Putting the 'role' back in role-playing games since 2002.
Donate to Codex
Good Old Games
  • Welcome to rpgcodex.net, a site dedicated to discussing computer based role-playing games in a free and open fashion. We're less strict than other forums, but please refer to the rules.

    "This message is awaiting moderator approval": All new users must pass through our moderation queue before they will be able to post normally. Until your account has "passed" your posts will only be visible to yourself (and moderators) until they are approved. Give us a week to get around to approving / deleting / ignoring your mundane opinion on crap before hassling us about it. Once you have passed the moderation period (think of it as a test), you will be able to post normally, just like all the other retards.

Pillars Combat Analysis-Review

Shevek

Arcane
Joined
Sep 20, 2003
Messages
1,570
I dont see players making tanks as a problem. If they want to go tank and spank let them. Those that want an offensive front line that plays with positioning to set up flanking or whatever can do that too. Some folks say, well, why go through the effort of making an offensive kill squad when tank and spank works? Well, its more fun thats why.

In the IE games, I often just made parties I could select-all attack my way through the game with. PoE is one of the few games where I actively make parties where I try to find ways to synergize ability use with my party members and set up flanking and so on. They have a solid system (admittedly which only shines in PotD) and it is flexible enough to allow multiple styles of play.
 

prodigydancer

Arcane
In My Safe Space
Joined
Feb 16, 2015
Messages
1,399
I do think that in 1.05 ,
engagement takes longer to *engage*, and the ai has been programmed to be much more diligent about
seeking out the back line.
Indeed. I was quite surprised to see post-release AI improvements.
 

Sensuki

Arcane
Joined
Oct 26, 2012
Messages
9,804
Location
New North Korea
Codex 2014 Serpent in the Staglands Shadorwun: Hong Kong A Beautifully Desolate Campaign
As far as numbers go, what deflection can you get your tanks up to in Pillars of Eternity?

120 Deflection is the equivalent of 34 AC in D&D (120 / 5 + 10), so you're essentially traversing into "Epic Level Numbers" by the end of the game (made worse by the existence of the attack resolution mechanics with crits and grazes).
 

tdphys

Learned
Joined
Jan 30, 2015
Messages
168
Location
the event horizon
As far as numbers go, what deflection can you get your tanks up to in Pillars of Eternity?

120 Deflection is the equivalent of 34 AC in D&D (120 / 5 + 10), so you're essentially traversing into "Epic Level Numbers" by the end of the game (made worse by the existence of the attack resolution mechanics with crits and grazes).

I'm too lazy to do the accounting, so I'm going to be really anectdotal here . but yeah, you can really jack up the DEF inordinately more then in pen and paper, so the scaling is pretty crazy. The difference between 4e and Pillars, though, is that you can stack debuffs. Some fights my accuracy is down to 14 starting above 60. That's a -10. I think the maximum debuff on to-hit in 4e was -5 ( combat advantage was -2) . I'm pretty sure you can get similar swings with deflection too.

(from the wiki... might be off)
So with Deflection : Base 25 , Stats (max per, res) 20, Leveling (3*12) 24, Large Shield 24,

Talents adding def:
Cautions attack +10
Weapon and shield +6
Superior deflection ( wiki doesn't say)

So yeah, optimal Deflection gain easily outstretches accuracy gain. The tank is real. Looking at this makes me want to *scream* balance, and do stuff like move per to just accuracy bonus, but really, the key is to neuter the tanks efficiency, and that
requires changes to engagement. I'm half minded to say that you get engagement only by landing a hit on a character. That would be some interesting mechanics. You'd have to juggle acc and def to make things work.
 

Shevek

Arcane
Joined
Sep 20, 2003
Messages
1,570
I think that, generally speaking, just reducing the accuracy and defense from level up is all thats needed to address a multitude of concerns. So, instead of 3 per level, make it 2. Basically, by level 12, it would amount to 11 less ac and def. That would firm up the end game, make tanks need a bit more support, continute to promote flanking and other debuffs, etc.
 

