Iterative Attacks

[Apalala]

Saw this in Trailblazer, basically a simpler and yet more radical alternative to pathfinder. They, rightly so, determined that iterative attacks were bunk. 0/-5 isn't so bad, but 0/-5/-10 is pretty dumb, and 0/-5/-10/-15 is just ridiculous. You have to keep track of different sets of modifiers, and you roll them one at a time, and most of the time that -10 and -15 are absolutely useless, but you'll never, ever just give them up.

So, they suggested an alternative. At 6 BAB, you can full attack for -2/-2. At 11 BAB, you full attack for -1/-1. And at 16, it's 0/0. One modifier to keep track of. Two less attacks to roll. And assuming you aren't hitting the enemy on a two, it actually BOOSTS damage output by about 15%.

It's so elegant...!

So, what does everyone else think of this solution?

[Ashtagon]

I found something even simpler.

Roll a number of d20s equal to the number of iterative attacks you would have, with attack modifiers as if each were you first attack. Take the best roll, and discard all other rolls. If that succeeds, you made a single successful attack.

Surprisingly, the maths of hit probability works out about the same as with the RAW iterative attacks.

[Hyooz]

Yay rogue nerf.

[Ashtagon]

Not a rogue nerf. If anything, it makes the rogue less swingy.

[FlamingKobold]

Definitely a rogue nerf. That said, I think that, in the same period of time, A fighter should be able to get off more attacks than a wizard, not just have a higher chance to connect.

[Ashtagon]

ok, let's take two hypothetical situations...

a) You make two attack rolls; each requires a 16+ to score a hit.

b) You make a single attack roll; it requires an 11+ to score a hit.

c) You make two attack rolls and take the best roll, discarding the other; that roll requires a 15+ to score a hit.

(Obviously, the actual maths of iteratives is more complicated than this. I'm using a deliberately simple example to help make the maths understandable).

Under a), you have a 56.25% chance of no hits, a 37.5% chance of one hit, and a 6.25% chance of two hits. The average number of hits per round is 0.5 hits.

Under b), you have a 50% chance of no hits, and a 50% chance of one hit. The average number of hits per round is 0.5 hits.

Under c) you have a 49% chance of no hits, and a 51% chance of one hit.

a) illustrates the situation with iterative attacks; b) illustrates the situation with Trailblazer's iterative variation; c) illustrates the situation with my variation on iterative attacks.

In all three cases (and yes, I did artificially choose numbers to illustrate my specific point; the actual numbers from the real maths is much more complicated, but gives approximately similar results), the average number of hits per round is exactly the same.

The two variant systems (b and c) have an advantage in that your character can more reliably expect to do damage on his attack sequence.



I am sincerely interested in why you consider this a nerf. About the only case where I can see it as a nerf is where an opponent won't be dropped by one hit, but would be dropped by two hits. But the odds of getting multiple hits in an iterative attack was never all that great to begin with (6.25% in my example).

As for 'getting off more attacks', I always assumed that characters are swinging their weapons multiple times in a round. Six seconds is an awful long time in a fight situation after all. Under the RAW system, the attack rolls merely represent those swings that had the potential to connect. The two variant systems simply change some assumptions about what the rolls represent.

[Fiery Diamond]

ok, let's take two hypothetical situations...

a) You make two attack rolls; each requires a 16+ to score a hit.

b) You make a single attack roll; it requires an 11+ to score a hit.

c) You make two attack rolls and take the best roll, discarding the other; that roll requires a 15+ to score a hit.

(Obviously, the actual maths of iteratives is more complicated than this. I'm using a deliberately simple example to help make the maths understandable).

Under a), you have a 56.25% chance of no hits, a 37.5% chance of one hit, and a 6.25% chance of two hits. The average number of hits per round is 0.5 hits.

Under b), you have a 50% chance of no hits, and a 50% chance of one hit. The average number of hits per round is 0.5 hits.

Under c) you have a 49% chance of no hits, and a 51% chance of one hit.

a) illustrates the situation with iterative attacks; b) illustrates the situation with Trailblazer's iterative variation; c) illustrates the situation with my variation on iterative attacks.

