Aggregate Attacks: How Should Volley Work?

[ModelCitizen]

PC-vs-army fights are cool and some PCs want to roll around with peasant militias or skeleton archer platoons, but nobody wants to roll 20 individual arrow attacks per turn.We've been talking about handling hordes of low-level mooks by having their individual attacks be worthless against mid-level targets but letting them gang up into one big attack. Think Aid Another in 3e, but a little more sophisticated and without the check vs AC 10.

Aggregate attacks are easy to write for melee because of positioning limits. You can't get more than 24 dudes on a single target and most of the time no more than 8 dudes, so the rule only needs to work for 2 to about 8 participants:

Dogpile: A bunch of dudes all whale on the same guy. One dude makes an attack roll (controller's choice which one), with a +1 bonus to attack and damage per additional dude. The total bonus of the attack can't exceed the attack bonus of the lowest participant +12.

A phalanx or a literal dogpile makes one attack, and in the unlikely scenario where 8 melee and 16 pikemen all surround the same target they still only make two or three. You might need another rule to keep them from making individual initiative rolls, but so far so good.


The problem is archers.

Archer Problem #1: A completely retarded number of archers can fire on the same point. I've personally seen a PC order 200 goblin archers to fire on the same gatehouse. Archer aggregate attacks need to work for 8 dudes, but they also need to work for 200 dudes. Obviously you don't want to roll an arrow attack at 1d20+203 for 1d6+200 damage. You could have large numbers of archers create an AoE damage zone that gets less than +1 per participant, but that still leaves problem #2:

Archer Problem #2: Every participant has to matter. If every 10 archers adds +1 to the volley, 59 archers make one volley and nine crit-fish attacks. You could let the 9 archers in the remainder use the melee rule and Dogpile, but then you have to worry that the 50 archers will split up into 5 Dogpiles themselves. Each dogpile has to matter in a mid-level encounter, which would mean a 50 archer Volley would have to be better than five weak but level-appropriate attacks. That sounds awful.


So, how do you solve this without resorting to DM handwaving?

[tussock]

IRL, archery plastered a distant area somewhat larger than the area the archers stood in. In D&D terms it's basically blind-firing (50% of arrows miss strait off) at +0 vs everyone in the area, assuming the archers are in a nice tight pack.

Which, if the PCs/monsters have AC > 20, you just ignore. They're immune to volley fire. All good.

At short ranges for direct fire, or to resolve lots of +0 attacks at range, work out how many hit on average and vary it a little, +1d6 -1d6 works well statistically for ~20 rolls. Average hits per 20 attackers are Attack Bonus +20 - AC (limit 0..20, or 1..19 if you give them auto-hits and misses).

For +1d6 -1d6, have the attacker add one and the defender subtract one. Use d4's for less attackers, d8's or more for very large numbers at once. Average damage works well enough after that, round down.

There should probably be some limit to the number of arrows raining on one square in six seconds, no more than 100 or so, and only for short-range direct fire attacks.

[Username17]

Having archery volleys attack larger areas rather than specifically getting larger bonuses against individuals is probably the way to go. You put some limit to the number of archers that can do one square, and further archers attack other squares.

The big key is that you want to incentivize the hordes to make singular volley attacks rather than have them make separate attacks looking for 20s. I think that implies the following:

• Nat 20s do not auto-hit.
• The bonuses for participating in volleys are larger than you can get for splitting up into aiding another and firing partners or other fiddly individual actions.
• Volleys can't be repeated for credit at the same square in the same turn. So no figuring out the magic magic breakpoint where you'd rather attack again than have a bigger volley and then doing that.
• Expanding the volley into more squars requies less extra archers than starting a new volley in the new squares.

That being said, I suspect that enemies would still routinely split themselves up into two volleys, one which fired on your position before your move and anoher that fired on your position after your move. I think that's acceptable.

-Username17

[MfA]

What about DR?

I'd say let a computer handle it

[Josh_Kablack]

Aggregate attacks are easy to write for melee because of positioning limits. You can't get more than 24 dudes on a single target and most of the time no more than 8 dudes, so the rule only needs to work for 2 to about 8 participants:

Smart Alec Edge Cases:
http://www.d20srd.org/srd/monsters/roper.htm
http://www.d20srd.org/srd/monsters/squidGiant.htm

I bring it up because in a fantasy RPG, you are going to have creatures with greater than 10' reach.
While it's unlikely to face a tendirculous formation, it's entirely possible that those Pikemen get a Mass Enlarge Person cast on them at the start of the fight, letting them reach out to 20' with their polearms.