Shevek

Arcane
Joined
Sep 20, 2003
Messages
1,570
Naw, since the bulk of base ac and def is gotten at level 1 and through items/abilities, reducing the effect of leveling makes it harder to trivialize encounters through excess xp. This will be important for them as they add xp in xpacs or currently as players do shit in different order, etc.
 

Sensuki

Arcane
Joined
Oct 26, 2012
Messages
9,804
Location
New North Korea
Codex 2014 Serpent in the Staglands Shadorwun: Hong Kong A Beautifully Desolate Campaign
No, not really. It's basically just moving the bracket. Changes to the amount of defenses per level also affect enemies and reducing numbers would make it easier to beat content above your level, which is I believe, the opposite of your intentions.
 

tdphys

Learned
Joined
Jan 30, 2015
Messages
168
Location
the event horizon
I think the balance comes not in rate of gain, but in the possible difference you can have between accuracy and deflection for a given level, Without considering buffs/debuffs. Removing base and level gains, in percentage, Accuracy approximately ranges from -8 from large shield to +24 ( superb,one handed, weapon focus). Def ranges from 0 to 44+ ( large shield + defensive stats, not considering talents) . So yeah, I think they should just remove the defensive stat bonus to def, or balance it with per getting acc and res def. And remove engagement, and have better encounter design. But that would be nirvana, and we don't want that , do we :).
 

tdphys

Learned
Joined
Jan 30, 2015
Messages
168
Location
the event horizon
It's harder to make a comparison with DnD here, but for 5e, you have a strictly equipment AC values maxing out at 22-24 for full plate, shield and maybe a feat or two, so that's +12 = +60 deflection. Accuracy maxes out at +5 stats, +3 weapon, +5 proficiency, at about the same level +13 . The difference is that critical hit damage is inflationary in Pillars with accuracy rise (it's amortized in DnD) , so that probably means they need to allow higher deflection levels, but it still seems a little off kilter to me. Especially if you're trying to use character level to inform monster level.
 

tdphys

Learned
Joined
Jan 30, 2015
Messages
168
Location
the event horizon
I think there's a decent mathematical base in terms of to-hit and dr. I'd be more inclined to refine then to scrap and redo. Hopefully they see the 45k downloads of the IEmod and take a seriously look at engagement.
 

Shevek

Arcane
Joined
Sep 20, 2003
Messages
1,570
I think there's a decent mathematical base in terms of to-hit and dr. I'd be more inclined to refine then to scrap and redo. Hopefully they see the 45k downloads of the IEmod and take a seriously look at engagement.
I use IE mod too (for fast stealth, ui and xp reduction)
 

Shevek

Arcane
Joined
Sep 20, 2003
Messages
1,570
No, not really. It's basically just moving the bracket. Changes to the amount of defenses per level also affect enemies and reducing numbers would make it easier to beat content above your level, which is I believe, the opposite of your intentions.

Hmm, I wasn't aware that enemy stats were calculated in the same manner as the player. I thought that enemies just had raw values inputed into their creature file not levels etc.
 

Sensuki

Arcane
Joined
Oct 26, 2012
Messages
9,804
Location
New North Korea
Codex 2014 Serpent in the Staglands Shadorwun: Hong Kong A Beautifully Desolate Campaign
Hmm, I wasn't aware that enemy stats were calculated in the same manner as the player. I thought that enemies just had raw values inputed into their creature file not levels etc.

Nah remember Sawyer's whole thing about unified mechanics? :P

Enemy classes are *exactly* the same as the player characters. Creatures have a table where they get some extra base stats but they still benefit from all of the level based bonuses to defenses and stuff like that.

That's why I'm saying that some of the reason why encounter design sucks is because the designers are limited by the system. In the IE games, they could make stuff that broke the rules, which in turn was fun/gave a little spruce to stuff that might have otherwise been bland. Designers can't do that in Pillars of Eternity. The only things they can do are place creatures and place some of the loot (Josh handled/oversaw the random loot table stuff). Perhaps they were even prevented by their producers from doing unique creatures/bosses because 'too much work'.
 
