has its pros and cons.

Class-based games are great when your group is a "Fellowship of the Ring", where everyone has different abilities that are clearly defined by their archetypes: the warrior, the mage, the ranger, and even the elf, the hobbit, etc.

Skill-based games excel in a "Knights of the Round Table"* scenario. Everyone has similar abilities**, but some are more skilled than others. It is also perfect for teams of detectives, soldiers, criminals ***, etc.

* I've found a similar comparison reading "Of Dice and Men", which inspired this post. Expect a review soon!

** Notice that Arthurian knights are also archetypes - but maybe these archetypes are less obvious, and also maybe not as strong as the "knight" archetype that includes all of them.

*** And elves! If you have an "elf" class, a band of elves can become too uniform; it would be better if they had different classes or skills. If you have a single elf in the party, however, it can be an archetype in itself.

Skill-based games seem suited for realistic games - because in real life, archetypes are vague and abstract, while in myth they are much stronger. In any case, archetypes are incredibly useful to create characters - even in skill-based games, it is good to have some archetypes to play with (which justifies hybrid approaches).

In theory, you could use professions or specialties instead of archetypes to create a team: say, a quarterback, a running back, a receiver, a kicker, a punter, etc. However, this cannot be "classes" in most games because a profession or job is insufficient to describe a real person - unlike archetypes, that are much broader. In other words, even games that have "profession" as an important part of PC creation usually include skills.

Classes are also very useful for world building; creating skills for every NPC is a hassle, but it is great for "player character building".

On the other hand, one should be careful to avoid creating a boring/weak class system by adopting classes do not represent strong archetypes. For example, archetypes such as "Strong Guy" or "Half-caster" might make sense within the rules, but are not by themselves strong enough to represent a class.  A "Witcher" class, on the other hand, is only a strong archetype because it has been drilled into our mind through books, games and the TV series. Likewise for paladins (D&D), Night's Watch (ASOIAF), etc. They are familiar enough to represent archetypes of their own. If you do not have specific in-universe archetypes, considering falling back into more recognizable ones: "Arcane Warrior", Holy Knight, Ranger, etc.

My preference? As suggested above, I like the hybrid approach. Start with an archetype that suggests some skills (and feats, etc.) and then add more details as you go. This allows quickly PC (and NPC) creation while also leaving room for customization. This is the approach I used with Old School Feats. Here is one example:

So, the "class packages" are ready for you if you just want to choose one. But you can also customize your characters or create multi-classes by simply picking form other lists. For me, this is the best of both worlds.