Dogpile: A bunch of dudes all whale on the same guy. One dude makes an attack roll (controller's choice which one), with a +1 bonus to attack and damage per additional dude. The total bonus of the attack can't exceed the attack bonus of the lowest participant +12.

Why +12?

Why base it off the lowest attack bonus? Is the intent to make Ray of Enfeeblement a whole-unit debuff that splits victims off to crit-fish.

A phalanx or a literal dogpile makes one attack, and in the unlikely scenario where 8 melee and 16 pikemen all surround the same target they still only make two or three. You might need another rule to keep them from making individual initiative rolls, but so far so good.

If you're talking about using this as a 3e houserule, not so good.

Provided that members of the group have at least a +1 attack bonus, then using Aid Another is going to work more than half the time for an average of +1.1 or better to hit per additional ally in melee reach with a max of +8 to hit. Compared to +1 per ally in melee reach with a max of +12, Aid Another is a better deal save for in cases where you have more than 8 allies. And in some of those cases it may still be better to split into groups to get the better bonus.

The fundamental problem is that whatever sort of volley attack rule you implement has to co-exist with any rules for crit-fishing and aiding another in a way that actually encourages groups of mooks to make volley attacks instead of using the other options. Otherwise it's tactically optimal to ignore volley attacks and keep on using the lengthier to resolve options.

Your rule is only worth considering if the 3.x Aid Another rule isn't in play.

And that's the easy one to evaluate, you're going to end up with cases where crit-fishing volleys are still mathematically superior unless you outright disallow them or ditch natural 20 auto-hits or match the odds via giving large enough volley-fire autohit/critical threat abilities or something.

[MGuy]

I'd have to ask if you want armies worth of people to "matter" to PCs after a certain level. As it stands in 3rd after a certain level mobs of regular people just can't do anything against higher level creatures. So to clarify how long do you want sending wave after wave of your own men at a problem to be a viable tactic against PCs?

[Josh_Kablack]

The weird thing about 3e is that it's expected to hit the "no matter how many of these dudes you send against one high level dude, it's never enough" breakpoint TWICE in the course of the supported level ups.

As per the CR and encounter tables, no number of CR 1 opponents is a significant challenge for a 9th or higher level character....and supposedly no number of 9th level opponents is a significant challenge for a 17th or higher level character. And well, that's just really hard for folks to wrap their heads around.

[Prak]

Well, another way to go would be to make volleys create a hazardous area, rather than be specific attacks. Arrow volleys were really closer to cover fire, designed to rain hell over an area. So what if archers basically created difficult terrain? The archers/controller select an area, like 1 square/10 archers, and anyone in the area has to make a save versus moving at reduced speed and taking some standard amount of damage.

[Grek]

Here's the rules you'd use if you just wanted to simulate the results without rolling dice.

For groups of archers less than 20, use the Dogpile rules as normal. For groups of archers greater than or equal to 20, use the following: Any group of 20 archers who can all fire at a target can come together to fire a Volley, which inflicts one die of weapon damage per point of average archer attack bonus to their target. The archers may instead opt to fire at a 10' radius, in which case the damage is divided evenly between everyone in the area volleyed at. Reduce this damage by half if the target(s) have partial cover and ignore it entirely if the targets have total cover vs. the archers. If more than three volleys have hit a given area, it becomes difficult terrain until the arrows are removed.

[ModelCitizen]

Dogpile: A bunch of dudes all whale on the same guy. One dude makes an attack roll (controller's choice which one), with a +1 bonus to attack and damage per additional dude. The total bonus of the attack can't exceed the attack bonus of the lowest participant +12.

Why +12?

Why base it off the lowest attack bonus? Is the intent to make Ray of Enfeeblement a whole-unit debuff that splits victims off to crit-fish.

Think of a mid-level fighter with a squad of retainers. The intent is to get the retainers to act separately from the fighter, rather than the fighter leading the dogpile himself for a huge attack bonus. And the reason I picked +12 and not +7 or something is to account for creatures with slightly different stats, so that little attack bonus differences like Ray of Enfeeblement or not having dwarf +1 vs goblinoids are less likely to force one guy to break off and critfish.