Last edited:

AwesomeButton

Proud owner of BG 3: Day of Swen's Tentacle
Patron
Joined
Nov 23, 2014
Messages
16,334
Location
At large
PC RPG Website of the Year, 2015 Make the Codex Great Again! Grab the Codex by the pussy Insert Title Here RPG Wokedex Divinity: Original Sin 2 A Beautifully Desolate Campaign Pillars of Eternity 2: Deadfire Steve gets a Kidney but I don't even get a tag. Pathfinder: Wrath
By now it seems typical of this game that they should claim "we're following a principle of unified mechanics" while in reality they are doing something not for gameplay or balance reasons but to conserve time.
 

Mangoose

Arcane
Patron
Joined
Apr 5, 2009
Messages
25,136
Location
I'm a Banana
Divinity: Original Sin Project: Eternity
My theorycrafted opinion is they needed a moderate number more Engagement-breaking abilities like the Rogue's Escape, Barbarian's Wild Sprint, and Monk's... whatever. This would allow for in-combat repositioning and in turn necessitate responsive repositioning. It may also give the player more to think about for initial formation.

Basically these Engagement-breaking abilities are a real-time emulation of the turn-based 5-foot-step, which is necessary when there are consequential AoOs. On that note every class should have at least one, though not necessarily in equal amounts or power.

For example this may necessitate a strategy like having mid-liners (like Rangers or something) to "catch" enemies that get past the front-line, and also make nuanced positioning important to reduce gaps (think Blood Bowl). Or even funnel enemies through an opening on purpose. Or perhaps you'd choose/develop front-liners that have better movement or disengagement capability rather than purely defensive "tanks" allowing the front-liners themselves to "catch" those same enemies. Or on the other hand you can design a very maneuverable party focused on getting around Engagements to get to the back line. Even then you may make your backliners maneuverable so that they can handle themselves without help.

I don't think decreasing the effectiveness of Disengagement Attacks is the right answer because I think it can be a good mechanic if implemented properly in that they are made avoidable if the right decisions are made in combat.
 

tdphys

Learned
Joined
Jan 30, 2015
Messages
168
Location
the event horizon
My theorycrafted opinion is they needed a moderate number more Engagement-breaking abilities like the Rogue's Escape, Barbarian's Wild Sprint, and Monk's... whatever. This would allow for in-combat repositioning and in turn necessitate responsive repositioning. It may also give the player more to think about for initial formation.

Basically these Engagement-breaking abilities are a real-time emulation of the turn-based 5-foot-step, which is necessary when there are consequential AoOs. On that note every class should have at least one, though not necessarily in equal amounts or power.

For example this may necessitate a strategy like having mid-liners (like Rangers or something) to "catch" enemies that get past the front-line, and also make nuanced positioning important to reduce gaps (think Blood Bowl). Or even funnel enemies through an opening on purpose. Or perhaps you'd choose/develop front-liners that have better movement or disengagement capability rather than purely defensive "tanks" allowing the front-liners themselves to "catch" those same enemies. Or on the other hand you can design a very maneuverable party focused on getting around Engagements to get to the back line. Even then you may make your backliners maneuverable so that they can handle themselves without help.

I don't think decreasing the effectiveness of Disengagement Attacks is the right answer because I think it can be a good mechanic if implemented properly in that they are made avoidable if the right decisions are made in combat.

The problems of engagement comes from the abrupt and incomprehensible way in which it disrupts character and ai movement. Something which can affect movement so strongly should be an active power ( or linked to one, like attacking) and be limited thereby. I'm all in favor for meaningful disengagement attacks that get the power and bonus they were meant to receive. Because of the poor engagement mechanics, disengagement attacks were succesively nerfed in the betas, to the point that Josh pointed to it in a forum response somewhere.

