What are the proper stats for a level 50 shield?

[ducci]

I'm continuing a conversation Xorex and I were having in chat the other day: how could shields become more useful? Obviously they aren't useless for Illithids, and there are some niche situations where a character might temporarily opt for a shield. But, particularly at the higher mortal levels, it is impossible to justify a 50% reduction in autoattack (AA) damage in exchange for the bonuses on even the most powerful shields in the mud. The easiest way to make shields competitive is just to add new shields to the mud with competitive stats or to increase the stats of some of the current shields. So what would the stats on a competitive shield look like?

Assuming the shield has the same damroll bonus as the weapon it's replacing, it gives a 50% reduction in AA damage for Dwarves, Kender, Elves, and Druids. Using some plausible numbers for level 50 Avians and Humans, the shield reduces AA damage by 45% and 37% respectively. If it reduces AA damage by 50%, a balanced shield ought to enhance armor class by the same proportion. So a competitive defensive shield for most races at level 50 would look something like this:

+15 damroll
+15 hitroll
-200 armor class

What if we want a more offensive shield (say, a buckler) that decreases AA damage by only 25% and consequently enhances armor class by only 25%? Believe it or not, that would require a damroll bonus of 90-95. Thus:

+93 damroll
+15 hitroll
-100 armor class

It looks so wrong, but unless my math is off that's what "balanced" looks like. If you give that shield to a Dwarf he will do 25% less damage and have 25% better armor.

[ducci]

Okay, my math was indeed off. I had always thought that second and third attack applied to the offhand weapon, too.

Here's my corrected numbers, which are a little less outrageous:

If a shield has the same damroll bonus as the weapon it replaces, it reduces the AA damage of Elves and Druids by 40% and of the other races by about 35%.

A level 50 defensive shield, which reduces AA damage and enhances armor class by 40% each, would have the following stats:

+15 hitroll
-160 armor class

A level 50 offensive shield, which reduces AA damage and enhances armor class by 25% each, would have the following stats:

+15 hitroll
+55 damroll
-100 armor class

I think the takeaway is the same as before: shields need to have either significantly more armor class bonus or a HUGE increase in damroll bonus to balance out the damage penalty they come with.

[Everybody]

Your numbers might still be off - there's a bunch of back-end code balancing that determines how often the extra attacks go off that's not visible to a player. To pull back the curtain a little bit:

Assuming that you perform a full combat round - i.e. the enemy doesn't die mid-round, you don't switch targets, flee, etc., your auto-attacks look like this:

• Every round you attack with your primary weapon.
• Every round you attack with your offhand weapon - note that this is a reduced damage hit based on your proficiency in dual-wielding.
• Second attack goes off in 50% of rounds (even if you have 100% skill level).
• Third attack goes off in 25% of rounds (even if you have 100% skill level).
• Avian claw goes off in 33% of rounds, plus 16% of rounds. Note that this gives you 56% of rounds where no claws go off, and 5% of rounds where two claws go off.
• Five style fist goes off in 33% of rounds.
• 1000 palms goes off in 33% of rounds.
• Mind thrust goes off in 25% of rounds.
• Racial autoattacks (demon limbs) always go off if you have the appropriate limb - this is part of why full auto limb builds for demons are so powerful.

There's some differences between the separate auto-attacks as far as how much damage they deal (even before considering poisoned weapons/enhanced damage), as all of the non-attacks don't get to add in weapon damage to them. So, the value of the offhand weapon actually increases marginally in those cases, as we're losing out on a bigger damage hit. We're also not considering talent levels here - elves, druids, and kender have 11% less likelihood of triggering each extra auto at max learn levels, compared to dwarves (pre-avatar).

Using a relatively average-ish equipment load-out for level 50 comes in around 110 damroll. So, ignoring weapon damage, an AA output per round with/without shield looks like:

• Martial - 302 with offhand, 192 with a shield. If we want the shield to put out competitive damage, it would need to provide ~60 extra damroll as compared to the weapon it was replacing. This presumes we're ignoring the defensive benefits momentarily.
• Spellcaster built for AA - 275 with offhand, 165 with a shield. For the shield to put out competitive damage, it needs ~70 extra damroll as compared to the weapon it was replacing.
• Spellcaster built for spells - competitive level 50 damroll here is around 85. Damage with offhand is 212, 127 with a shield. For competitive AA damage with a shield, we'd need +55 damroll on the shield. Note that this doesn't consider the loss of a flaming spell weapon for extra spell damage procs (and assumes the shield is flaming).

As a final note, I didn't spend any time analyzing the defensive benefits of a shield - using a level 50 shield provides 14-15 AC at base, where a level 50 character I'd expect to have somewhere in -100 to -180 AC pre-spells. Since AC starts at +100, this is effectively 200-280 total AC. At -100 AC, 15 extra AC is a 7% defensive improvement.

