Monday 16 September 2013

Monitors: combat maths

Maths!

So far I've got one injury model I'm mostly looking at, with one backup. I need to look at the numbers for weaponry, to try and get them about as effective as I want. Fair warning, this is a fairly long and relatively dry post, so if you're looking for reptile jokes and shoehorned song references, I'd wait for another one.

Let's start with some assumptions.

The system needs to cover targets wearing anything from 'nothing' to heavy battle armour, with common targets including criminals (with ordinary clothing), pirates (lightly armoured), wild beasts (unarmoured) and military threats (moderate armour). In addition, many creatures will have some degree of innate armour from scaly skin, barklike flesh, thick blubber, being a robot and similar protection.

Weaponry needs to cover a range from bare-knuckle brawling, through primitive clubs and slings, focus heavily on futuristic handguns and blasters, and allow breathing space for lizard-portable heavy weaponry and vehicle-mounted hardpoints. While the setting features starship-mounted artillery and the like, I don't need a system to rigorously model what happens when you fire those at a person: I think we can guess.

Attack Roll, Wound Roll model

Attack Rolls

  • An attack roll involves a roll under the attacker's skill. The defender may roll to evade if in melee or if they spent an action for evasion on their previous turn. If they fail, they are affected by the attack.
  • A wounding model handles most physical attacks, while blinding damage is modelled with a blind die and slowing with a slow die. Restraint by ropes, webs and so on requires a Strength check to escape (with occasional allowance for Houdini stunts).
  • A wound requires a roll of 11+ on the d20, with weapon strength as a bonus and armour (or other defence, depending on the weapon) as a penalty. This number should be tweaked so that Monitors have around a 75% chance of wounding an average target - coupled with the need for skill rolls to hit, they should end up with about one wound per two actions, which is a single round.
  • An unsuccessful roll leaves the target pinned: they suffer a penalty until they spend an action to recover.
  • Monitor-grade characters have 3 wounds, civilians 1, and hulking alien monstrosities 5 or more.

We're looking at 1d20 rolls of 11+, with armour and weapon as opposing modifiers. If we want to allow some scope for any weapon to affect any target, then the greatest possible difference between armour (including hide) and weapon strength must be 9. At a pinch, we might allow heavy armour to provide complete protection from unarmed attacks and possibly from primitive weapons like clubs - not sure about this. Given the system is non-lethal anyway, I think allowing a mob of animals or civilians to overwhelm a character is probably okay. Moreover, because we have Pinning rules, we could allow very minor creatures to have a bigger differential than that, overcoming the classic Wizard/Cat phenomenon while allowing for particularly puny NPCs to, in fact, be temporarily disabled by a housecat.

Let's assume that armour 0 represents soft squishy creatures with no armour, 1 indicates thick civilian clothing or scales, 3 is basic protective gear, 5 is riot armour and 7-8 is military grade. Armour values above that are restricted to heavy infantry, vehicles and serious monstrosities and will be occasional occurrences rather than regular opponents.

See below for number-crunching. I couldn't easily get the table to look how I wanted; while the colours provide a broad indication, it's important to note that we're actually not aiming for the attractive green zone of 100% kill rates, but actually more of a fetching yellow.

Chance of causing a wound by modifiers

Weapon bonus
Armour penalty   0 1 2 3 4 5 6 7 8 9 10
0 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100%
-1 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95%
-2 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90%
-3 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85%
-4 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80%
-5 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75%
-6 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70%
-7 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65%
-8 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60%
-9 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55%
-10 0% 5% 10% 15% 20% 25% 30% 35% 40% 45% 50%

Analysis

Looking at this table, I think we're looking at a weapon strength of 7-8 for the go-to Monitor weapons if we want a ~70% wound rate on average targets. This will allow... okay, someone with Skill 4 (basic training only) will hit 20% of the time, and so cause a wound 15% of the time overall, which with two actions a turn means they're likely to wound someone every third turn; if they can get the drop on them or otherwise gain an advantage, though, that'll increase significantly (because accuracy is the main problem, with a +4 bonus to hit, every second shot is likely to cause a wound). A more military character who actually takes some weapons training (Skill 10) has only a 39% chance of failing to wound with two shots.

