Protection and Defense

[Amorphous]

Originally Posted by Peter Ebbesen

Err, no, or rather yes to the last postulate and no to the earlier "vanishingly unlikely" one.

There are several SC chassis that start out with extremely high defense and routinely sport well over 30 defense, the most obvious one being our friend, the Djinn, who starts with a defense of 20, hits 30 with ease with one shield and possibly a bit of experience, and is often seen dualwielding shields, but more generally any SC type unit devoted to anti-SC duties where not being hit at all by an opponent's weapon of doom takes on increased importance over being able to survive the weapon hitting you. (Fragile SCs like the queens of elemental air are other obvious candidates for stacking defense rather than protection). Shields like Shield of the Accursed are in their element here, as are Shield of Valor and Shield of Gleaming Gold.

There are lots of defense 15-16 chassis in the game, many good for a pair of boots of speed as part of their thug/SC duties, that will hit 30+ defense with a bit of experience under their belt should you be willing to invest in them and while it may not happen in your MP games, it certainly happens a lot in mine.

It certainly isn't the most common way to see thugs and SCs kitted out, but it has its place and isn't vanishingly rare.


As an example, here's my favourite anti-SC SC from an ongoing game (where his job is obviously to jump anybody using an SC or the like in my dominion). Everything is easily and cheaply replacable (except for the Mage Bane, which is in no way essential to making him exceedingly nasty - it just makes him even nastier versus his intended prey), and his high defense means that most SCs and even armies of normal units are likely to miss.

http://img213.imageshack.us/img213/1889/antisc.jpg (picture is with an E4D4B4 minor blessing as prophet in dom 10 and with heroic toughness 56)

Any attacks that does go through his does not have to deal with much protection but does have to deal with 81 hp regeneration/round. Other armours used to counter whatever the opponents are fielding (resistances, even more defense, reinvigoration+even more MR - the hydra skin is just the basic fallback for the last ~24 hp/round). This one is definitely a "nuke with Gift from the Heavens" type of SC, for overwhelming him in melee with thugs/SCs (probably armed with morningstars if they want to hit) is likely to end up with the surviving winners horrormarked very feriously.

Which non-artefact shield would be a better choice for him in general than the Shield of the Accursed he is using? Some other shield might be better for specific niché uses, but that's what swapping equipment to better face known threats is for.

I can of course not speak for Agema, but I find it likely that he did not include a shield's parry value in the defence value, which you evidently do. The thing is that the parry part of the defence is not as good as pure defence. A shield hit is still a hit.

The added protection is nice, of course, but nowhere near as good as not getting hit. AP weapons halves the protection provided, AN weapons makes it irrelevant and as long as some damage goes through all sorts of nasty special effects still apply.

Your SC is a rather good example of the state of things, I think. It does have a high defence in the upper twenties, but a lower protection. A good shield makes the effective protection a lot higher in most instances, but 30 protection is still not that much considering it includes the shield protection. At the level of equipment indicated by his gear punching through 30 protection is rather easy. Just put him up against himself - he has a strength of 31. Using AP or anti-undead weaponry will make it even more effective.

Also, do not make too light of the Mage Bane. The weapon has one of the best defence values in the game. Without it he is much easier to hit.

[Peter Ebbesen]

Originally Posted by Amorphous

I can of course not speak for Agema, but I find it likely that he did not include a shield's parry value in the defence value, which you evidently do. The thing is that the parry part of the defence is not as good as pure defence. A shield hit is still a hit.

Do you have some sort of evidence that the parry part of the total defense attribute is not counted as part of the defense? My understanding is that it is added on top of the total defense (which already includes the shield parry attribute) for purposes of checking for a shield hit.

The reason that I am of this belief is that it seems to fit what I observe and that, in addition, that is how the mechanic is described very explicitly in the manual.

