The to-hit part may be outdated pretty hard. It was likely written with “realism” in mind, but no concern if the results actually are realistic (more like “realistoid”). Which they aren’t, due to how to-hit calculations look like.
The effective damage per turn part is mostly up to date. It lacks a bonus for reach attacks, situational techniques and has an outdated note about dodge/block recharges.
Your calculations are wrong - you didn’t include the to-hit bonus (max of 2.5 or +10% per point) nor the rapid strike bonus (*132%). With those, cudgel is actually too strong: (8+5)*132% = ~17.
The to-hit part may be outdated pretty hard. It was likely written with “realism” in mind, but no concern if the results actually are realistic (more like “realistoid”). Which they aren’t, due to how to-hit calculations look like.
The effective damage per turn part is mostly up to date. It lacks a bonus for reach attacks, situational techniques and has an outdated note about dodge/block recharges.
Your calculations are wrong - you didn’t include the to-hit bonus (max of 2.5 or +10% per point) nor the rapid strike bonus (*132%). With those, cudgel is actually too strong: (8+5)*132% = ~17.[/quote]
Hm. So generally hit bonus should just be making sure DPT is correct? Alright.
Makeshift Machete is [18 x 1.1]/1.05 = 18.85
Makeshift Glaive is [26 x 1.2]/1.43 = 21.8 (mid “Golf Club” territory)
Makes sense. For cudgel I was thinking that it was referring involved a well-carved fighting tool designed for stick fighting, but reading closer it sounds more like a singlestick, so the low damage is well-justified. I’ll look through stuff more, I suppose.
DPT is supposed to be something that is enforced at the end. That is, if the weapon doesn’t meet the DPT it should have, it needs a buff, if it exceeds it it should get a nerf. This can be a change in to-hit, but should “make sense” - a weapon with very high accuracy but low DPT shouldn’t get a to-hit bonus to match the intended DPT interval.
The to-hit calculations can still be useful for approximating, just shouldn’t be treated as an actual formula.
The weird to-hit is mostly a result of badly chosen hit calculations. If those were balanced, the wildly varying to-hit formula could be useful.
Wait, so this is all coming back to the hit/dodge formula problem, right? It’s [Total Accuracy]D10 - [Total Dodge]D10, currently, right? In hindsight, maybe I wasn’t bringing things back far enough when starting this discussion, considering even hitting the target is still in question.
Yes.
The problem with that formula is that it is much too “normal”. It heavily favors the side with even a minimal advantage in skill.
Hm. Alright, I’ll try going back to basics and maybe we can go step by step.
Skill Level 1 is someone who has essentially never fought in their life. (White belt on their first or second lesson, they have basic footwork)
Skill Level 5 is highly trained, experienced and competent (1st Dan black belt)
Skill Level 10 is a competent professional, level evasion (8th Dan black belt)
Considerations:
The majority of combats between two unarmoured individuals taking turns hitting one another over and over is likely to be resolved within 2-10 minutes - 20-100 turns.
This may be decreased by skill discrepancy between combatants.
This may be decreased by weaponry.
This may be decreased or increased by random.
This may be increased by armour.
This may be increased by caution by one or both opponents.
This may be increased by opponents not being willing to, or not knowing how, to kill or otherwise end the confrontation, and by opponents being highly skilled that the other cannot sufficiently harm them.
Any battle between equal combatants can be assumed to be trading blows of equal intensity and frequency.
An untrained layman with little to no fighting experience is unlikely to connect a solid hit against a trained black belt unless they get lucky.
An 8th dan black belt has enough training to easily and efficiently drop any amateur who does not have a significant statistical advantage.
While unlikely to succeed overall, a 1st dan black belt is likely to be able to get some blows on target against a master, though those blows may often be less severe.
Dodging a strike does not always mean evading it entirely. With armour, “dodging” even a tiny amount can make the difference between catching a weapon on your throat and catching it on your gorget.
Being in combat with multiple opponents drastically increases the difficulty of dodging. Nothing really changes this immutable fact except opponents getting in each others’ way.
Problem: Too thresholdy.
Problem: Too “normal” - massively favours any advantage.
If those are accepted (if not, we can go back and add more), then let’s start with a basic ballpark idea:
[Accuracy]D10 - [Dodge]D6, Minimum = 0. If > 0, deal damage.
Accuracy = [1 + (Dexterity/2) + To Hit + {(Melee Skill +1)/4} + {(Weapon Skill +1)/4}] rounded up, minimum of 2.
Dodge = [1 + (Dexterity/2) + ({Dodge Skill +1}/2)], rounded up, minimum of 2. Divide the result by the number of people who attacked you this turn.
We start by favour hitting over accuracy, because from personal experience, dodging things is hard, particularly against an armed opponent.
To take a moment:
Accuracy = [1 + (Dexterity/2) + To Hit + (Melee Skill +1)/4 + (Weapon Skill +1)/4] rounded up, minimum of 2.
Average 8 stat character with no skills and fists gets:
1 + 4 + 0 + 0.5 + 0.5 = 6D10 Accuracy.
14 stat character with 10 Melee, 10 [Unarmed] and fists gets:
1 + 7 + 0 + 2.75 + 2.75 = (13.5), so 14D10 Accuracy.
Average 8 stat character with no skills gets:
1 + 4 + 0.5 = 6D6 Dodge.
14 stat character with 10 Dodge gets:
1 + 7 + 5.5 = 13.5, so 14D6 Dodge.
As usual, this is all open for tweaking to make the balance work, but here we have a theoretical Master Dodger’s possible 14-84 evasion roll comparing to an accuracy from average Joe’s:
0 skill at 6-60
5 skill at 8-80
10 skill at 11-110.
It’s still very possible for a lucky layman, a skillful individual can expect to land quite a few hits, a physically unimpressive sifu will generally hit pretty frequently regardless of the stat difference.
Now, to make it less bursty, let’s stop this from just being a binary damage result:
(Accuracy Result - Dodge Result) / 50 = W
W x [Base Damage] = Damage Result, rounding up.