In all three cases (and yes, I did artificially choose numbers to illustrate my specific point; the actual numbers from the real maths is much more complicated, but gives approximately similar results), the average number of hits per round is exactly the same.

The two variant systems (b and c) have an advantage in that your character can more reliably expect to do damage on his attack sequence.



I am sincerely interested in why you consider this a nerf. About the only case where I can see it as a nerf is where an opponent won't be dropped by one hit, but would be dropped by two hits. But the odds of getting multiple hits in an iterative attack was never all that great to begin with (6.25% in my example).

As for 'getting off more attacks', I always assumed that characters are swinging their weapons multiple times in a round. Six seconds is an awful long time in a fight situation after all. Under the RAW system, the attack rolls merely represent those swings that had the potential to connect. The two variant systems simply change some assumptions about what the rolls represent.

Well, in those six seconds, if you choose a single attack rather than full attack, you are also moving 30 feet. Assuming you aren't dashing (since you can go even faster if you forgo your attack), that's a pretty decent movement speed.

...I've forgotten what my point was. I think it had something to do with "only having one good swing in 6 seconds is actually pretty reasonable".

In fact, something that I've always thought was silly about iteratives is that you have to have a sufficiently high BAB to do them on a full attack. If you're going to have massive penalties anyway, shouldn't you be allowed to make your maximum iteratives at low BAB but just have a negative attack bonus for the successive attacks?

In other words,
A level 1 fighter with 18 Str would have +5, +0, -5, -10 (instead of just +5)
A level 6 fighter with 18 Str would have +10, +5, +0, -5 (instead of just +10, +5)
A level 11 fighter with 18 Str would have +15, +10, +5, +0 (instead of just +15, +10, +5)
And a level 16 fighter with 18 Str would have +20, +15, +10, +5 (as normal)

To me, that makes more sense.


Anyhow, I suspect they say it is a rogue nerf because a rogue can do sneak attack damage on every hit and usually makes use of ways to lower the AC of the target.

[Ashtagon]

Anyhow, I suspect they say it is a rogue nerf because a rogue can do sneak attack damage on every hit and usually makes use of ways to lower the AC of the target.

If that is the reason, then it is mistaken.

A normal, non-sneak-attack hit may do 10 hp damage on average (numbers totally pulled out of thin air here). Let's assume our sneak attack hit does 30 hp damage on average.

using the hit probability numbers in my first example, the SA guy under system a) will score no damage 56.25% of the time, 30 damage 37.5% of the time, and 60 damage 6.25% of the time, for an average of 15 hp damage per attack sequence. The average, you should not be surprised to learn, is the same 15 hp per attack sequence for the two variant systems as well.

In cases where SA-style damage only applies to the first attack in a multiple-attack attack sequence (such as that of ninja and scout classes, I think (not looking it up)). those classes will actually be boosted.

If we assume a normal, non-iterative attack scores an average of 0.5 hits per round, the following numbers are also true:

1-attack iterative: 0.5 hits per round
2-attack iterative: 0.75 hits per round
3-attack iterative: 0.85 hits per round
4-attack iterative: 0.9 hits per round

Numbers rounded to the nearest 0.05.

Allowing a level 1 character to get 4-attack iteratives (or even 2-attack iteratives) from the start is a major power boost, even if some of those attacks are at a negative bab.

[Lord Vukodlak]

Don't claim your math is right without showing your work, If I said I knew the formula for generating cold fusion you'd ask me for it. Even if most of use don't understand it a few D&D players do, one of my player's is a physics major and works in research for 3M, I'm willing to bet he could say if your math right or not.

Do you take into account TWF rogues? or just TWF in general.
What about haste? or rapid shot, manyshot? and a crap load of other abilities that give extra attacks.
I was in a party with a TWF who on average hit two point or more times a round. Trading the potential of six attacks[seven when hasted] in order to be assured of one hit isn't worth it.

What about monsters like hydra's, dragons, or even ghouls who have multiple natural attacks, suddenly there damage potential is much higher.

I also think its wrong to assume 11 is the base to hit in a standard encounter, the fighter should have a good deal more then a 50/50 shot to hit on his first attack for mooks, especially if he's got buffs like heroism or a bards singing, flanking etc.
And once again, haste, rapid-shot, and flurry of blows gives multiple extra attacks at little to know penalty.