On page 75 in the manual, as the explanatory example, there's a Jotun Axeman vs a Black Centaur - the Centaur with defense 17, the Jotun with attack 9. Those 17 in defense includes +4 parry (the unit does indeed have base 14, -1 for armour+shield combo, +4 parry, for a total of 17). It is stated as an example how, if the Jotun rolls a total of 18 and the Centaur a total of 9, i.e. 27 vs 26, it is a hit exceeding the Centaur's defense by one and hence a shield hit because it is a hit that is below the defense value (which included the parry value) plus the parry value (which determined the interval in which to check for shield hits), and it would have needed to be five higher (i.e. 31) to be a normal hit. In other words, the defense roll is based on the defense attribute as you can see it in-game, shield parry included.

So according to the manual, at least, your interpretation is completely wrong and mine is right - but the manual has not proven without error in the past, and my observing what I expect to observe (the game working as is stated in the manual) is certainly no proof that it is actually doing so.

It is possible that it is generally "known" by the community that the manual is wrong and that things work differently and the shield parry isn't counted for complete defense, but I haven't come across this knowledge and, since the manual is generally trustworthy, I tend to trust it when lacking evidence to the contrary.

Do you have some sort of good tests showing that it works the way you believe to or word of god from our friendly developers (i.e. the standard "we think we implemented it this way, perhaps, somewhere in the code, I've forgotten where, I'll forget my own head next" or "the manual is wrong") on this issue?

Your SC is a rather good example of the state of things, I think. It does have a high defence in the upper twenties, but a lower protection.

The defense is in the upper thirties and the upper forties for checking shield hits... should I be right. Should you be right it is a great deal lower.

In either case, the protection value is poor - in cases where I expect many enemies to beat through it heavily, I do of course equip a heavier armour (Elemental Armor is always a favourite on SCs for general armour and resistance purposes).

A good shield makes the effective protection a lot higher in most instances, but 30 protection is still not that much considering it includes the shield protection. At the level of equipment indicated by his gear punching through 30 protection is rather easy. Just put him up against himself - he has a strength of 31. Using AP or anti-undead weaponry will make it even more effective.

Most certainly - I am not in any way, shape, or form claiming that it is a unit that cannot be defeated or won't take severe damage from thugs or SCs kitted out to deal with the undead - so long as they are capable of hitting it with some reasonable chance of success.

However, given that the hydra skin gives him about 24 hp/round and body protection 8, while an Elemental Armor would give him body protection 18, he needs on average to be hit more than 2.4 times per turn by attacks big enough to deal significant damage - and in a world with armour negating or penetrating weapons the advantage of the heavier armour with defense penalties dwindles even further, those 10 extra subtracted per attack may turn into 5 extra or even, worst case, 0 extra, boosting the value of the hydra skin considerably. (This all because of his outrageous hitpoint total - the frailer the target, the higher the value of protection over regeneration)

Also, do not make too light of the Mage Bane. The weapon has one of the best defence values in the game. Without it he is much easier to hit.

It is an incredibly good weapon for many reasons including its +6 defense - but again, unless I am shown to be wrong on the defense issue (in which case I may well weep a few manly tears), I'll keep on believing that his defense to be hit at all is closer to 40 than 30 and that the +6 rather than the normal +2 to +4 for most non-artefact swords he might be using instead, while certainly wonderful a another perk of a great weapon, isn't a major reason for his being hard to hit.

[Amorphous]

Originally Posted by Peter Ebbesen

Do you have some sort of evidence that the parry part of the total defense attribute is not counted as part of the defense? My understanding is that it is added on top of the total defense (which already includes the shield parry attribute) for purposes of checking for a shield hit.

The reason that I am of this belief is that it seems to fit what I observe and that, in addition, that is how the mechanic is described very explicitly in the manual.

I cannot offer any substantial evidence, but I can tell you why I think that shields work as I said.

There are essentially two reasons for my line of though:
1. The description in the manual in combination with how defence is displayed in game makes me think so.

As you correctly noted, the given example does not really support my reading, but that presupposes that we already know that the 17 defence of the Black Centaur includes the parry value of the shield. It is not mentioned anywhere in the example text.