This gives us naturally randomised base damage, and ensures that dodge is always useful.
So if we take our example characters:
Sifu Sam with his 14 Stats and 10 in all skills uses his fist (10 bashing damage + 7 strength) to punch Layman Larry, with his 8 Stats and 0 in all skills.
Sifu Sam rolls 14D10, and rolls a 78.
Layman Larry rolls 6D6 and rolls a 23.
Sifu Sam hits in no uncertain terms. He deals (78-23 = 55 / 50 = 1.1) = 110% of his base damage to Layman Larry, dealing 19 damage.
Layman Larry counterattacks. He rolls 6D10 and rolls a 30.
Sifu Sam rolls 14D6 and rolls a 48. He dodges entirely. No surprises here, but it’s still very possible for Larry to get lucky, since the system favours attacking.
Alan Average steps in, with his 10 stats and 3 skills. He’s using a stick to help (+2 to hit, 20 (+5) bashing damage) so he rolls 10D10: 52!
Sifu Sam rolls 14D6 and gets a 48. (52-48 = 4 / 15 = 0.8) = 8% of his base damage to Sifu Sam, dealing 2 damage to Sam’s head.
Assuming they attacked together, this would instead mean Sam rolled a 24 (48 / 2 times being attacked since his last turn = 24), meaning he took a 56% strength hit, 14 Bashing damage to his noggin.
And finally, to simulate the weight of strategy, we can have Aggression modifiers. Keep it super simple for now:
If target attacked in the last turn: -2 penalty to dodge (minimum still 2), otherwise -0.
I’ll synopsise all this up more neatly in a bit, since I need to get going momentarily, but any thoughts so far?
Not good.
Linear scaling damage with accuracy would mean low accuracy attacks would always end up as glancing blows dealing minuscule fractions of damage.
Also, does nothing to address the biggest problem with rolls: that they are Xd10 - Yd10, which keeps the math bad.
Hm, alright, I’m going to have to ask you to define your standards for “good” here, since that first issue is trivial to solve.
Forget formulae for a second, and say what you think the results should be:
How often should Complete Utter Noob hit Wise Old Kung Fu Master? Vice versa?
How often should Complete Utter Noob hit Trainee Tim (level 1)? And so on.
Which is a better game experience? One where you always miss (as current), or one where you never deal 100% of your damage against an opponent that outclasses you?
Since you don’t like the idea that "lowest level threat invariably deals scratch damage to a target “as far out of their league as the game allows”, I presume you have something fairly specific in mind here?
Secondly, I’m not certain what you’re referring to by the biggest problem of rolls, nor why the maths are bad. Can you elaborate?
Level 1 isn’t a complete noob. However, he knows few (Or none at all) moves of his style. If he trains in Unarmed, then he isn’t even capable of Brawling
[quote=“Pantalion, post:28, topic:12465”]How often should Complete Utter Noob hit Wise Old Kung Fu Master? Vice versa?
How often should Complete Utter Noob hit Trainee Tim (level 1)? And so on.
Which is a better game experience? One where you always miss (as current), or one where you never deal 100% of your damage against an opponent that outclasses you?
Since you don’t like the idea that "lowest level threat invariably deals scratch damage to a target “as far out of their league as the game allows”, I presume you have something fairly specific in mind here?
Secondly, I’m not certain what you’re referring to by the biggest problem of rolls, nor why the maths are bad. Can you elaborate?[/quote]
There should be some low cap of accuracy or a formula that doesn’t exponentially increase the “expected number of misses per hit” with increasing dodge.
D&D has 5% (natural 20), something like that would be useful here.
0 skill and 1 skill should differ by something like 5%-10%. That is, 0 skill character should have 40% chance to hit and 40% chance to dodge against 1 skill character. Not exactly that, just something along the lines.
Glancing blows by themselves aren’t necessarily a bad idea, but they blur the line between damage and accuracy too much if not done right. Having some excess dodge guarantee glancing blows from few first low accuracy attacks per turn is by itself OK, but linear scaling would turn it into “dodge is extra armor”. Instead some hard thresholds here would be more manageable: 50% damage for hits with less than 5 points of margin, for example. And for the ultra-high accuracy, the bonus should be crits, not damage scaling.
In your example, it would not allow more hits against dodgy targets, it would just turn the existing hits into glancing blows, making dodge even more overpowered.
Math is bad because of distribution of rolls in the form Xd10. Or more specifically, Xd10 - Yd10. For low values of X and Y, a single point of difference gives 20%-30% difference in chance of success. That is, character with 3 dodge (overall, not skill) has something near 80% chance of dodging an attack from a character with 2 accuracy. 2 points of difference result in ~90% chance of success and so on.
This does get better at very high values, but only because it doesn’t scale. That is, 1000d10 - 1001d10 is very close to 50%.
As a result, the expected behavior of dodging, which is dodging some hits but taking others, only happens when you’re at most 1 skill from opponent’s skill. You either dodge almost all or take almost all.
Good in-game example: skeletal dog. Those can shred no-skill, low-dex characters, but as soon as they gain 2 skill or so, skeletal dogs no longer pose any danger and can be disposed at minimal risk.
I rather like this idea, and you’ve obviously put alot of thought into it. I don’t have much to add beyond a general +1.
Alright, there’s some good stuff in there.
1: Accuracy chances should remain reasonable.
I would suggest that this doesn’t need to be something like “guaranteed automatic 5%”. The D&D idea that the ultimate swordsman having a 1 in 20 chance of whiffing every swing and the villager with a spear having a 5% chance of hitting a winged M1 Abrams are banal constructs that cause severe issues within the system without DM fiat.
2: Low level skills should not differ excessively from one another.
In a way, this is reality ensues. A person with no experience whatsoever will have more trouble against a person with a little experience than a person with middle experience will have against an expert. If you’re not much into martial arts, the same applies to any competitive game.