[ducci]

Thanks for weighing in EB! It seems like our numbers are fairly close. Just for posterity, here's a little more detail about the assumptions I made:

I did take the probabilities of the different autoattacks into account, based on some comments you made a few years ago The only difference was in the probability of the second claw--I had that as 1/12 probability instead of 1/6. And I included the enhanced damage multiplier and the effect of critical hit for Kender, although I don't think those affected the final percentages. I assumed Frostbite for the weapon (25 average damage and +15 damroll bonus). I assumed total dual-wielding damroll was 130 with a net armor class of -200. Your numbers are probably a little more realistic in that respect.

I didn't take the racial differences in practice level for dual, second attack, and third attack into account. And of course I didn't consider the lower damroll a spellcaster build would have.

All in all it seems like it's hard to make an "overpowered" shield. At level 50, big damroll bonuses in the range of 40-60 should still cause a noticeable drop in damage, to be offset by big AC bonuses in the range of -100 to -200, possibly in combination with HP or MP bonuses to make things interesting.

[Everybody]

Overall, agreed. As far as baseline stats - I'd consider it fairly hard to achieve 130 damroll and -200 AC (pre-spells) at level 50, given that going for damroll typically means the evil build and thus a much worse AC and hitroll to compensate. Since I'm discussing it, considering AC without spells is reasonable, but it means that any defensive bonuses on the shield are actually weaker than their numerical value suggests due to diminishing returns. I.e. a 40% defensive buff when calculated against -120 AC is only a 25-30% buff when calculated against -220 AC. (This gain from spells, of course, presumes access to the full range of defensive buffs across all races - bless, combat mind, armor, shield, intellect fortress, etc.)

In numerical values, for a 40% gain in defensive utility to offset the offensive loss (taking into account the ~15 AC the shield naturally gives us):

• At a -100 AC baseline, we'd need -65 additional AC from the shield
• At -150 AC, we'd need -85 additional AC
• At -200 AC, we'd need -105 additional AC

And, finally, our analysis didn't include anything on hitroll percentage calculations - presumably since we're not changing hitroll in this thought experiment, all of the individual attacks would continue to hit at the same rates, thereby maintaining our calculations accurately. I mention this because bumping the hitroll up on shields also provides us an option for returning offensive stats, although it's somewhat harder to quantify.

[xorex]

I have never studied the combat math well enough to contribute to the detailed analysis. It is an interesting question though of what stats a shield would need to be competitive with dual wielding. One other aspect you might consider is that if you are balancing the ratio of damaged dealt to damage received in the two cases then for the shield case both numbers will be lower leading to a slower combat and so more time for extra actions between rounds like a spellcaster casting a damage spell to be performed within a combat session.

Are shields just pure stat boosts now? I could imagine some game mechanic that gave an extra save roll when using a shield that would block an attack (and say "You block mob's attack with your shield!") with a chance of damaging the shield. My comment in the chat was that it was a little weird how heavily skewed towards dual wielding the game is since in real life I would guess most fighters would be better off with a shield than a second sword (though not every one).

[ducci]

And, finally, our analysis didn't include anything on hitroll percentage calculations - presumably since we're not changing hitroll in this thought experiment, all of the individual attacks would continue to hit at the same rates, thereby maintaining our calculations accurately. I mention this because bumping the hitroll up on shields also provides us an option for returning offensive stats, although it's somewhat harder to quantify.

Haha yes, I was keeping the hitroll equal for the sake of simplicity. My attempts to figure out hitroll have not exactly succeeded.

Are shields just pure stat boosts now?

As I read this, I started to say, "yes, they're just armor for your offhand", but it hit me that you could make an item that is a shield in its description but is actually a weapon with "pound"-type damage. So now you're carrying a sword and shield, but as long as you have dual, you're pounding your targets with that shield every round. Suddenly, balancing the stats on the shield gets a lot easier. You might even put a negative damroll value on it, depending on the other stats. And if it's a spell-weapon, with self-targeting defensive spells for example, it's doing two things in addition to boosting your stats. It might not be the ideal solution, but you wouldn't have to change the code.

[Everybody]

One other aspect you might consider is that if you are balancing the ratio of damaged dealt to damage received in the two cases then for the shield case both numbers will be lower leading to a slower combat and so more time for extra actions between rounds like a spellcaster casting a damage spell to be performed within a combat session.

Are shields just pure stat boosts now?

Considering damage ratio (dealt to received) is certainly a more overall accurate analysis than what we've done so far. To do so, however, we'd have to get into mob damage and to-hit calculations, and likely perform them at several comparative points - same level, below level by 2, above level by 2, above level by 5, etc.

As far as the code's concerned, yes, shields are pure stat boosts, like any other piece of armor. Adding them into the offhand slot provides you with a modicum of additional base armor as compared to dual wielding (as noted below), but otherwise has no particular advantages. In some of the various race/skill proposals, utilization of shields to offer an additional dodge/parry-like skill have been fairly common, but none of those are currently implemented.

My attempts to figure out hitroll have not exactly succeeded.

We can take the full hitroll calculations into another thread, if you want the details around it. As a brief, we take a random number from 1 to the target's AC (utilized as a positive number), and if it's less than the attacker's hitroll, it hits. Various affects modify this, depending on the affect, and there's a flat miss chance built-in for the case that hitroll is higher than the target's AC. We additionally taper off the high end of AC in a few cases to prevent mobs from being unhittable.