Of course, there will be other factors to consider. Some weapons should be weak against certain enemies and strong against others. Large weapons are unwieldy, but may be more powerful or simply affect more targets. Light weapons can easily be concealed, drawn and even used in a brawl. Some weapons are useless in the wrong circumstances, or have side-effects.

A problem here (in one sense at least) is that the numbers involved mean weapons will be heavily concentrated towards the top of the table, simply because of probability. With a minimum skill of 4 for Monitors, rolling under skill on 1d20 to hit and needing an 11+ on a d20 Wound roll, they have only a 10% chance of causing a Wound. This means that as soon as armour comes into the equation, weapons need to be at least equally good to offer a significant chance of success. A +4 overall modifier will increase the chance to 15%. Of course, one way to look at this is that the minimum skills are exactly that - representing a Monitor whose skills lie elsewhere, and who uses weapons as a fall-back. If they've upped their skill to 8, they're already succeeding 20% of the time when weapon strength equals armour, which tends to mean about a 35% Wound rate per round if they're just attacking. Still not huge, admittedly.

Sanity check

Pause a minute. Is this actually a problem? What success rate will be satisfying?

D&D is my go-to for combat success rates.

  • AD&D. A 1st-level fighter has THAC0 20 with a +1 specialisation bonus, and most likely no Strength bonus. Against a level-appropriate goblin (AC 6) this calls for a roll of 13, giving a 40% chance of goblin-brutalising. A wizard has a 35% chance. Either will likely kill the goblin in one hit.
  • D&D 3.5. A 1st-level fighter probably has Str 16 and a +1 BAB, for 1d20+4 overall. Against a level-appropriate goblin (AC 15) this gives exactly a 50% chance of hitting, and almost certainly killing it. A wizard is more likely to have Str 8 and a +0 BAB, for 1d20-1 overall, which gives a 25% chance of success.
  • D&D 4E. A 1st-level fighter probably has Str 18, 1/2 level (+0) and a +2 proficiency bonus , for 1d20+6 overall. Against a level-appropriate goblin (AC 16-17) this gives a 50%-55% chance of a hit. A minion will be killed instantly, but other types may easily have 30 hit points, and will survive three or four hits (at 1d8+4) on average. This means the fighter will take around eight attacks to kill the goblin! That's a very dramatic difference from earlier editions, and results in extended fights that I for one found tedious. I think three to four is more my speed for your thug-level creatures. Of course, fighters aren't the big damage-dealers... a striker will add extra damage, around 1d10+4 to 2d8+4, dropping that to two or three hits. A wizard functions very much like a fighter here so there's no real difference.

At this point, I'm inclined to think that it's not a huge problem if unoptimised characters have a low chance of success. I'd perhaps prefer a 20-25% level, simply because Monitors are supposed to be competent and it's annoying to constantly fail. On the downside, that's actually more or less impossible in the current model because they have only a 20% chance of hitting the target, and so would need a more or less 100% wound chance. The other problem here is that differentials have different effects at different skill levels: each +1 to weapon is roughly +1% for a skill 4 character, but a full +5% for a skill 20 character. Of course, I could simply up the skill level of Monitors. Or I could decide that if a professor of physics who passed basic weapons training ,but mostly investigates space-time anomalies and negotiates with local officials, can only drop a hostile moving target one time out of five, that isn't actually a problem. That decision is likely to come in a future post.