This said, while I agree that the difference should not be ridiculous, 0 Dodge should not be rolling 1D10 against 1 Dodge’s 2D10, that’s a 100% increase. Rolling 5D10 vs 6D10, however, is still the largest increase possible (20%), but does not represent an insurmountable hurdle.
3: Hard Thresholds are preferable to soft scaling.
Fine by me. We can work this pretty easily into any system after the basic “to hit” roll is agreed, though I did like the way that this handled damage randomisation for us.
4: In your example, it would not allow more hits against dodgy targets.
Are you sure you’ve not misread the example? The first given change is that Dodge is a D6 roll, while Accuracy remains D10. That change alone should drastically increase the number of hits occurring in the system. I’ll show some probabilities below.
From there, adding in scratch damage (which, from the rest of the thread, potentially causes injuries that slow them down) is still drastically more damage against dodgy targets, owing to more hits in general, while giving more value to low level dodge (which always has some value, even against high level accuracy, by diminishing the amount of damage dealt).
5: XD10 + YD10 averages are problematic…
We can resolve that by adding more dice to the base roll (4+X)D10 - (4+Y)D10. The more dice you have as your “base”, the less extreme low level skills can get without devaluing high level skills. We could move onto something like D100 - 3D[Dodge] +3D[Accuracy] or something if we need to, but that sort of highly luck based roll gets tedious.
So:
Goal 1: Any formula must offer a character with 4 Dexterity, 0 skills, and a +0 weapon a minimum 1% chance to connect with a target with 15 Dexterity and 10 skill.
Goal 2: Any difference between levels must be reasonable.
Goal 3: Thresholds, not scaling damage.
Goal 4 (My own goal): Dodge should serve a useful purpose against all opponents, without granting invincibility.
Let’s start by tweaking the number of dice getting rolled:
Dodge Roll = {5 + [Dexterity/2] + [(Dodge +1)/2]}D6
[spoiler]0 Dexterity, 0 Skill Traffic Accident rolls 5+0+0.5 = 6D6
4 Dexterity, 0 Skill Little Old Lady rolls 5+2+0.5 = 8D6
8 Dexterity, 0 Skill Nubcake rolls 5+4+0.5 = 10D6
11 Dexterity, 0 Skill Bystander rolls 5+5.5+0.5 = 11D6
15 Dexterity, 0 Skill Uberbaby rolls 5+7.5+0.5 = 13D6
0 Dexterity, 5 Skill Witness rolls 5+0+3 = 8D6
4 Dexterity, 5 Skill Vietnam Vet rolls 5+2+3 = 10D6
8 Dexterity, 5 Skill Sarge rolls 5+4+3 = 12D6
11 Dexterity, 5 Skill Arnie rolls 5+5.5+3 = 14D6
15 Dexterity, 5 Skill Chuck rolls 5+7.5+3 = 16D6
0 Dexterity, 10 Skill Grasshopper rolls 5+0+5.5 = 11D6
4 Dexterity, 10 Skill Shaolin Cripple rolls 5+2+5.5 = 13D6
8 Dexterity, 10 Skill Old Mentor rolls 5+4+5.5 = 15D6
11 Dexterity, 10 Skill Bruce Lee rolls 5+5.5+5.5 = 17D6
15 Dexterity, 10 Skill Albert Wesker rolls 5+7.5+5.5 = 18D6[/spoiler]
And the same for accuracy, except for D10s, instead.
With this, the probability of Traffic Accident landing a punch is:
Albert Wesker (18D6): 0.1%
Sarge (12D6): 15.23%
Nubcake (10D6): 39.03%
Another Traffic Accident (6D6): 91.89%.
The chance for Little Old Lady fighting Albert Wesker is: 3.65%. < That’s our Goal #1 met. Our lowest possible starting character (and very easy to achieve for a traffic accident with a stick) has a frankly unrealistic chance of landing a punch against a character which, let’s be honest, is probably only ever going to be the player character itself anyway.
Minimum number of dice that may be rolled: “5”.
So you can wear 99 layers, your dodge skill can be negative, and you have a -4 to hit weapon, and your minimum roll will be those five basic dice. Our theoretical extreme, Traffic Accident, wears 900 torso encumbrance. Their probability becomes:
Wesker (18D6): 0.0058%
Sarge (12D6): 4.29%
Nubcake (10D6): 17.24%
Unladen Traffic Accident (6D6): 78%
Goal 2: Any difference between levels must be reasonable:
Every hit roll in this system falls between these two extremes, so that’s a good start. We’re never going to have more than around 95% of hits by little old ladies missing expertly training demi-gods, 93% of hits by little old ladies hit other little old ladies, and little old ladies have the following chances:
]5D6 baseline: 99.9% chance of hitting
6D6 traffic accident: 99.3%.
7D6: 98.5%.
8D6 Little Old Ladies: 95% chance of hitting.
9D6: 89.3% of hitting.
10D6: 80.6%
11D6: 69.16% chance of hitting.
12D6: 55.9%
13D6: 42.3%
14D6: 29.88% chance of hitting.
15D6: 19.66%
16D6: 12%
17D6: 6.87%
18D6: 3.64%
19D6: 1.8%
20D6: 0.83%
Note that there’s a lot less outright dodging going on, particularly when you start including penalties for swarming, but this shows that at even the bottom of the scale, where the system is the swingiest, there’s still a manageable, logical progression, with diminishing rewards at both ends of the scale.
Goal 4: Dodge must serve a useful purpose.
All these changes to avoid immortal characters introduces a new potential problem here. Perhaps Dodge doesn’t do enough?
Wesker has a 99.9999997% chance of hitting Little Old Lady. He knocks her out.
