Torment Armor "Penalty" -- May Not Exist?

[Slawbug]

Lilith and I did some poking around and we have a theory.

 

Our theory is that there is NO penalty to armor or resistance on Torment. However, enemy damage is doubled (plus a little; see below) and there is a text bug.

 

Basically, the game engine doubles the enemy damage, and correctly calculates how much of it you block. It then prints the damage you take correctly, but for blocked damage, it makes a mistake. Instead of printing (Attack Damage - Damage Taken) it prints (Attack Damage ON NORMAL - Damage Taken ON TORMENT). This is why damage blocked will usually be 0 when your armor is below 50% on Torment. You are actually blocking damage, it just isn't displaying due to this bug.

 

The other piece is that we think (as I have been saying since A4) that Torment gives enemies the effect of being at a higher level in many ways, raising their auto-calculated stats and resistances. Not all ways: it doesn't raise experience or, presumably, Capture Soul requirements. Since level and base stats all factor into attack formulae, this will also raise total damage slightly on top of the 200% effect; but the precise percentage of this raise can vary depending on attack base damage, etc.

 

We aren't quite certain this is what's going on but it seems to fit better than any other explanation we've tried to fit to the data (including the old chestnut subtractive -30% explanation).

 

Anyway, if this is correct, it means that armor and resistance are even more important on Torment, if anything.

 

If anyone observes anything that is clearly NOT in the ballpark for this theory, we'd love to hear about it.

[Lilith]

Basically, the game engine doubles the enemy damage, and correctly calculates how much of it you block. It then prints the damage you take correctly, but for blocked damage, it makes a mistake. Instead of printing (Attack Damage - Damage Taken) it prints (Attack Damage ON NORMAL - Damage Taken ON TORMENT). This is why damage blocked will usually be 0 when your armor is below 50% on Torment. You are actually blocking damage, it just isn't displaying due to this bug.

 

By the way, if you want a quick demonstration of this at home, try taking off all your armour at a point early in the game when you're still seeing 0 for damage blocked. Your damage blocked will still display as 0 either way, but you'll take considerably more damage without armour than with it.

[Bonam]

Should be easy to test?

 

If your armor is at 90%, take off equipment to make it 80%. If there's a linear 30% subtraction, your effective armor will go from 60% to 50%, so the damage you take will increase from 0.4 to 0.5, or 25% (i.e. if you were taking 100 dmg from someone you will now take 125).

 

On the other hand, if there is no linear subtraction and its just a reporting issue, you should take double the damage at 80% armor compared to 90% armor (if you were taking 100 damage you will now take 200)

 

Given the huge difference we are talking about, would probably only need a very small sample to find out for sure.

[Lilith]

On the other hand, if there is no linear subtraction and its just a reporting issue, you should take double the damage at 80% armor compared to 90% armor (if you were taking 100 damage you will now take 200)

 

I've tested it and this is more or less what I observed. Testing against Athron on Torment with no buffs or debuffs active, her bite does about 200 damage to a character with 90% armour, about 400 damage to the same character with some equipment removed to reach 80% armour, about 600 damage to the same character with further equipment removed to reach 70% armour, and about 800 damage to the same character again with even more equipment removed to reach 60% armour. If I remove all her armour, she has a residual 44% armour rating from Hardiness and Luck, and sure enough, as the model predicts, she takes about 1100 damage (with 0 damage reported as blocked). If I use the editor to remove as much Hardiness and Luck as possible (all points except for those gained from trainers and special encounters), her residual armour rating is 11% and she takes about 1700 damage, also with 0 blocked damage reported. If I try the same tests on Normal, in each case the damage taken is a little less than half the amount on Torment, and a nonzero amount of damage is always reported as being blocked. The sample sizes are smaller than would be ideal and there's enough variation in the data that I can't say for sure that the Torment damage multiplier is exactly 2x (especially since we suspect additional effects are in play from difficulty-based modifications to monster levels), but the pattern is pretty clear.