Regardless, it was just an example for some vague hypothetical D&D clone. It's very possible you could demonstrate that I picked bad numbers if this D&D clone existed outside of my imagination.

If you're talking about using this as a 3e houserule, not so good.
...
Your rule is only worth considering if the 3.x Aid Another rule isn't in play.

Yeah, I know. I wasn't thinking about it as a 3e house rule or coexisting with Aid Another.

I'd have to ask if you want armies worth of people to "matter" to PCs after a certain level. As it stands in 3rd after a certain level mobs of regular people just can't do anything against higher level creatures. So to clarify how long do you want sending wave after wave of your own men at a problem to be a viable tactic against PCs?

I'm thinking about level 5. A level 5 wizard has access to Fly, Fireball, and Wind Wall (or Protection from Normal Missiles if you're into that sort of thing). A 5th level cleric can walk into a graveyard and come out with 20 skeletons. Pretty much any PC past about level 3 can hire a bunch of mercs, because in the adventurer economy hired soldiers work for chump change. It's not quite "mass combat," but even at what I consider starting level D&D expects more mooks on the field than I want to roll dice for.

[Duke Flauros]

For simplicity's sake, you could say that each 10 archers adds a five- foot square to the volley, which has a reflex save of dc 10+ archer's attack bonus; on a successful save, you are struck by 1d3 arrows (or none at all with evasion), and on a failed save, you are struck by 1d10 arrows.

[ModelCitizen]

Alright, it looks like no one can think of a way to get each individual one of 200 archers to want to participate in a volley rather than critfish. I mean, I can't either, that's why I asked. A bigger AoE is something, but it's not necessarily better. A lot of the time they're only going to be trying to hit one wizard.

Looks like the way forward is to have schlub archers fall off the RNG fairly early, and not let them auto-hit on 20s. This would need to happen by low-mid levels (i.e. level 1 goblin archer hits level 5 PC on a 19 and auto-misses level 7 PC). Assuming a new game with new core math, do you guys think that would be too harsh of a design constraint?

[MGuy]

Well if you're just straight up making a new game then the constraints can be whatever you want. What a "Level" means is entirely dependent upon your system. If you want volleys of archers to not matter a lot to higher level characters but you still want them to be able to hit them you just have to adjust your numbers to make it fit and get rid of rampant aid another actions and auto-hit crits.

[Grek]

Alright, it looks like no one can think of a way to get each individual one of 200 archers to want to participate in a volley rather than critfish.

Did you completely miss my post or something?

[ModelCitizen]

Alright, it looks like no one can think of a way to get each individual one of 200 archers to want to participate in a volley rather than critfish.

Did you completely miss my post or something?

Nah, I read it. It's alright, but it still has the the thing where (num_archers % 20) can't contribute to a volley and have to break off to do something different. I'm trying to figure out if there's a way to avoid that.

[Grek]

Alright, it looks like no one can think of a way to get each individual one of 200 archers to want to participate in a volley rather than critfish.

Did you completely miss my post or something?

Nah, I read it. It's alright, but it still has the the thing where (num_archers % 20) can't contribute to a volley and have to break off to do something different. I'm trying to figure out if there's a way to avoid that.

The spare 1-19 archers do the dogpile rules. They still come together, they just use a slightly different abstraction.

[RobG]

It sounds like you need to come at this backwards, result first. Figure out how much damage you think your PC should take if he's fired on by 20,100 or 1000 archers. Use your result to figure out a system to reflect that.

[tussock]

Real medieval knights walked slowly through arrow storms from longbows where not one man dropped (though it was noted to hurt like fuck and scare the shit out of them all, to paraphrase). The trick was to look down at all times and walked hunched at just the right angle. The near-eastern armies were perplexed by the western knights inane tactics of weighing themselves and their horses down with so much armour than they couldn't catch you, but noted it also made them invulnerable to archery.

Heh, as the commander of the Byzantines noted in a report home, the French knights had no tactics, no sensible divisions, fell for every trick in the book, and paid no heed to numbers, terrain, or weather. All they did, over and over, was line up, salute, and charge en masse with their heavy horse. The shame of it was it was working.

When the Welsh Longbow dominated such battles it was thousands of them concentrating fire on smaller forces from behind prepared defences, and every single time the contest was decided by the resulting melee (where those who'd walked through the storm, it being dangerous and highly disruptive to horses, tended to be rather disorganised). At Crecy the longbowmen decided the fight by charging both flanks of the disorganised French with axe and maul, not with the bow (and the vast majority in contact surrendered anyway).