She goes on a montage and tries again at each distinct skill level:
1 Skill - 5+2+1 = 8D6: No change.
2 Skill - 5+2+1.5 rounds up - 9D6: 99.9999996%
3 Skill - 5+2+2 - 9D6: 99.9999996%
4 Skill - 5+2+2.5 - 10D6: 99.99998%
5 Skill - 5+2+3 - 10D6: 99.99998%
6 Skill - 5+2+3.5 - 11D6: 99.9998%
7 Skill - 5+2+4 - 11D6: 99.9998%
8 Skill - 5+2+4.5 - 12D6: 99.9992%
9 Skill - 5+2+5 - 12D6: 99.9992%
10 Skill - 5+2+5.5 - 13D6: 99.998%
Extra half levels aren’t an issue. Half levels giving 0.5 means you can give yourself extra gear without penalty, or benefit more from a small stat bonus.
Clearly, however, there’s not much point in a binary hit/miss system. Unless your dodge stat exceeds your target, you won’t be dodging much.
Great!
That’s perfectly sensible. Dodging is hard, and the ability to no-sell anything is extremely, amazingly powerful. We can do better. Let’s check out:
Goal 3: Thresholds, not scaling damage.
The chance of rolling 50+ with 18D10 vs 5D6 baseline is 99.42%.
If we assume, for now, that 50 is our primary threshold for full damage (easily tweaked later), Wesker will pretty much always hit his target, and will always deal full damage.
Every 10 points above fifty, Critical Hit Stage +1
[spoiler]60: 95.75% chance of Critical Hit Stage 1.
70: 80.42% chance of Critical Hit Stage 2.
80: 56.18% chance of Critical Hit Stage 3.
90: 26.68% chance of Critical Hit Stage 4.
100: 7.98% chance of Critical Hit Stage 5.
110: 3.57% chance of Critical Hit Stage 6.
120: 0.125% chance of Critical Hit Stage 7.
So let’s check out our Little Old Lady again:
She has a 95% chance of being hit for full damage.
80% Crit#1
54% Crit#2
25% Crit#3
8% Crit#4
1.4% Crit#5.
0.145% Crit#6
Even the jump to 9D6 drops this to 91% chance of full damage, 72% C1, 44% C2, 18% C3, 5% C4, 0.74% C5.
Damage thresholds can do what you like. Call it 20% per 10 for example:
0-10: Deal 20% of base damage.
11-20: 40%
21-30: 60%
31-40: 80%
41-50: 100%
C1: 120% (And here you can start adding crit bonuses).
Our little old lady has a 3.64% chance of dealing 20% base damage to Wesker, 0.3% chance of dealing 40%. (little old ladies don’t get to hit Wesker for full damage, that would be unreasonable).
Our post-montage Lady has 13D10 and against Wesker, who is superhuman and just as skilled, gets:
46% chance of 20% damage (I forgot to mention above, this is “or better”, since I’m calculating this as probability of at least X).
19% chance of 40% damage.
5% chance of 60% damage.
0.67% chance of 80% damage.
0.04% chance of 100% damage or better.[/spoiler]
Using the “problematic” nature of dice results and thresholds together, Dodge, even small amounts of dodge, has a tangible effect, even at low levels, and vice versa. Our actors will generally be able to hit their targets, but if they are helplessly weak then they will be at a consistent disadvantage, as they should be.
Meanwhile our Skeleton Dog may not ever be an insurmountable horror, but if we say, for now, that monsters are [10+Dodge]D6, their 12D6 makes them consistently better against every opponent - being less likely to take higher amounts of damage even against our Wesker.
That’s Dodge Roll = {5 + [Dexterity/2] + [(Dodge +1)/2]}D6
Results:
Damage is linked to your ability to attack your enemy, ameliorated by your enemy’s ability to minimise harm to themselves.
Immortality is impossible without armour. Two absolute minimum 5D6 enemies attacking Wesker at the same time gives each enemy a 28.54% chance to hit for 20% damage, 3.59% for 40%.
All actors within defined parameters (8D6-18D6) maintain at least minimum baseline chance of success interacting with one another.
Dodge is valuable at all levels, but not overwhelming because it is no longer binary.
Characters close to each other in skill remain close to one another in ability.
All goals appear to be met. If you’d like to shoot up some examples for probability testing we can see how this would work and refine it further if need be.
We already have that included in XP ratios. Getting from level 3 to 4 takes more practice than from 0 to 3.
No need to represent that twice, especially considering that it wrecks granularity and makes accuracy vs. dodge math bad for gameplay.
The first given change is that Dodge is a D6 roll, while Accuracy remains D10.
OK, I missed that 6 part. I thought it’s both D10, like it is right now.
scratch damage
Could be interesting if we did scratching damage and glancing blows as slightly different. As in, different damage types having different glancing blow scaling - fire/electricity/acid having near 100%, cutting having something like 75%, stabbing 50% and bashing 25%. Or maybe the other way - glancing bashes dealing more than glancing cuts, since weak bashes still bruise, but weak cuts may hit with non-bladed parts or be pushed away.
The latter option sounds better, since best player weapons are cutting.
We can resolve that by adding more dice to the base roll (4+X)D10 - (4+Y)D10. The more dice you have as your "base", the less extreme low level skills can get without devaluing high level skills.
It’s still misapplication of the formula. I’d rather see it get more random instead. For example, rolling a set number of dice, but increasing the sides.
The problem with rolling lots of dice is that it has low variance and is heavily weighted towards the expected value. Variance of a sum of dice rolls grows linearly with number of dice, but quadratically with number of sides. Meaning that standard deviation grows linearly with sides, but only as square root of number of dice.
Another problem is that it doesn’t work at both edges - too little or too much causes it to freak out. And it both cases in a different way - too many dice causes diminishing returns, too little causes huge changes.
Those two may sound like good things, but they are actually hard to manage. And they are hard to manage because they involve non-trivial math that only looks trivial.
If we want diminishing returns, it would be better to represent them by lowering high accuracy/dodge values according to some simple function. ToME4 does it. I don’t really like the way it is done in ToME (makes changes feel meaningless if you aren’t very observant), but it works.
If we want weighting towards expected value and low variance, it would be better to use an explicit normal distribution and avoid having to pad the dice rolls to make it work for low numbers.