When you look at the break down on the defence value in game, the parry value is added separately as parry value and not anything else. It seems very strange to me that the parry value should be counted twice in addition to the defence modification - a much simpler way would have been to instead have a defence value that included the defence bonus from parry and just count parry once.

If you discount the example, just counting parry once seems in line with the rest of the text.

2. The performance of shields I have observed in game seems in line with my interpretation.

I know that you say the exact opposite, but consider e.g. the performance of light or medium shield-bearing infantry armed with spears. As a general rule they should not be able to hurt each other at all for quite a few rounds if your interpretation were correct. This is not consistent with what I have seen.


I know that this cannot be said to be a strong argument, so you should in no shape or form feel obligated to accept it. However, I am sure enough of my in game observation that I will not abandon my belief on the strength of the example in the manual. (I should add that, in general, I am not at all averse to count the manual as significant evidence - it is more often correct than not.) If enough members of the community tell me I am wrong, I will, though.

[Peter Ebbesen]

Originally Posted by Amorphous

When you look at the break down on the defence value in game, the parry value is added separately as parry value and not anything else.

So is extra defense from Quickness +3, which also gets its own line, or Prophet +2, Experience + X, and several other (all?) defense modifiers that get added separately and aren't labeled +defence on equipment, and the sum of all those are listed as the "Current Defence and Parry skill"

It seems very strange to me that the parry value should be counted twice in addition to the defence modification - a much simpler way would have been to instead have a defence value that included the defence bonus from parry and just count parry once.

I can understand what you are saying, but it also makes excellent sense the other way around.

The way you are thinking of it (parry value used once to determine whether an opposing attack is a shield hit that reduces the damage you take, but not adding to defence skill) a shield does not actually help prevent you from being hit - it merely reduces damage you take if you manage to interpose it between yourself and an incoming attack that would otherwise deal damage to you. The damage may be reduced to 0 (in line with the general protection mechanism) but probably not in case of really strong opponents.

This seems in stark opposition to not only several shield descriptions in-game (the Shield of the Accursed, the Shield of Gleaming Gold, and the Aegis all explain how the one using them becomes harder to hit because it is difficult to focus on the wielder), but also how shields have traditionally had dual use in warfare, both deflecting and absorbing blows (some better at one than the other and also depending on their martial use).

As such it makes excellent sense to have shields' both add to defense (value A) and to determine a range in which there's a shield hit (value B). This requires two different values to really fine tune - and shields have that: the parry value (x) and the defense value (y); in other words x = b, y = A - x

I know that you say the exact opposite, but consider e.g. the performance of light or medium shield-bearing infantry armed with spears. As a general rule they should not be able to hurt each other at all for quite a few rounds if your interpretation were correct. This is not consistent with what I have seen.

Independent Light Infantry using a spear: attack 10, defense 13 - including shield +4, and protection 15/8 [S15]. Damage is 10 + 3 + DRN - DRN - relevant_protection at location hit when considering shield protection into the mix.

Defense decreases by 2 per multiple attack and by fatigue/10 and attack by fatigue/20.

Under these conditions and given two groups of statistically significant size of Light Infantry, I most certainly expect a fair number of hits from the very first round they are in melee range of each other, for the difference between attack and defense and even attack and defense+parry is rather small (3 and 7 respectively) and the RNG will have its say in the +DRN - DRN business; moreover, the little light infantrymen won't square off one by one - given an equidistribution over each infantryman over those enemy infantrymen he can attack, it is practically certain that some infantrymen will be attacked more than others (and hence be easier to hit for every attack after the first one) and given the number of size 2 figures per square, some a lot more.

It might come out differently if I actually spent the time on doing a real statistical analysis of the problem but at least my rough mental napkin map (without a napkin ) comes nowhere close to your conclusion that "LI vs LI with spears and shield should not be able to hurt each other at all in the first few rounds if non-shield hits only occurred at above another +4 on top of the 13 defense rather than at above 13.