Obligatory table of wounding chances:

1d20 wound roll cumulative wound chances by target number and skill

Target number (as modified)
Skill 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
1 5% 5% 5% 4% 4% 4% 4% 3% 3% 3% 3% 2% 2% 2% 2% 1% 1% 1% 1% 0%
2 10% 10% 9% 9% 8% 8% 7% 7% 6% 6% 5% 5% 4% 4% 3% 3% 2% 2% 1% 0%
3 15% 14% 14% 13% 12% 11% 11% 10% 9% 8% 8% 7% 6% 5% 5% 4% 3% 2% 2% 1%
4 20% 19% 18% 17% 16% 15% 14% 13% 12% 11% 10% 9% 8% 7% 6% 5% 4% 3% 2% 1%
5 25% 24% 23% 21% 20% 19% 18% 16% 15% 14% 13% 11% 10% 9% 8% 6% 5% 4% 3% 1%
6 30% 29% 27% 26% 24% 23% 21% 20% 18% 17% 15% 14% 12% 11% 9% 8% 6% 5% 3% 2%
7 35% 33% 32% 30% 28% 26% 25% 23% 21% 19% 18% 16% 14% 12% 11% 9% 7% 5% 4% 2%
8 40% 38% 36% 34% 32% 30% 28% 26% 24% 22% 20% 18% 16% 14% 12% 10% 8% 6% 4% 2%
9 45% 43% 41% 38% 36% 34% 32% 29% 27% 25% 23% 20% 18% 16% 14% 11% 9% 7% 5% 2%
10 50% 48% 45% 43% 40% 38% 35% 33% 30% 28% 25% 23% 20% 18% 15% 13% 10% 8% 5% 3%
11 55% 52% 50% 47% 44% 41% 39% 36% 33% 30% 28% 25% 22% 19% 17% 14% 11% 8% 6% 3%
12 60% 57% 54% 51% 48% 45% 42% 39% 36% 33% 30% 27% 24% 21% 18% 15% 12% 9% 6% 3%
13 65% 62% 59% 55% 52% 49% 46% 42% 39% 36% 33% 29% 26% 23% 20% 16% 13% 10% 7% 3%
14 70% 67% 63% 60% 56% 53% 49% 46% 42% 39% 35% 32% 28% 25% 21% 18% 14% 11% 7% 4%
15 75% 71% 68% 64% 60% 56% 53% 49% 45% 41% 38% 34% 30% 26% 23% 19% 15% 11% 8% 4%
16 80% 76% 72% 68% 64% 60% 56% 52% 48% 44% 40% 36% 32% 28% 24% 20% 16% 12% 8% 4%
17 85% 81% 77% 72% 68% 64% 60% 55% 51% 47% 43% 38% 34% 30% 26% 21% 17% 13% 9% 4%
18 90% 86% 81% 77% 72% 68% 63% 59% 54% 50% 45% 41% 36% 32% 27% 23% 18% 14% 9% 4%
19 95% 90% 86% 81% 76% 71% 67% 62% 57% 52% 48% 43% 38% 33% 29% 24% 19% 14% 10% 5%
20 100% 95% 90% 85% 80% 75% 70% 65% 60% 55% 50% 45% 40% 35% 30% 25% 20% 15% 10% 5%

In case anyone needs to do anything similar...

The chance of wounding is calculated as (chance of rolling X or better on 1d20)*(chance of rolling Skill or less on 1d20).

For a one-die roll, the chance of rolling X or better is calculated as (1 - (X-1)*(1/sides)) and the chance of rolling Skill or less is (Skill*(1/sides))

For multi-die rolls, I produced a table of possible 2d6 rolls, and used a CountIf function to tot up occurrences of each roll. A secondary row then simply added up each number's probability and the previous cell's cumulative probability (zero in the first cell), to produce a cumulative probability.

Maybe I'll do another post just for this stuff...

Alternatives

Curved model

One possibility would be to switch damage rolls to a 2d10 roll or similar. What would that do?

With a 2d6 roll needing a 7 or better, a skill 4 character would need a +5 on weapons to get a 20% Wound rate. That being said, a mere +1 or +2 would allow around a 15% rate, which is better than the previous option. To hit 50%, you'd need to have about skill 12 and a +2 weapon.