The hardest thing about engagement is the geometric problem of when to give it, and when to proc the attac. If you wait for the target to leave range, but the attacker follows it, then you never proc. If you proc on the slightest movement, you essentially disable in combat movement... and get the ensuing ai-nightmare when a single target moving generates a flurry of instantaneous attacks and madness. I'm half convinced that the solution is tracking the distance an engaged character has moved away from its targeting enemy and then rewarding when that distance = some modifier times the targets reach. I still think engagement should be rewarded only by a successful hit/crit (maybe graze). Currently, It seems that engagement is rewarded by just having an enemy in range for a certain amount of time. This is way too passive, opinion, though you can shift engagement by targeting another enemy in range. It's all really confusing and haphazard. They should just stick with the targeting mechanic.
 

tdphys

Learned
Joined
Jan 30, 2015
Messages
168
Location
the event horizon
That being said, I'm now at the "illithids" in Od Nua, and a lot of Od Nua has provided the most interesting combat/encounters to date on POTD. If anything, I'm enjoying it more than BG2 at this point, though for different reasons.
 

Mangoose

Arcane
Patron
Joined
Apr 5, 2009
Messages
25,136
Location
I'm a Banana
Divinity: Original Sin Project: Eternity
My theorycrafted opinion is they needed a moderate number more Engagement-breaking abilities like the Rogue's Escape, Barbarian's Wild Sprint, and Monk's... whatever. This would allow for in-combat repositioning and in turn necessitate responsive repositioning. It may also give the player more to think about for initial formation.

Basically these Engagement-breaking abilities are a real-time emulation of the turn-based 5-foot-step, which is necessary when there are consequential AoOs. On that note every class should have at least one, though not necessarily in equal amounts or power.

For example this may necessitate a strategy like having mid-liners (like Rangers or something) to "catch" enemies that get past the front-line, and also make nuanced positioning important to reduce gaps (think Blood Bowl). Or even funnel enemies through an opening on purpose. Or perhaps you'd choose/develop front-liners that have better movement or disengagement capability rather than purely defensive "tanks" allowing the front-liners themselves to "catch" those same enemies. Or on the other hand you can design a very maneuverable party focused on getting around Engagements to get to the back line. Even then you may make your backliners maneuverable so that they can handle themselves without help.

I don't think decreasing the effectiveness of Disengagement Attacks is the right answer because I think it can be a good mechanic if implemented properly in that they are made avoidable if the right decisions are made in combat.

The problems of engagement comes from the abrupt and incomprehensible way in which it disrupts character and ai movement. Something which can affect movement so strongly should be an active power ( or linked to one, like attacking) and be limited thereby. I'm all in favor for meaningful disengagement attacks that get the power and bonus they were meant to receive. Because of the poor engagement mechanics, disengagement attacks were succesively nerfed in the betas, to the point that Josh pointed to it in a forum response somewhere.

The hardest thing about engagement is the geometric problem of when to give it, and when to proc the attac. If you wait for the target to leave range, but the attacker follows it, then you never proc. If you proc on the slightest movement, you essentially disable in combat movement... and get the ensuing ai-nightmare when a single target moving generates a flurry of instantaneous attacks and madness. I'm half convinced that the solution is tracking the distance an engaged character has moved away from its targeting enemy and then rewarding when that distance = some modifier times the targets reach. I still think engagement should be rewarded only by a successful hit/crit (maybe graze). Currently, It seems that engagement is rewarded by just having an enemy in range for a certain amount of time. This is way too passive, opinion, though you can shift engagement by targeting another enemy in range. It's all really confusing and haphazard. They should just stick with the targeting mechanic.
No, the problem with engagement is that the current implementation is way too static. That's clear as fuck.