Bowmen caught without defensive preparations were destroyed by direct cavalry charges, and routed by any serious foot formation. Swords, spears, bills, and so on kill people dead as shit, arrows mostly do not, though mostly because the goal in a melee was to drive forward over all the fallen so your back ranks could kill the enemy and save their fallen neighbours. Skeletons uncovered on medieval fields of battle usually have multiple massive skull and face fractures, heads pounded until they split wide open, and you can't do that with an arrow.

Figure out how much damage you think your PC should take if he's fired on by 20,100 or 1000 archers.

Yeh, realistically, fuck all, assuming you're covered in anything like decent armour, and you remember to cover your eyes.

[sabs]

Even the most famous battle of Longbow vs Knight:

THe Battle of Agincourt. The arrows killed very few knights. The horses panicking and throwing knights, and stampeding through the advancing Heavy footmen killed more than the arrows. The hunched forward slog through the mud fatigued more men, than the arrows killed. The french men at arms, got to the English line so exhausted, that the English pushed them over, and they couldn't get back up.

Your problem is that D&D doesn't model certain things:
Hit locations
Armour weakpoints
stamina damage

A D&D knight doesn't have to talk about having his visor down, and his head down to avoid arrows to the face. A D&D knight doesn't take fatigue damage from slogging through the mud and arrow strewn ground wearing 60lbs of armor in the suffocating confines of his metal helmet with the visor down.

Tussock is both right and wrong. A man in full plate isn't going to get pierced by arrows. But he is going to take 'damage' in the form of exhaustion, and lack of oxygen etc.

[Ikeren]

Usually when I'm dealing with high numbers of volleys against PC's, the attack bonus is pretty low. Say I've got 100 goblin archers with attacks of +5 against a bunch of PC's with AC 20, 21, 23 and 25.

I roll 25d20, filter min15, min16, min18 and min20.

Then for the list of rolls I get up, I then roll damages; crits I just roll the damage twice.

For example:

http://www.wizards.com/dnd/dice/dice.htm
19,6,3,18,16,13,20,2,3,11,17,17,1,11,5,19,16,14,7,13,2,13,1,17,4,+0
Total:268

That's 10 damages on min15.
Roll 10 damage: 3,3,1,6,2,2,5,3,1,3,+0
Total:29
(add +10, if they all have +1)


DR makes you have to filter the damages on a min number, but doing this by eyesight for groups smaller than 200 takes no more than 3-4 minutes. Usually I'm faster than PC's on their turns. If I miss one or two, w/e.

This is indeed a crit-fishing style. But there is nothing wrong with crit-fishing in my books; worked on Smaug.

[Koumei]

worked on Smaug.

No it didn't. A whole bunch of people pinged arrows at him for no effect, making a futile gesture because it looks better than sitting down and saying "Well, we're fucked". Then one skilled archer made a called shot based on the secret information about Smaug's weakpoint using a special named arrow.

[CCarter]

I've mentioned this before on the board (ages ago, though) but it is possible to compress a large number of d20 attack rolls into a single d100 roll, if you're willing to use a table. For a given number of attacks (say 10 or 20) and a target number you can work out in advance the likelihood of getting N successful attacks.
Someone did that in Dragon #113: it has tables that cross-index the target number with # attacks and lists the equivalent d100 result you need to get that many, so you can tell how many arrows hit or whatever.
For instance looking at it now there's a table for doing sets of 20 rolls; if an archer needs a 15+ to hit on the raw die roll (e.g. the archers get +10 to hit, and the target AC was 25), and you roll an 88 on d100, that would mean 8 successful attacks.

Within a 1% margin of error, its identical to just rolling the 20d20 and counting the hits, although it doesn't do criticals - you get how many rolls were successful, but it can't tell you what those rolls were i.e. how many were 20s.

[Hicks]

Well, statistically speaking 30% of any number of attack rolls that hit on a 15+ will successfully strike their target, 1 in 20 total attacks will have rolled a "20", and of those 30% of those attacks would be a confirmed critical hit, which works out to 1.5% of all attack rolls score a critical hit if an archer hits on a 15+.

In D&D 3.5 that would be:
TOTAL ATTACKS*0.285*(4.5+DAMAGE BONUS)+TOTAL ATTACKS*0.015*3*(4.5+DAMAGE BONUS)