Changing the die size doesn't improve matters - you need a bigger modifier with d10s. However, one thing it does do is create a bell curve. Is this a good thing? Well... if I'm not using variable damage then a bell curve is irrelevant, since we're modelling only change of overall success rather than degree of success, and a flat graph with cutoffs does that perfectly well. If I do want to use variable damage, then a bell curve would be worth investigating, but at the moment, no.

2d6 wound roll cumulative probabilities

Target number (as modified)
Skill 1d20 2 3 4 5 6 7 8 9 10 11 12
1 5% 5% 5% 4% 4% 3% 2% 1% 1% 0% 0%
2 10% 10% 9% 8% 7% 6% 4% 3% 2% 1% 0%
3 15% 15% 14% 13% 11% 9% 6% 4% 3% 1% 0%
4 20% 19% 18% 17% 14% 12% 8% 6% 3% 2% 1%
5 25% 24% 23% 21% 18% 15% 10% 7% 4% 2% 1%
6 30% 29% 28% 25% 22% 18% 13% 8% 5% 3% 1%
7 35% 34% 32% 29% 25% 20% 15% 10% 6% 3% 1%
8 40% 39% 37% 33% 29% 23% 17% 11% 7% 3% 1%
9 45% 44% 41% 38% 33% 26% 19% 13% 8% 4% 1%
10 50% 49% 46% 42% 36% 29% 21% 14% 8% 4% 1%
11 55% 53% 50% 46% 40% 32% 23% 15% 9% 5% 2%
12 60% 58% 55% 50% 43% 35% 25% 17% 10% 5% 2%
13 65% 63% 60% 54% 47% 38% 27% 18% 11% 5% 2%
14 70% 68% 64% 58% 51% 41% 29% 19% 12% 6% 2%
15 75% 73% 69% 63% 54% 44% 31% 21% 13% 6% 2%
16 80% 78% 73% 67% 58% 47% 33% 22% 13% 7% 2%
17 85% 83% 78% 71% 61% 50% 35% 24% 14% 7% 2%
18 90% 88% 83% 75% 65% 53% 38% 25% 15% 8% 3%
19 95% 92% 87% 79% 69% 55% 40% 26% 16% 8% 3%
20 100% 97% 92% 83% 72% 58% 42% 28% 17% 8% 3%

2d10 wound roll cumulative probabilities

Target number (as modified)
Skill 1d20 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
1 5% 5% 5% 5% 5% 4% 4% 4% 3% 3% 2% 2% 1% 1% 1% 1% 0% 0% 0%
2 10% 10% 10% 9% 9% 9% 8% 7% 6% 6% 5% 4% 3% 2% 2% 1% 1% 0% 0%
3 15% 15% 15% 14% 14% 13% 12% 11% 10% 8% 7% 5% 4% 3% 2% 2% 1% 0% 0%
4 20% 20% 19% 19% 18% 17% 16% 14% 13% 11% 9% 7% 6% 4% 3% 2% 1% 1% 0%
5 25% 25% 24% 24% 23% 21% 20% 18% 16% 14% 11% 9% 7% 5% 4% 3% 2% 1% 0%
6 30% 30% 29% 28% 27% 26% 24% 22% 19% 17% 14% 11% 8% 6% 5% 3% 2% 1% 0%
7 35% 35% 34% 33% 32% 30% 28% 25% 22% 19% 16% 13% 10% 7% 5% 4% 2% 1% 0%
8 40% 40% 39% 38% 36% 34% 32% 29% 26% 22% 18% 14% 11% 8% 6% 4% 2% 1% 0%
9 45% 45% 44% 42% 41% 38% 36% 32% 29% 25% 20% 16% 13% 9% 7% 5% 3% 1% 0%
10 50% 50% 49% 47% 45% 43% 40% 36% 32% 28% 23% 18% 14% 11% 8% 5% 3% 2% 1%
11 55% 54% 53% 52% 50% 47% 43% 40% 35% 30% 25% 20% 15% 12% 8% 6% 3% 2% 1%
12 60% 59% 58% 56% 54% 51% 47% 43% 38% 33% 27% 22% 17% 13% 9% 6% 4% 2% 1%
13 65% 64% 63% 61% 59% 55% 51% 47% 42% 36% 29% 23% 18% 14% 10% 7% 4% 2% 1%
14 70% 69% 68% 66% 63% 60% 55% 50% 45% 39% 32% 25% 20% 15% 11% 7% 4% 2% 1%
15 75% 74% 73% 71% 68% 64% 59% 54% 48% 41% 34% 27% 21% 16% 11% 8% 5% 2% 1%
16 80% 79% 78% 75% 72% 68% 63% 58% 51% 44% 36% 29% 22% 17% 12% 8% 5% 2% 1%
17 85% 84% 82% 80% 77% 72% 67% 61% 54% 47% 38% 31% 24% 18% 13% 9% 5% 3% 1%
18 90% 89% 87% 85% 81% 77% 71% 65% 58% 50% 41% 32% 25% 19% 14% 9% 5% 3% 1%
19 95% 94% 92% 89% 86% 81% 75% 68% 61% 52% 43% 34% 27% 20% 14% 10% 6% 3% 1%
20 100% 99% 97% 94% 90% 85% 79% 72% 64% 55% 45% 36% 28% 21% 15% 10% 6% 3% 1%