[Amorphous]

Originally Posted by Peter Ebbesen

So is extra defense from Quickness +3, which also gets its own line, or Prophet +2, Experience + X, and several other (all?) defense modifiers that get added separately and aren't labeled +defence on equipment, and the sum of all those are listed as the "Current Defence and Parry skill"

Exactly, it is called "Current Defence and Parry skill", not "Current Defence skill" or "Current Defence, Blessed, Experience, X and Parry skill" - parry is singled out.

I can understand what you are saying, but it also makes excellent sense the other way around.

The way you are thinking of it (parry value used once to determine whether an opposing attack is a shield hit that reduces the damage you take, but not adding to defence skill) a shield does not actually help prevent you from being hit - it merely reduces damage you take if you manage to interpose it between yourself and an incoming attack that would otherwise deal damage to you. The damage may be reduced to 0 (in line with the general protection mechanism) but probably not in case of really strong opponents.

This seems in stark opposition to not only several shield descriptions in-game (the Shield of the Accursed, the Shield of Gleaming Gold, and the Aegis all explain how the one using them becomes harder to hit because it is difficult to focus on the wielder), but also how shields have traditionally had dual use in warfare, both deflecting and absorbing blows (some better at one than the other and also depending on their martial use).

As such it makes excellent sense to have shields' both add to defense (value A) and to determine a range in which there's a shield hit (value B). This requires two different values to really fine tune - and shields have that: the parry value (x) and the defense value (y); in other words x = b, y = A - x

The magical shields you talk about have special effects that fulfil the description. Fear and awe perform the mentioned function admirably. I do not know the description of the Accursed Shield, but I expect it has to do with the horror marking.

As for your explanation about needing two values, it fails to convince me because of the simple fact that shields also have a listed defence value. If they did not, it would be plausible (albeit a bit inflexible), but as is, it is a part of the representational argument for my interpretation. The standard shield has a defence value of -1 and parry value of 4. If shields were intended to add to defence as you say, it makes much more sense to display the defence value as 3, the parry value 4 and only count the parry value once. You do not lose any flexibility but gain a lot in clarity. And you need to do less overall work, if I understand you correctly. First deciding a defence value from which you then subtract the parry value in order to display it does not make much sense when the persons looking at the display are then supposed to add that parry value back in order to determine what is a shield hit.

Independent Light Infantry using a spear: attack 10, defense 13 - including shield +4, and protection 15/8 [S15]. Damage is 10 + 3 + DRN - DRN - relevant_protection at location hit when considering shield protection into the mix.

Defense decreases by 2 per multiple attack and by fatigue/10 and attack by fatigue/20.

Under these conditions and given two groups of statistically significant size of Light Infantry, I most certainly expect a fair number of hits from the very first round they are in melee range of each other, for the difference between attack and defense and even attack and defense+parry is rather small (3 and 7 respectively) and the RNG will have its say in the +DRN - DRN business; moreover, the little light infantrymen won't square off one by one - given an equidistribution over each infantryman over those enemy infantrymen he can attack, it is practically certain that some infantrymen will be attacked more than others (and hence be easier to hit for every attack after the first one) and given the number of size 2 figures per square, some a lot more.

It might come out differently if I actually spent the time on doing a real statistical analysis of the problem but at least my rough mental napkin map (without a napkin ) comes nowhere close to your conclusion that "LI vs LI with spears and shield should not be able to hurt each other at all in the first few rounds if non-shield hits only occurred at above another +4 on top of the 13 defense rather than at above 13.

I was unclear. Sorry about that.

Since we are talking about a decent number of troops, there will of course be some troops that are hurt, but relative to the number it will not be significant. My observation is of course that troops suffer more losses than they should, if parry was counted twice. Beating a difference of 7 is not easy (about 2/25) and though the majority of troops struck in a round will be struck twice (reducing their defence for the second attempt by 2), few will be struck more often than that. Beating a difference of 5 is also not that easy (a bit less than 1/7). Counting on that a rough 2/3 of those targeted are targeted twice, about 2 in 21 are actually hit this way. If we then, to simplify, assume that all hits are on protection 8, the chance to kill someone this way is about 1/73.
Defence drops with fatigue, but so does attack. In effect, relative defence drops by one after three rounds and goes down a notch again after eight rounds.