 

The take-home message:

 

1) To a first approximation, you can expect to take about twice as much damage per hit on Torment as on Normal difficulty. (There are also other effects: enemies on higher difficulties have faster initiative, better accuracy, and sometimes additional abilities. But right now we're just looking at raw damage per hit that lands.)

 

2) Even if armour or resistances don't appear to be blocking any damage on Torment, they are: you'd be taking even more damage without them.

[Bonam]

Got it, so the take away is armor works correctly by the listed % on Torment, despite what the damage blocking text says. So armor rating and getting it as close as you can to 90% is very important (since the difference between 85% and 90% is huge compared to the difference between 55% and 60%). I assume this all applies for elemental resistances as well.

[Slawbug]

Yes to the take away but not to the parentheses. Yes, the difference between 85% and 90% is huge in terms of damage taken, but it's also huge in terms of accruing armor and resistance. Remember that those are acquired multiplicatively, i.e., if I have two pieces of armor that each protect 40%, I don't have 80% protection; rather 60% is applied to my damage taken twice, leaving me with 36% taken, or 64% protection.

 

Thus, in order to get from 85% to 90%, you need to add approximately another 33% protection -- that means either one piece/effect that supplies 33%, or multiple pieces that supply rather more in sum. That extra 33% is equal to the damage decrease you get going from 85% to 90%, too. The difference is just that you lose efficiency for spreading it out over multiple pieces/effects.

 

This is also part of the reason why high Hardiness and Resistance are so great: they each count as a single piece. Likewise for Mental Resistance from Intelligence.

[Hume]

Should be easy to test?

 

If your armor is at 90%, take off equipment to make it 80%. If there's a linear 30% subtraction, your effective armor will go from 60% to 50%, so the damage you take will increase from 0.4 to 0.5, or 25% (i.e. if you were taking 100 dmg from someone you will now take 125).

 

On the other hand, if there is no linear subtraction and its just a reporting issue, you should take double the damage at 80% armor compared to 90% armor (if you were taking 100 damage you will now take 200)

 

Given the huge difference we are talking about, would probably only need a very small sample to find out for sure.

 

Armor penalty does never mean linear subtraction in any Avernum game. The armor penalty in A5, A6 means you take double damage if your armor resist more than half of the damage, or take the full damage if otherwise.

The only difference between how it is in Avernum 2, is that armor penalty treat you as 50% armor resistance when you actually have less.

 

Armor is not therefore a lot more important - you will pass 50% armor soon anyway. Then they will be identical.

 

Slartibus is right that the importance of an armor piece has nothing to do with your existing armor level. A 10% armor piece will reduce damage taken by 10%, no matter you already have 50% or 85%.

[Bonam]

Yes to the take away but not to the parentheses. Yes, the difference between 85% and 90% is huge in terms of damage taken, but it's also huge in terms of accruing armor and resistance. Remember that those are acquired multiplicatively, i.e., if I have two pieces of armor that each protect 40%, I don't have 80% protection; rather 60% is applied to my damage taken twice, leaving me with 36% taken, or 64% protection.

 

Yes, I'm aware of the mechanic. The point I was making was that without the 30% subtraction that was supposed, getting to 90% is really important... whereas if there was a 30% subtraction, getting to 90% would not be nearly as important, since all that huge effort to accrue those last few % of armor would just be subtracted down anyway to a range where it matters a lot less (hence my statement that the difference between 85% and 90% is a lot bigger than the difference between 55% and 60%).

[Slawbug]

Bonam: Ahh, gotcha.

 

Armor penalty does never mean linear subtraction in any Avernum game. The armor penalty in A5, A6 means you take double damage if your armor resist more than half of the damage, or take the full damage if otherwise.

The only difference between how it is in Avernum 2, is that armor penalty treat you as 50% armor resistance when you actually have less.

No, it does not. The DISPLAY is bugged, but the actual calculations are done correctly. You can test this yourself: on all difficulty levels, a character with 0% armor takes twice as much damage as a character with 50% armor, while a character with 25% armor will take 1.5x as much as the 50% armor dude. The difference is just that the display is bugged on Torment.