Bigger modifiers

Another possibility is to decide that armour and weapons can have substantially bigger modifiers. Maybe +5 is a weak weapon, +10 a normal one and +20 a heavy weapon.

With a normal +10 weapon and a +5 lightly armoured target, you'd have a... 15% chance again. A solid skill 12 shooter would have a 45% chance per shot, and a 16 skill sharpshooter a 60% chance. That's actually not too bad. Against an unarmoured target you'd automatically succeed. A heavy weapon would injure anyone without at least 11 points of armour. This might work okay.

Sharpshooting model

We could also allow some interplay between the hit roll and the wound roll. At the most basic level, if you roll half your skill or less, you get a +5 bonus on the wound roll. I'm keen to avoid odd little aspects of fixed-roll Critical Hit rules, such as the situation where all hits on a tough monster are critical, so let's see how this one might work out.

Under this model, our skill 4 professor needs a 4 to hit (a 9 against a straightforward target, or a 14 against an easy target). Normally she needs an 11 to wound, but half the time she hits she'll score a critical and need only a 6+. This looks a bit fiddly to model, but broadly speaking we should be able (I think) to average the values for a 4/11 and a 4/6 (slashes are successive rolls), which gives us a 13% in d20/d20 or a 15% in d20/2d10. For higher values, a skill 8 gets 25, a skill 12 34% and a skill 16 50% in d20/d20. 29%, 40% and 58% in 2d10.

I'm actually quite pleased with this one. I think we can work with this.

Percentage chance of wounding with +5 bonus for half-skill hit

Skill d20/d20 d20/2d10
1 3% 4%
2 6% 7%
3 9% 11%
4 13% 15%
5 16% 18%
6 19% 22%
7 22% 25%
8 25% 29%
9 28% 33%
10 31% 36%
11 34% 40%
12 38% 44%
13 41% 47%
14 44% 51%
15 47% 54%
16 50% 58%
17 53% 62%
18 56% 65%
19 59% 69%
20 63% 73%

Single Roll model

Single Success

The Single Success Roll model might be something like: 1d20 + AttackerSkill + weapon modifier - DefenderSkill - armour = damage. So 1d20 + 12 Ballistics + 2 (heavy blaster) - 8 Dodge - 4 armour = 1d20 + 2 damage (however that translates). My main objection here is it's fiddly maths! In practical terms, the player would roll 1d20 + skill + weapon = damage, and the GM would apply damage - defence - armour = result. So it's two separate lots of maths, but it's still something to bear in mind. I'm a bit tempted by this model, though. It's straightforward.

Let's see how this version might play out, adapting it for the current wound model so the overspill from the attack roll becomes the Wound roll.