In cases where SA-style damage only applies to the first attack in a multiple-attack attack sequence (such as that of ninja and scout classes, I think (not looking it up)). those classes will actually be boosted.

The scout requires you move to gain the bonus, so they could only make one attack anyway. If they can't make multiple attacks they shouldn't benefit from rolling multiple times. Ninja's can get the bonus on every hit its just more difficult but its more difficult to deny the target their dex bonus for every attack. But it is possible.

Lastly trading the randomness for certainty takes some fun out of combat.

[FlamingKobold]

I'm unsure of how your examples at all illustrated the 3 systems... In A, people would need 11+ on the first, and 16+ on the second, not 16+ on both. In B, trailblazer's system involves 2 attacks, not one, and they are at a penalty (since you're assuming only 2 attacks). In C, I'm not sure where 15+ comes from, as your system involves no penalties and the bonus is equal to that of the first iterative.

If you care to post more in depth math, don't worry about me not understanding it, unless you have a degree in pure math or something. I'd like to see your full computations :smallsmile:

So here's the scenario that I'm running. You have a combatant wielding one weapon under perfectly fair conditions (no miss chance, no surprise round, no haste, etc) and full attacking an enemy. The character in all three scenarios has identical abilities and BAB, which is somewhere in the 6-10 range. There is a 50% chance that a standard action attack will hit. That is, you need an 11+ before any penalties. Here's the math:


a) Normal iterative system: need 11+ on first attack, 16+ on second. So 50% on first, 25% on second. This means 1/8 both hit, 3/8 both miss and 1/2 one hits.

b) Trailblazers: Need 13+ on both attacks. 40% chance of hitting on both. Thus 16% both hit, 36% both hit, and 48% one hits. About the same likelihood of one hit, approximately flips likelihood of both hit vs. both miss. Power up to unoptimized people full attacking against high AC enemies that have between 6 and 10 BAB in fair condition.

c) Your system: Need 11+ on at least one die. 75% to hit, 25% miss.

Assign damage will be similar and valued at “x”

a) .75x damage (.125*2x + .5*x)
b) .8x damage (.16*2x + .48*x)
c) .75x damage

All in all, trailblazer’s system yields greatest damage output.

However, that's vanilla damage for an unoptimized, full attacking creature that isn't smart enough to be using tactical advantage or, you know, having a decent to hit bonus. Here are a variety of factors not accounted for:

1. Having a high BAB/low AC opponent. If you are a level 10 fighter, you probably have all sorts of buffs and feats that add to what you do. For example, if you could computationally hit on a -1, yours and trailblazer's are the best for hitting, because you only have a 5% chance of missing on a given attack, while the standard iterative system can still miss with a 3 or less on the second attack. However, since in your system, you still only get to deal damage once, it's vastly suboptimal compared to either of the other two.

2. Very High AC creatures. As in, you need a 20 to hit. Your system hits once 9.75% of the times, while the other two hit at least once 9.75% of the time, but have a .25% chance of hitting twice. Not the best odds, but it still favors both iteratives and trailblazers.

3. Creatures with AC such that you need 16+ to hit. Need 16+ to hit

a) .0125*2x + .275*x = .3x
b) Need 18+ to hit, (15%), .0225 * 2x + .255*x= .31x
c) .2775x

This favors trailblazer, slightly, while putting yours in last place.

4. Criticals. This is assuming your critical threat hits (which will happen with most characters). The most commonly used weapons have threat ranges 19-20 at worst, and even a 14+ should be hitting appropriately CR'd monsters, assuming you're smart) all three systems have the same likelihood to threaten a critical (they all roll two dice). In your system, you can only threaten once, while in the other two, you can do it twice. Then we go back to hitting AC, but it's just one attack. Yours hits most often, trailblazers 10% of the times less, and standard iteratives 15% less. Trailblazer has a slight advantage over standard D&D in this area, but the fact that you can only crit once makes it a lot closer. All in all, it depends on the weapon.