It may sound realistic or elegant to use multiple dice rolls, but this amount of realism and elegance is certainly not worth having to deal with sums of dice rolls, which are relatively hard to predict and thus hard to actually make realistic and elegant in different ways.
Also, about the formula making dexterity equal to dodge:
Dodge is already pretty good. It’s better than perception in most ways. It doesn’t need that buff. In current code, 2 points of dex give the same bonus as 1 point of dodge. 4 points of dex give 1 accuracy, like 2 points of melee or 3 points of weapon skill.
We already have that included in XP ratios. Getting from level 3 to 4 takes more practice than from 0 to 3.
No need to represent that twice, especially considering that it wrecks granularity and makes accuracy vs. dodge math bad for gameplay.[/quote]
I happen to disagree, since both are still true, it took very little time to master the basic footwork, movements and strikes. Doing so took extremely little time - a few training session - but still represented a significant difference over where I was before, and where I would be in a few more sessions when I started mastering the basic forms. This mirrors both how quickly a character increases and learns the basics and how much of a difference those basics make. Skill level 0: Doesn’t even know how to kick. Skill level 1: Knows basic striking and attacking stances that give power to a blow. Skill level 2: Knows basic blocks and forms.
If you want to keep it specifically and purely with levelling speed it can be possible, but while there’s no “need” to represent the difference between skill levels in terms of reward, it’s not something we need to worry about too much so long as it’s not too extreme.
Could be interesting if we did scratching damage and glancing blows as slightly different. As in, different damage types having different glancing blow scaling - fire/electricity/acid having near 100%, cutting having something like 75%, stabbing 50% and bashing 25%. Or maybe the other way - glancing bashes dealing more than glancing cuts, since weak bashes still bruise, but weak cuts may hit with non-bladed parts or be pushed away. The latter option sounds better, since best player weapons are cutting.
This is a great idea, and it would work excellently with the threshold damage results we’re looking to implement. Something liike a tazer dealing full damage whatever you roll and blades striking differently depending on your result could really bring some variety to the different types of weapon in the game.
Even two swords might be different, rapiers would generally use the point, rather than the blade, unless they scored particularly low, which cutlasses never thrust, but even scratches tend to deal more damage.
Another problem is that it doesn't work at both edges - too little or too much causes it to freak out. And it both cases in a different way - too many dice causes diminishing returns, too little causes huge changes.
Sure, these are definitely potential problems if you extend the count too far in either direction. In practise though, if everything works “within the boundaries we need”, any such downsides may as well not exist, because they wouldn’t exist within the game. This is why we have a minimum 5 dice, and possibly a maximum 20 dice.
We can bring the bands even tighter by reducing the value of dexterity (this also fixes the balance issue you pointed out, though we’ll probably need to do some legwork to give value to odd numbers again elsewhere). Dexterity / 3 means our range is between 6 dice Traffic Accident and 16 dice Wesker, which works out to make even our maximum possible difference between two characters a 0.7% chance of hitting.
It's still misapplication of the formula. I'd rather see it get more random instead. For example, rolling a set number of dice, but increasing the sides.
What a coincidence, I was working on this one before my last post:
D100 - 3D[Dodge] - [Dodge/2] + 3D[Accuracy] + [Accuracy/2]
Round up the Dodge/2 and Accuracy2, so 3 Dodge is “3D3-2”, for example.
It doesn’t normalise enough for my tastes, but let’s see:
This is with Dexterity/3 - if Perception needs a boost, we can tweak how it modifies damage thresholds later.
Aggression: -2 if you attacked in the last turn.
Dodge: Dexterity/3 + [(Dodge+1)/2]
Aim: Dexterity/3 +([Melee + Weapon Skill + 2]/4)
Round the
result.
Worse than baseline: “0” minimum. Weapon penalty, high encumbrance, and extremely terrible stats all at once.
Baseline: 0 anything: “1”
Joe: 8 stat/3 = 2.6, 5 skills = 3 = “6”
Wesker: 15 stat/3 = 5, 10 skills = 6: “11”
Two baselines or worse have a 100% chance of hitting each other, a 50% chance of rolling 50+.
Baseline will never effectively never dodge, but thresholds become more likely. I’ll give their chance of hitting a target with the skill and the chance of 50 or above when defending against those hits:
2: Attacking: 98.5% chance of hitting (49.5% chance of 50+). Defending: 50 = 52.5%
3: Attacking: 96% chance of hitting, (47% chance of 50+). Defending: 55%
4: Attacking: 94.5% chance of hitting (45.5%: 50+). Def: 56.5%.
11: Attacking: 80% chance of hitting, 31% chance of 50+. Def: 71%
Joe has a 98% chance of 1+ against another Joe.
Wesker has a 95.74 chance of rolling 1+ against another Wesker.
As you can see it’s all very low key and linear, but it’s pretty bleh at the moment, since Baseline 0 stat 0 skill shouldn’t be hitting Wesker 4/5 times.
I’m thinking it should be:
1D100 - 5D[dodge] - 2xDodge + 5D[accuracy] + 2xAccuracy.
That gives a ~5% chance for full damage Baseline vs. Wesker, 55% of hitting overall, and gives Wesker a 95% chance of dealing full damage to the baseline with decent chances for various stages of critical hit.
Meanwhile 0 vs 1: Att: 95.5% chance to hit, 45.5% of 50+, Def: 55.5% of 50+,
1 vs 2: Att: 95.5% chance to hit. 46.5% of 50+. Def: 99.9% chance to hit, 55.5% of 50+
All very stable and consistent. The higher you go, the more scope for two equal opponents to “get lucky” against one another.
Goal 1: Check, every character has a high overall chance of striking anyone else.
Goal 2: Check, Differences between levels are extremely smooth with this.
Goal 3: Check, Thresholds work fine with this system, since it’s effectively a percentile check.
Goal 4: Check, Since every level works out around 1.5-2.5% difference to the chances, every point of dodge and accuracy matters.