1d20 + AttackerSkill + weapon modifier - DefenderSkill - armour = Wound roll

Assuming our professor's weapon and armour cancel out, the average will be 10.5 + 4 + 0 - 4 - 0, giving a score of 10. For an equally puny opponent (DS4), that's about a 50% chance of a successful Wound. More likely, the enemy will have a bit of an edge here, and the odds will drop somewhat. A skilled character with skill 10 has around an 80% chance to take down an ordinary criminal (DS4), and around a 50% chance to drop a mercenary (DS10). A real sharpshooter with 15-20 skill can drop a mercenary (DS10) with military armour (+5) one shot out of two and drop henchmen like flies.

Another version would scrap DefenderSkill (for simplicity), and would allow a wider range of armour values to compensate. I think basically this would work either way around, so we can decide later whether or not defensive attributes are a good idea.

Here there are a lot of factors at play in a single calculation, and so it's a bit tricky to decide how they interact. The relative frequency of various armour types will make a significant difference to success over time. Monitor weaponry is likely to make less difference, as PCs tend to broadly stick with the gear they have except where there's a level-based upgrade model.

There's not a huge amount of maths I can usefully do here. Broadly speaking, though, I'm slightly less happy about it for some reason. Not sure why. Any ideas?


That really is an awful lot of stuff (and it takes freakin' ages to do those tables if you're fussy like me and insist on peeling off all the revolting useless bits of code that importing tables adds, and making the cell/row/table structure halfway respectable) so I'll leave it there for now. I still need to think about what sort of narrative results I'd want from various situations, in order to try and juggle the numbers appropriately.

4 comments:

  1. Tricky one.

    FWIW I *do* favour one-roll mechanics. The maths can get a bit dirty though.

    Really simple kludge - since Wounds are in such short supply, make Armour add Wounds directly. You get the same hits-to-kill effect, but lose a chunk of dice rolling.

    Bigger weapons could then just do more wounds (which can easily be 1/2/3 for Light/Med/Heavy).

    Another option might be to swap the wound roll for a WH40K-style Armour save, with bigger weapons simply negating bigger saves. Or keep the Wound roll, but instead of 1D20+Damage-Defence make it 1D20 vs Defence, with bigger weapons providing auto-success. This eliminates *some* of the excess rolling at least.

    ReplyDelete
    Replies
    1. Interesting. There are potentially some issues with armour-wounds feeling weird in play, I think, unless you go for ablative armour that actually degrades. Otherwise it could be fun.

      I've been thinking about allowing high damage rolls to inflict extra wounds (which under current models would mean heavy weapons have more wounding capability in general).

      The armour save one seems quite neat. I quite like the idea (which I think I've floated elsewhere) that being hit by weapons is something that just flat-out hurts you unless you can block the hit, so an adjustable Armour save could work well in place of a Wound roll. Maybe weapons have a modifier to saves, while particularly puny ones like bites from tiny animals actually give a bonus and can't realistically hurt anything that isn't stark naked. Again, a really bad save result could result in additional wounds.

      Okay, clearly I need to play with this. Thanks!

      Delete
  2. Mathematically, of course, save rolls are identical to wound rolls.

    The advantage of auto-penetration is that it speeds things up.

    I actually think armour-as-wounds could work in this setting, because your tech is very much *part* of your character. In a sense it's no weirder than hitpoints (then again, you aren't *using* hit points so... yeah).

    ReplyDelete
    Replies
    1. So there's a couple of reasons I hesitated. One is that yeah, I'm not using hitpoints. The other thing is that I feel like as a player, I'd find it unintuitive that my armour soaks up the first hit from a weapon and is afterwards useless. That would work fine for forcefields or medical nanobots, mind! But I'd expect armour to protect me continuously in an unpredictable way, rather than predictably against a specific number of hits. Although, I suppose some high-tech kinds could be skinned with all that "Warning! Armour status 24% and falling" sort of thing. But that doesn't help with things like thick scales.

      OTOH if I was using a random hit location model for wounds, instead of the cumulative status boxes I've got now, armour could add more boxes to that. Possibilities, possibilities... seems a bit fiddly though TBH.

      Delete