Yes, you do want engagement to disable freely moving around in combat. You need to have the ability to out-position (out-think) the opponent and make it punishing, and the opponent should be able to do the same to you. Just like AoOs in tabletop. However there also needs implementation of ways to break engagement with little penalty, a la the 5-foot-step. That is how to preserve the dynamism in that you are competing for position throughout an encounter.
And more active shit in a real time game is shit. Micro is not fun. And yes, it is with pause, and more active shit means more micromanagement of pausing. Passive is good. Then when necessary you make an active decision in order to change the "passive" status quo.
 

tdphys

Learned
Joined
Jan 30, 2015
Messages
168
Location
the event horizon
I don't think rewarding engagement based on hit mechanics is unnecessarily adding to micro. My point about reducing engagement to an active ability was more so that it could be "limited" not infinite. Are you not suggesting a 5ft step option as an added "micro" to evade engagement? There's unlimited engagement, there's a lot of trash mobs, they . are . going . to . get . you. ( and engage you ) . I don't suppose you want unlimited 5ft step buttons?

The problem with most of the actual in game engagement evades , is that the AI maintains its target and re-engages is short order. This isn't because of the power itself, but because engagement is so easily gained. The rogue switch places, for example, is horribly situational because of this. There's no real mechanic that's going to subvert the rush of oncoming trash mobs and their engagement freeze rays. You can't fight running battles in this game, that's for sure.
 

Johannes

Arcane
Joined
Nov 20, 2010
Messages
10,556
Location
casting coach
No, not really. It's basically just moving the bracket. Changes to the amount of defenses per level also affect enemies and reducing numbers would make it easier to beat content above your level, which is I believe, the opposite of your intentions.
Making it easier to attack higher-level enemies, and harder to steamroll lower-level ones, would be an improvement all things considered. If some encounters felt then too easy, you could give them an extra level. I see that as a definite improvement, shifting the focus from raw numbers to tactics a bit. Even if that'd hardly make the combat interesting by itself, but it'd have a nice effect on the game by making the progression between areas a bit less linear.
 

gestalt11

Arbiter
Joined
Apr 4, 2015
Messages
629
As far tanks go:

I killed the sky dragon in 30 seconds with no knockouts (and cleared od Nua 1-13) with a team that was completly DPS based on hard using:

1 priest of skaen
1 chanter
4 rangers

These guys all had min resolve and min con with high int, might and max perception/dex. The whole team was built to use interrupt. While I am not gonna say this setup had no tank, because 4 animal companions + priest makes a pretty good meat shield it did have absolutely no character made as a tank. I am not sure you can really play this game without either massive aoe/cc or some kind of meat shield but you are not confined to requiring a player made tank of some sort, although its probably the easiest and most effective way.

As for counters:
One thing the above party showed me is that the counter system in PoE is just plain flawed. When I killed Undead Raedric with this party countering the fampyr charms is essentially impossible even though I have a priest. I buffed the +50 defense against charm spell and it does absolute jack didly. Why? There are multiple reason:

1) every single fampyr targets my priest everytime and does so immeidately
2) grazes still charm you
3) the charms are so fast you can't actually interrupt them even with an entire party with excellent interupt

The chances of 4 charms all missing and not grazing a character with even defenses as high as 100 (which is what +50 on a medium Will character is (high int low resolve)) is extremely low. For a fucking 6th level spell that is completely fucking ass.
 

Shevek

Arcane
Joined
Sep 20, 2003
Messages
1,570
I think grazes for things like paralyze should give lesser effects (like slow or stuck or something) instead of just being half duration.
 

gestalt11

Arbiter
Joined
Apr 4, 2015
Messages
629
My theorycrafted opinion is they needed a moderate number more Engagement-breaking abilities like the Rogue's Escape, Barbarian's Wild Sprint, and Monk's... whatever. This would allow for in-combat repositioning and in turn necessitate responsive repositioning. It may also give the player more to think about for initial formation.

Basically these Engagement-breaking abilities are a real-time emulation of the turn-based 5-foot-step, which is necessary when there are consequential AoOs. On that note every class should have at least one, though not necessarily in equal amounts or power.

For example this may necessitate a strategy like having mid-liners (like Rangers or something) to "catch" enemies that get past the front-line, and also make nuanced positioning important to reduce gaps (think Blood Bowl). Or even funnel enemies through an opening on purpose. Or perhaps you'd choose/develop front-liners that have better movement or disengagement capability rather than purely defensive "tanks" allowing the front-liners themselves to "catch" those same enemies. Or on the other hand you can design a very maneuverable party focused on getting around Engagements to get to the back line. Even then you may make your backliners maneuverable so that they can handle themselves without help.