That sounds better.
I prefer the version with less divisions. Divisions are a problem since they introduce thresholds in stats, for example rewarding even dexterity values.
If we are going full D100, we could also make it about floating point values, to gain any arbitrarily big granularity.
[quote=“Coolthulhu, post:35, topic:12465”]That sounds better.
I prefer the version with less divisions. Divisions are a problem since they introduce thresholds in stats, for example rewarding even dexterity values.
If we are going full D100, we could also make it about floating point values, to gain any arbitrarily big granularity.[/quote]
Great, so we can use this as a rough basis to work from?
1D100 - 5D[dodge] - 2xDodge + 5D[accuracy] + 2xAccuracy.
Some refining:
Dodge = {(Dexterity/2*) + (Dodge Skill+1)/2 +/- Penalties and Bonuses.} / Attacks Since Last Turn (Not # of attackers, but attacks, you can be overwhelmed by rapid flurries from a single superfast enemy).
Bonus to Dodge: Waited Last Turn: Dodge +2
Penalty to Dodge: Attacked Last Turn: Dodge -2
Grabbed: Dodge -3 to grabber, -2 to everyone else.
Grabbing: Dodge -2 to everyone else, -1 to grabber.
*: Dexterity/2 does make Dexterity equal to points in dodge. This is potentially problematic, but means that “odd” dexterity is useful, since that extra 0.5 interacts with (Dodge Skill +1)/2.
Accuracy = {(Perception/2) + (Melee + Weapon Skill+2)/4 +/- Penalties and Bonuses.}
It is not a great feat of agility to swing a stick. You do not become more likely to hit your foe by being more graceful, you hit your foe by observing their movements and striking in the right location - Perception.
With this, Perception becomes a lot more useful, Dexterity becomes less of a Godstat, and we can get all of the stats involved:
Strength: Deal more raw damage, wield better, harder to block, weapons, have more hitpoints.
Dexterity: Dodge incoming damage, minimise damage taken, more easily recover from things like “tripping”.
Perception: Spot openings and opportunities to strike.
Intelligence: Know where best to strike. Increase the likelihood and severity of critical damage to vulnerable points. We can include this when he cover any threshold consideration.
Martial Arts:
Bonus to Dodge: “Moved Last Turn”.
Improvement to defending bonus.
Improvement of attacking penalty.
Dodge bonus against a specific target - Whether the last target to attack you or the last target you hit.
Improvement of either grabbed or grabbing penalty (Judo style)
Divide the number of incoming attacks: This could either be a flat reduction (take 1 from the number of attacks, minimum 1), a conditional reduction (for each distinct target you have hit successfully within the last X time, reduce the number of attacks by 1 - tiger style), or skill dependent (divide the number of attacks by your unarmed skill - Zul Quan).
[Stat] to Dodge/Accuracy (Sumo gives [(strength/2 + dexterity/4)] to dodge instead of Dex/3), and [Strength/2+Perception/4] to Accuracy. Strength is now key to the sumo fighting style.
Special:
“Block”: Happens after Reduce severity by one threshold and change the location to “Arm”, % chance based on unarmed skill modified by # of attacks and the power of the incoming blow.
“Leg Block”: Same thing, except leg.
“Weapon Block”: Same thing, except weapon, chance from weapon skill, damage dealt to weapon if it surpasses the weapon’s hardness.
“Parry/Redirect”: Soft style martial arts. Harder to accomplish than blocking, and with the same difficulties, but can both stop incoming force entirely without damage and put your opponent off balance, giving them dodge and/or aim penalties.
“X and hold”: Block or redirect a strike, then use that to grab your opponent or their weapon, either disarming or grabbing them.
Any other modifiers I might have missed, or suggestions to improve the system so far? Or should I move on to defining damage in this new, threshold based system?
well, fire electric and acid wont always have 100% certain things negate their properties. ANBC suits perfectly insulate from electrical attacks, things with NOMEX as a material would resist fire really well. rubber boots and such negating acid.
Would be better to get rid of divisions and just increase the granularity.
Making odd dexterity better instead of having even dexterity be better doesn’t fix anything.
Dividing dodge by attacks makes it horrible against fast opponents: if a critter has 2 attacks per turn, the second one would be significantly more likely to hit.
Dodge is like this at the moment. This makes it hard to balance, as it is incredibly useful against single enemies, but stops working as soon as more than one enemy comes near.
Would be better to make the per-attack penalty more linear. That the effects would be more stable and wouldn’t suffer so much from the “stub your toe and lose all will to fight” thing that plagues DDA.
There is also a problem with dodging weak critters and then having no dodges for the strong ones. But that’s relatively minor as special attacks don’t respect dodges left and melee attacks of even hulks and predators are relatively weak.
Additionally, there could be some difference between dexterity and skill here. For example, making dexterity not lose its value with extra attempted dodges.
Mixed feelings about moving accuracy to perception.
This would make melee fighters need to pick a specialization: attacks or dodges.
This in turn would mean that light armors would lose usefulness.
But zombies have bad dodging skills, so most characters would pick dodge anyway, then suffer when a fast dodgy enemy (wasp or manhack) attacked them.
Crits from intelligence would be weird.
It would spread melee combat rather thin - one stat for every aspect of it.
It would give intelligence a long-term application, but nerf “crit combat” (especially unarmed). Though it can be overpowered at times, as there are no resistances to status effects.
Though it doesn’t make much sense, as striking weak spots isn’t rocket surgery.
[quote=“Coolthulhu, post:38, topic:12465”]Would be better to get rid of divisions and just increase the granularity.
Making odd dexterity better instead of having even dexterity be better doesn’t fix anything.[/quote]
Sorry if I was unclear, but in the current system even and odd dexterity are useful.
Even dexterity starts out higher than odd dexterity compared to its investment. 8 dex = 4, 0 Dodge = 0.5, so Even Dexterity starts at “5”.