5. Being sneaky. This is where the whole "bad for rogues" thing comes in, and applies more clearly at higher levels. Flat footed AC is lower than normal AC. In many cases, significantly lower. Your system can, at maximum, deal 1 sneak attack per round (two if you're dual wielding). Trailblazers can do it twice per round, and standard D&D does it between 2 and 4 times, depending on level. Let's take a PC. At level 16, unless you wear full plate, your flat-footed AC will be between 5 (little magic support) on up to 15 (with classes like monk or swordsage) points lower than normal AC. These figures assume relatively smart play choices (hmmm... getting hit less would be good...). This means that more attacks are drastically superior to hitting more often, because you're hitting a lot anyway. Even assuming being flat-footed only drops AC by 5, the D&D iteratives system still makes it more likely to land attacks, thus makes sneak attack/sudden strike damage apply.

[Ashtagon]

I did specifically note that it was a simplified example to illustrate the concept. I have a more worked example using the actual attack probabilities from RAW somewhere on my hard drive that I did months ago, but it'd probably be easier to work it from scratch over again than to search for it :smalleek:

I did not do any calculations for monsters with multiple attacks, or for characters with TWF, as the game treats those quite differently from iterative attacks. I would not recommended using that "roll twice take best result" option to resolve TWF extra attacks for that reason.

[Agi Hammerthief]

not going for the higher math here, I'm with FlamingKobold on the points 4 and 5


You have to keep track of different sets of modifiers, and you roll them one at a time,
why do you roll one at a time?

just get 4 d20 with different colors and assign a color to each order of hit.

one roll for the attacks, count number of hits -> roll as many dX for damage

or get your damage dXs in the same colors as the d20s and be done with it in one roll.

[Apalala]

Mm. Things like Haste and TWF would work much the same under any of the rules.

Using Trailblazer, if you were swinging with a bonus of +10/+10 normally, the attack bonus and extra attack would make that +11/+11/+11. Whereas in the standard format +12/+7 would turn into +13/+13/+8.

If the same fighter had TWF and a light weapon and such, he'd be swinging for +10/+10/+5 in standard, +8/+8/+8 in trailblazer. Add haste into the mix and standard is now +11/+11/+11/+6 and trailblazer is +9/+9/+9/+9.

At higher level, a hasted TWF fighter with an attack bonus of +20 would swing for +21/+21/+21/+16/+11/+6 in standard, and +21/+21/+21/+21 in trailblazer.

Given the choice between +21 and +16/+11/+6, I'll take the 21.

[Apalala]

not going for the higher math here, I'm with FlamingKobold on the points 4 and 5


why do you roll one at a time?

just get 4 d20 with different colors and assign a color to each order of hit.

one roll for the attacks, count number of hits -> roll as many dX for damage

or get your damage dXs in the same colors as the d20s and be done with it in one roll.

Decent way to keep track of things, but still more trouble than its worth. You' roll 4 dice at once. Then you have to individually check off each roll to determine whether or not it hit. And no matter how efficient you are about it, it's still rolling three dice instead of one.

[FlamingKobold]

Given the choice between +21 and +16/+11/+6, I'll take the 21.

Why? Did you do the math, or what? This really depends on the situation.

On another note: can someone check my math? I want to make sure it's right.

[Lord Vukodlak]

More efficient doesn't necessarily translate to better.

And you can roll your attack rolls during someone else's turn and keep track of what AC was hit each time.
When it comes to my turn playing my fighter, I can simply list off what AC"s I hit then the damage.

[Apalala]

Why? Did you do the math, or what? This really depends on the situation.

On another note: can someone check my math? I want to make sure it's right.

Well, yes, it depends on the situation. If something has 5 AC, then of course the more attacks the better. But if it's something with, say, 31 AC, I'd much rather just make the one roll.

Hell. Even if I was hitting on a 2, I'd still want the one attack.

[Agi Hammerthief]

Well, yes, it depends on the situation. If something has 5 AC, then of course the more attacks the better. But if it's something with, say, 31 AC, I'd much rather just make the one roll.

Hell. Even if I was hitting on a 2, I'd still want the one attack.

I'd rather have my 4 rolls and a chance to 4 natural 20s (or 4 chances to 1 natural 20 respectively)