 

I strongly suspect that this same mechanism was at work in A5 and A6 as well -- it's probably the same display bug inherited from one game to the next.

[Hume]

Bonam: Ahh, gotcha.

 

 

No, it does not. The DISPLAY is bugged, but the actual calculations are done correctly. You can test this yourself: on all difficulty levels, a character with 0% armor takes twice as much damage as a character with 50% armor, while a character with 25% armor will take 1.5x as much as the 50% armor dude. The difference is just that the display is bugged on Torment.

 

I strongly suspect that this same mechanism was at work in A5 and A6 as well -- it's probably the same display bug inherited from one game to the next.

 

Sorry my wording is not clear, and you misunderstood me. I was describing what is happening in A5 and A6, and arguing that in A5 and CS, the difference is NOT important, since they work the same once you have over 50% armor.

[Slawbug]

It looks like the same thing is happening in all the games from A5 through A2:CS, though. Double damage on Torment plus a display bug. There is no actual armor penalty, nor is there a 1/3 increase in total damage.

[Hume]

Just tested Avernum 6, seems you are right about it is a display bug. Armor does help even when it is low. But you are wrong about doubling damage without another 1/3. My number suggest it may be greater than 8/3 multiplier

 

Normal

Armor Damage

33% 11.7

50% 10

68% 6.8

 

Torment

Armor Damage

33% 46.3

50% 35.7

68% 23.5

[Bonam]

Torment

Armor Damage

33% 46.3

50% 35.7

68% 23.5

 

So looks like armor is very much just a straight up reduction by the listed % on Torment without anything funny going on. With a base damage of ~71, the numbers above are well explained:

 

71 * 0.67 = 47.6 ~46.3

71 * 0.50 = 35.5 ~35.7

71 * 0.32 = 22.7 ~23.5

[Slawbug]

So, the other part of the OP is that changing the difficulty changes the auto-calculated stats for every NPC. That is what is responsible for any increase in damage beyond the double damage effect of Torment. It's likely that this has a different impact on enemies of different levels, and also affects different attacks differently.

 

In other words, if you can show me a single consistent 8/3 multiplier (or whatever) on damage from a variety of different monsters, of different levels, and with different attacks, then I'll believe it. Otherwise it's just 200% plus NPC modifiers.

 

(Also, out of curiosity, what the heck were you fighting that did 20 damage on Normal and 71 damage on Torment?)

[Hume]

Well, they are the bats in Blosk guarding the iron cache.

[Bonam]

So, the other part of the OP is that changing the difficulty changes the auto-calculated stats for every NPC. That is what is responsible for any increase in damage beyond the double damage effect of Torment. It's likely that this has a different impact on enemies of different levels, and also affects different attacks differently.

 

In other words, if you can show me a single consistent 8/3 multiplier (or whatever) on damage from a variety of different monsters, of different levels, and with different attacks, then I'll believe it. Otherwise it's just 200% plus NPC modifiers.

 

Why wouldn't it just be ALL NPC modifiers rather than a combination of those plus some hidden 2x damage factor?

[Lilith]

Why wouldn't it just be ALL NPC modifiers rather than a combination of those plus some hidden 2x damage factor?

 

Well, for one thing, because that wouldn't provide a plausible explanation for the errors we see in the displayed amount of damage blocked by armour/resistances on Torment -- our working hypothesis for the error is that the game is calculating the damage blocked by comparing what the base damage would be on Normal to what the actual inflicted damage is on Torment, and it doesn't make sense for that kind of error to even be possible unless there's a difficulty-based modifier applied in the damage formula itself, not just to enemy stats. For another thing, because accuracy and damage are linked to the same stats, NPC level being increased enough to double the damage they inflict would also increase their accuracy to the point that it'd be absolutely impossible for even a Dex-based character to see enemy hit rates below 90% at higher levels on Torment, which is demonstrably not true. In short, doubling all NPC stats would produce knock-on effects that we don't see, and fail to produce effects that we do.

[Bonam]

Good points. Hypothesis does not match the available evidence. Thanks