With fatigue and damage accumulation, deaths will be increasingly likely, but it takes a number of rounds for that to happen.

I will not pretend to have done a complete statistical analysis, because I have not, but the (very) rough numbers above just do not correspond with my in game experiences.

Edit:
For what it is worth http://forum.shrapnelgames.com/showthread.php?t=43615#5 is a post that explains defence and parry as I think it works.
Take it for what it is worth. I have not made an extensive search of the boards, I just took the first one I found that explained it.

[Peter Ebbesen]

What we probably need, should we desire to know how it works rather than rely on our anecdotal evidence, is for us or somebody else to mod in some degenerate shield with an absurd defense value to lower base defense or set it to some convenient total number when considering the attacker for the test and a grotesque high parry value and run a series of structured tests. I won't have time to do this today, but later this week should be an option.

[Amorphous]

Some shield-modding seems to confirm my beliefs.

I modded the machakan hide shield (armour number 105) to encumbrance -2 (to eliminate fatigue from the machakan warriors), defence 20 and protection 0, which ends up as defence 2, parry 18 and protection 0.

With the above, the machakan warrior without javelin ends up with 11 defence not including and 29 defence including parry.

If you were correct, they should be practically impervious to normal troops. In reality they die like flies.

These are the lines that modify the shield:

#selectarmor 105
#prot 0
#def 20
#enc -2
#end

I attach a small mod containing only this change, so that you can try it yourself. Just start a game as Machaka and go to town with the warriors.

[Agema]

Originally Posted by Agema

Not including shield...

So no, I wasn't including shields.

* * *

Shields simply cannot be "double counted" as per Peter Ebbesen, The example in the book with the black centaur should not be 17 Def including shield. I would rationalise it as follows.

Let imagine a Hastatus/Principe is being attacked by a fairly generic unit (10Def, 10Str, 5 damage weapon). The Hastatus has 12 protec armour and a tower shield, 15prot, +7parry. This gives a total Def (including shield) of 15.

To hit the Hastatus at all, the attacker needs a DRN roll 6 higher (10Att v 15Def). That occurs 11% of the time.

If the shield parry is added to Def including shield, then shield hits occur when the attacker's DRN roll is 6-12 higher, which occurs 10% of the time. Then, the attacker needs to beat the Hastatus's protec by 13 on a DRN roll (10Str+5Weapon v 12armour+15shield). That would mean a shield hit has a 1% of inflicting damage.

To hit armour naturally and bypass the shield entirely, the attacker would need to win the Att v Def DRN roll by a massive 13. That's 1%.

Consequently, you'd expect a Principe or Hastatus in combat with a generic infantryman to take damage well under 2% of the time. It is entirely obvious that this is not the case, and they take far, far more hits in melee.

[MaxWilson]

Originally Posted by Agema

I would expect here again over 30 Def is vanishingly unlikely. That will partly be because there are much better ways to equip an SC rather than Def boosters: 30Def is made useless by all sorts of spells like earth meld, tangle vines, paralyze, and so on, and due to -2Def per attack, it quickly degrades if the SC is swarmed by many attackers.

If an attack is stopped by luck/ethereal, it doesn't degrade defense. Lucky, ethereal thugs/SCs with high defense are basically impossible to swarm unless you happen to have troops with magic weapons (like Dawn Guards). That makes high Defense a worthwhile investment, and it doesn't degrade with fatigue as fast as Prot does. (Ideally you'd have both but sometimes you have to choose.)

-Max

[Peter Ebbesen]

Oops, I forgot to thank Amorphous properly for actually bothering to run a test to discover whether the manual was wrong or not on the shield issue (it clearly was; A good test beats anecdotes or revealed wisdom any day of the week).

So, thank you Amorphous.