Quote from: ggroy;396766What would be the probability to successfully dodge/parry/block an attack?

For example, with a 97% probability to-hit an opponent with a 60% probability to dodge/parry/blocking the attack, the probability of doing damage to the opponent is 39%.  (ie. 0.97 * 0.40 = 0.388).

If I did a Defense roll, it'd be based on Agility, I think.  So 2d10 + Agility mod, and there's a few class abilities I'd probably tweak to add bonuses to that like the Hunter's Aspect of the Monkey.

So for our unarmored commoner, it would still not given him a great chance of escape (2d10 + 0), but it would seem to at least prevented from being an insta-paste.

One type of defense roll I used back in the day for 1E AD&D, was to have a target roll a d20 and add 10-AC to the roll.  This d20 + (10-AC) was basically the "AC" for that particular attack.  This is somewhat easier to manage, than the Runequest combat system.

My previous example of a 60% dodge/parry/block defense, was more like how it's done in Runequest.

Are you using a level advancement system (ie. like D&D) or no levels at all (ie. like Runequest)?

Quote from: ggroy;396771Are you using a level advancement system (ie. like D&D) or no levels at all (ie. like Runequest)?

There is level advancement, 10 levels.  Health and Mana go up by a die code per level, and they get X number of points to spend to improve attributes.  I'm thinking 3 or 5.  5 seems a little high though.

You guys have been busy...
OK a few thoughts; Using 2d10 rather than d20 gets trickier since each additional number increases the % by more each time...1% chance of the furthest-out roll (2s, 20s), 2% chance of the next one in (3s, 19s) and so on...which is cumulative so that going from 20 crit range to 19-20 bumps your odds from 1% to 3%, 18-20 is 6% etc.

Just as some ideas:
*If you were using a roll-under type mechanic rather than roll-over, a critical threshold would be slightly easier to work out in play...e.g. something like d20 roll less than [Stat/5] = critical. Gets messy with 2d10 for the reason above.

*Or, you can have a system that distinguishes between "change in target number" and "bonus to the roll"...they might have a target number that varies by skill, while attribute adds a + to the die roll instead (which counts against getting a 20+ and criticalling).

*Or, you could shift criticalling from the attack roll to the damage roll. For example if you always used d10+weapon type modifiers for damage you could have a "threat" range (rolling up on damage) that improves from 10, to 9-10, to 8-10, etc. varying by attribute. That'd let you skip the complications from the 2d10 V-curve, if nothing else.

*Or (I think you had this idea already, though), but you could also just use a D&D type system where the chance of "critical threat" is fairly generous, but gets a confirmation roll based on the attribute score. If you wanted to heavily weight this by attribute, might need a different mechanic for confirmation than the normal d20 roll; something more like a die pool roll (e.g. roll 1d10 per 5 stat points, getting any roll of 8+ in the pool confirms; extra dice give diminishing returns as far as bumping up the probability goes with this sort of system).

JDCorley in the Haven thread pointed out a clever trick that I think is what I'm going to go with.

Since I'm using 2d10 added for my attack rolls already, this basically means I already have a percentile roll on the table too once I roll the dice.  

Which means we can essentially make the critical chance a percentile roll over.  Take the appropriate stat (Agi for melee/ranged, Int for magic) out of 100, and find the number to beat that way.

Decoupling the to-hit roll from the crit, may be easier to do.

For a tricked out rogue with a 50% crit probability, one can use d% to determine whether the rogue's successful attack is also a crit.

The question then becomes  how to map the stat mods to a particular % value for the separate crit roll using a d%.

Does the rogue (or other classes) have multiple attacks per combat round?

If the rogue with a 50% crit probability has multiple attacks per round, the crits can become quite frequent.

Rogues have Dual-Wield, which allows them an additional attack per round with their off-hand at a penalty to hit and damage, like -5 abouts.

I may also give them some additional ability(s) that allows them another free attack under some conditions, as I've done for several other classes.

Yep that's quite clever. Needs one of the d10s to be of the kind that has the tens written on it so you can tell its a 97 not a 79, but that's not hard these days.
Interestingly, does partly (but not entirely) disconnect the attack roll from the critical roll, since one of your 2d10 gets a much higher weighting. For example a 9+1 is only a 10 on 2d10 (not that high) but is a quite high 91% roll if translated to percentile.

Sort of reminds me of how hit location is figured in Dark Heresy.

One possible mapping between stats mods and % crit, could be multiplying the stat mod by 7.

For example, a tricked out rogue with a +7 agility mod could have a 49% crit probability.

A fighter with +5 strength mod could have a 35% crit probability.

A mage with a +3 intelligence mod could have a 21% crit probability.


Though this "multiply 7" mapping breaks down if the stat mod is +15 or over.  (ie. 15*7 = 105).

An even simpler mapping between stats and % crit, would be to multiply the stat in question by 2.5  (ie. not the stat mod).

For example, a tricked out rogue with an agility of 20 would have a 50% crit probability.

A fighter with a strength of 16 would have a 40% crit probability.

A mage with an intelligence of 12 would have a 30% crit probability.

EDIT:  A commoner with every stat at 6 would have a 15% crit probability, while a "boring" player character with every stat at 8 would have a 20% crit probability.


This "multiply stat by 2.5" mapping breaks down if a particular stat is 40 or over (ie. 40*2.5 = 100).