Odd dexterity starts out lower than even dexterity. 9 dex = 4.5. Dodge = 0.5, so Odd Dexterity costs more, but also starts at 5.
At the same time, odd Dexterity progresses earlier, so 1 Dodge skill makes them increase to 6 Dodge at level 1, while Even takes until level 2 to catch up. Finally, at level 10 dodge the system stabilises with both actors at 5.5+4/4.5 giving them 10 in a dodge.
This means that having an odd number in a stat makes you consistently ahead for the majority of the game, while even is constantly playing catch up until the very end, when the two equalise.
On the other hand I doubt it would break anything, so sure, let’s try no/div.
1D100 + {3D[Per + Aim] + [Per + Aim]} - {3D[Dex + Dodge] + [Dex + Dodge]}
Wesker always hits Trainwreck, and scores over 100 65% of the time.
Trainwreck hits Wesker 36% of the time, and scores over 50 1% of the time.
Trainwreck hits “Zombie Dog” (3D2+2) 93.5% of the time, 50+ 44.5% of the time.
Yep, formula still works fine (everything else must fall within these two extremes, and these extremes aren’t problematic), so let’s go with that if you prefer it. It might need bigger bonuses and penalties, but that’s hardly difficult.
Dividing dodge by attacks makes it horrible against fast opponents: if a critter has 2 attacks per turn, the second one would be significantly more likely to hit.
Well, yes. Just like in reality, if you’re fighting a big cat, you’re probably going to be busily keeping its jaws off your throat, while its claws are going to be raking your stomach. I could suggest “try not to fight big cats”, but that’s perfectly logical. Any animal with multiple attacks should be attacking with the main attack first, of course.
But let’s run the maths. I don’t think it’s actually that bad in our current system.
Player (skill 5, Dexterity 10) is fighting two skeleton dogs (skill 2).
First dog has a 15% chance of missing, 36% chance of 50+, 26% chance of 60+, 16% chance of 70+, 6% of 80+.
Second dog has a 5% chance of missing, 47% chance of 50+, 37% chance of 60+, 27% chance of 70+, 16% of 80+, 6% chance of 90+.
Because this system isn’t very threshold-y, the effect of halving their dodge isn’t very extreme, and it doesn’t get much worse:
Third dog has a 1% chance of missing, 50.6% chance of 50+.
Fourth dog has a 52.5% chance of 50+.
I think this may even be using the version of the formula that uses Dodge/2 + Dex/2, which is harsher on the player than the one above, but yeah, either way the example shows what I’m talking about.
Being swarmed or outsped still sucks. That’s good, since this is a zombie game, and zombie games are about large hordes of zombies tearing hopelessly outnumbered survivors apart, but, in effect, because this system is more linear, the results are more linear too.
It’s also pretty thematic for small obnoxious attackers to distract and interfere with a victim while a large enemy sets up to attack. Kind of makes me want to see a monster based around that, like a tentacle fiend whose jaw is surrounded by three tentacle creatures that do little damage but drag victims up to the hard hitting maw.
Additionally, there could be some difference between dexterity and skill here. For example, making dexterity not lose its value with extra attempted dodges.
I like this as a possible, but if I might suggest a bold and radical alternative, how about the other way around?
High Dexterity = Good against single targets because you’re more flexible.
That won’t help much against multiple targets, but knowing how to dodge so that your enemies get in each others way might.
Much easier to balance around stats that way, right?
Mixed feelings about moving accuracy to perception. This would make melee fighters need to pick a specialization: attacks or dodges. This in turn would mean that light armors would lose usefulness. But zombies have bad dodging skills, so most characters would pick dodge anyway, then suffer when a fast dodgy enemy (wasp or manhack) attacked them.
Remember that I’ve shown the absolute minimum in my examples here. If you want to check probabilities yourself to check me, try somewhere like http://anydice.com/.
Because I’ve shown you this is viable even if you have 0 Stat, 0 Skill fighting against a 15 Stat, 10 Skill superbeing, I can safely say your concern here is unnecessary.
Melee fighters do not need to pick a specialisation - the system doesn’t penalise low stats enough.
The difference against a manhack (20 Dodge) with a 4 Stat and 14 stat character with 0 skill is a 40% miss, 11% 50+ vs a 16% miss, 36% 50+. So 4 in ten vs 1.6 in ten, and 1.1 in ten vs 3.6 in ten.
So:
1: A player can specialise in dexterity and dodge.
2: A player can specialise in perception and accuracy.
3: A player can specialise in being able to deal and take a lot of damage.
1: Dodges more and takes less damage, but kills opponents more slowly, and therefore has more opportunities to get hurt.
2: Hits more often and kills enemies quicker so fewer opportunities to be attacked, but dodges less often and generally takes more damage.
3: Doesn’t dodge OR hit more often, but they can survive more hits, and kill in fewer hits,.
Tradeoffs. Each one wishes that they had something each of the others gets, and those who put 2 points into each stat will feel the benefits of generalisation.
Crits from intelligence would be weird. It would spread melee combat rather thin - one stat for every aspect of it. It would give intelligence a long-term application, but nerf "crit combat" (especially unarmed). Though it can be overpowered at times, as there are no resistances to status effects. Though it doesn't make much sense, as striking weak spots isn't rocket surgery.
Ever heard the phrase “fight smarter, not harder”? Intellect isn’t just about wearing spectacles and doing sums, it’s about quick thinking, coolness under pressure, split second analysis and knowing what a uvula is and when it’s best to stab it (always). Murder takes brains, as any cephalopod will tell you.
Striking weak spots may not be rocket surgery, but what’s the cliché archetype for the guy who’s hitting vital points? The Cunning Rogue. The smart guy who fights dirty. The coolheaded genius who’s just waiting for a single opening to finish the entire combat in a single, devastating strike. It’s even generally the smart guy in zombie movies who figures out that you’re supposed to shoot them in the head.
In the end, though, meh, that’s fluff. Nobody would really care or argue much, intelligence being tied to “clever” damage shows up in a lot of systems (including Cata already), so we can justify it easily if it makes for a better game, and to decide that, I’m of the opinion we’d need to establish what our damage system is up to first.