I don't think decreasing the effectiveness of Disengagement Attacks is the right answer because I think it can be a good mechanic if implemented properly in that they are made avoidable if the right decisions are made in combat.

The problems of engagement comes from the abrupt and incomprehensible way in which it disrupts character and ai movement. Something which can affect movement so strongly should be an active power ( or linked to one, like attacking) and be limited thereby. I'm all in favor for meaningful disengagement attacks that get the power and bonus they were meant to receive. Because of the poor engagement mechanics, disengagement attacks were succesively nerfed in the betas, to the point that Josh pointed to it in a forum response somewhere.

The hardest thing about engagement is the geometric problem of when to give it, and when to proc the attac. If you wait for the target to leave range, but the attacker follows it, then you never proc. If you proc on the slightest movement, you essentially disable in combat movement... and get the ensuing ai-nightmare when a single target moving generates a flurry of instantaneous attacks and madness. I'm half convinced that the solution is tracking the distance an engaged character has moved away from its targeting enemy and then rewarding when that distance = some modifier times the targets reach. I still think engagement should be rewarded only by a successful hit/crit (maybe graze). Currently, It seems that engagement is rewarded by just having an enemy in range for a certain amount of time. This is way too passive, opinion, though you can shift engagement by targeting another enemy in range. It's all really confusing and haphazard. They should just stick with the targeting mechanic.

IMO as someone who has done various forms of combat-ish sports (wrestling, karate, brazilian-jiujitsu) the real problem with the engagement system is that it overlooks how movement in a fight actually works. It tries to do something good which is to model that simply ignoring an opponent is a deadly thing to do which is why you almost never want to show someone your back.

But it absolutely fails to model actual movement. In a one-on-one fight people don't stand still and trade blows. They dodge and weave and circle (unless they are in formation). If I start to circle then my opponent will also move, if he does not he will almost certainly give me an opening. Similarly if I back up a step they will often move.

Controlling the angle and distance of a one-on-one fight is controlling that fight. Movement provokes movement and only idiots or incompetents stand still. Footwork is a fundamental thing in fighting, easily one of the top 3 most important things.

The engagement system models incompetent combatants, it essentially assume you turned your back and ran which is a flat out stupid thing to do. In the engagment system I literally can't back up a step, which is stupid. If I am engaged in a sword fight someone they literally cannot stop me from backing up. I have to be careful and I can't do it by sprinting but I can back up and they will either follow me or disengage. At best an engagement may stop you from moving forward, but even that is iffy as you could circle. Certainly being engaged means you must move slower than a run or maybe even a walk.

Not only can an engagement system not realisticly prevent sideways or backward movement but that sort of movement should actually drag the opponent with them (if he wishes to continue to attack). So while the guy initiating a backward movemnt might lose an attack sequence, he should not be penalized for "disengaging" (and you could even have some sort of ability to say move back and then do a lunging attack if the opponent follows you). Similarly the opponent could follow and not lose an attack sequence or opt not to follow and break the engagement.


So while I laud it as an attempt at something I think should exist instead of some dumb "aggro" mind control thing the model itself is fundamentally flawed since it make the characters behave in what would be an incompetent manner and fundamentally makes fight behave in a way even amateur feel in some manner is "wrong". Note I am not saying the system should be realistic simply that the model should follow some semblance of reality. When you look at a model and the result of following that model is that the actors behave like idiots its simply not that compelling.

Unfortunately most developer's simply don't understand the intricacies of this sort of stuff. They are not to be blamed for this. Mastering this idea in practice is something that is above Black belt level in a system like karate. This is like 5th dan level stuff. But its also considered the absolute heart of a fight as well.

Combat is a flow of things, anything that forces it into something "static" is goning to feel "wrong" or "stupid".
 

As an Amazon Associate, rpgcodex.net earns from qualifying purchases.
Back
Top Bottom