For now, let’s explore your threshold damage idea, since it works great, makes sense, is very easy to balance (just change your threshold) and ties in very neatly with the rest of the thread where we talked about damage types.
0-50: Scratch damage. You didn’t quite miss, but you didn’t get a solid hit. You deal a small amount of damage. At 10, 20, 30 and 40 you deal 20/40/60/80% damage, and have 20/40/60/80% your usual chance of X.
So every time you hit your target with:
Cutting weapon: Chance to cause bleeding if you deal damage.
Piercing weapon: Chance to bypass armour (decrease coverage by 50%). Chance to cause bleeding if you deal damage.
Bashing weapon: Chance to reverberate, dealing some of its damage through armour by striking a point without a lot of “give”,
Burning weapon: Chance to cause temporary blindness. Chance to cause a burn. Chance to ignite a bodypart.
This could vary by weapon, but let’s start out assuming a 1 in 5 chance. Intelligence adds 1% chance of it happening, and is divided or multiplied by the result as well.
Jack is using a rapier, which has a 20% chance of avoiding armour if it hits normally. He has 14 intelligence, giving him a 34% chance normally.
Jack is fighting a manhack. He rolls an 18. Minor damage: He deals 40% of his damage - “8”. The manhack’s metal armour can stop 10 damage, so it stops this entirely.
40% of 34% is 13.6%, however. Jack rolls a 90 and succeeds. Manhacks are a monster, which have 100% coverage, so Jack has a 50/50 chance of ignoring its armour entirely. Jack rolls a 78. Success! He bypasses the manhack’s armour and deals 8 damage to it directly. It takes all 8 damage, damaging it. Robots can’t bleed, however, so the manhack survives, smoking heavily (and potentially less mobile from the damage to its motor) as it loops around to retaliate.
Later, skill 0 Jack learns some skills. He adds his now 10 melee skill to chances as well, so his base chance becomes 44%, (and he deducts his piercing weapon skill from the base time per attack, so both skills play a role in making Jack a better swordsman beyond just damage and accuracy).
Now we can certainly still have critical hits and techniques on the upper end of the scale too:
Jack lands a critical hit (75+) on a Lobsterman, dealing 50% more damage, and increasing the chance of regular effects by 50%. It also gives a 10% base chance of a Critical Tech over and above the damage.
The same basic effects happen:
Bypass Armour: 50% bonus chance over 44% is 66%. Jack has a 1 in 3 chance of completely ignoring the lobserman’s absurdly thick armour.
Piercing Damage: Cause Minor Bleeding if they take damage: Again, 44% becomes 66%. If Jack deals any damage through the Lobsterman’s armour, it has a 2/3 chance of causing the lobsterman to bleed every turn. If the target was another human they would have the chance of causing an injury, depending how severe the damage was.
Apply Critical Technique: 10% damage modified by Jack’s crit bonus is 34%. 1 crit in 3, Jack may deal one of the Critical Techs he knows: Impale, Riposte, and Highlander Special.
So finally we have #4: A player can specialise in Intellect. They will not hit hard, they will not hit often, and their damage and survivability aren’t the best, but when they do hit they are more likely to cripple or maim their target, hindering their ability to fight back and killing them faster over time.
This gives four viable choices to melee characters.
We can’t quite give four viable choices to ranged characters in the same way, but we could potentially give then something to benefit from each stat as well:
Strength: More damage and range when firing bows. Throw items further and deal more damage.
Dexterity: Recover from being jostled faster, throw items more accurately, reduce the chance that incoming ranged damage is severe.
Perception: Aim non-throwing weapons more accurately (how flexible you are has very little to do with using a bow and arrow well, let alone a handgun).
Intelligence: Increase chance of dealing critical ranged damage against targets (lead the target better, pick your shots better, determine parabolic firing arcs better, all work together to hit more vulnerable spots, opposed by your targets ability to mislead you).
If that works, then there’d be less “this character is designed purely for melee”, and more “this character is good at certain things and bad at others”.
Thoughts?
Trainwreck hits Wesker 36% of the time, and scores over 50 1% of the time.
This means pure dodging is never a good idea and you always should wear light armor to stop the constant scratches.
This is already the case and it doesn’t work too well. It makes optimal play very uniform - you either wear a light armor and train some dodge or wear a medium armor and train some dodge.
Another problem with “linear scratches” is that we still need a different formula for special attacks.
Just like in reality, if you're fighting a big cat, you're probably going to be busily keeping its jaws off your throat, while its claws are going to be raking your stomach.
No. In reality multiple attacks tend to be one “compound attack”. If you’re fighting a big cat, you don’t dodge the jaws, you dodge the entire cat.
Because this system isn't very threshold-y, the effect of halving their dodge isn't very extreme
So half the dodge isn’t much worse than 100% of the dodge? That IS a problem.
Being swarmed or outsped still sucks. That's good, since this is a zombie game
It is already way too pronounced to make a good gameplay mechanic. It’s enough that being grabbed sucks hard and being in pain sucks even harder.
how about the other way around
I’d prefer dexterity allowing more dodges, but skill allowing better dodges. That way picking low dex at chargen would still allow good dodging, just not against groups. This would be easier to balance against hulks, predators, barfers etc., while high dex would allow dodging to complement armor instead of running out as soon as more than one zed gets near.
2: Hits more often and kills enemies quicker so fewer opportunities to be attacked, but dodges less often and generally takes more damage.
But zombies generally don’t dodge well.
3: Doesn't dodge OR hit more often, but they can survive more hits, and kill in fewer hits
HP damage to player generally doesn’t mean much. It’s the pain which kills.
You deal a small amount of damage. At 10, 20, 30 and 40 you deal 20/40/60/80% damage, and have 20/40/60/80% your usual chance of X.
No, I mean one hard threshold for scratch/glancing, everything below being a miss.