Larian Studios Forums Divinity: Original Sin 2 Divinity - Original Sin 2 - General Response to Swen Vincke PCGamer Article

So I was helping my girlfriend with statistics for last couple hours, but I'm back!!

@Limz
My Chess analogy was loosely based. I said as much.

First, there is no such thing as live combat not having elements of chance.

Secondly, my entire point has and will continue to be: the lack of elements of chance create less pressure and take away from the fact that this is "combat" and thus detracts from the feelings of enjoyment of playing a tactical turn based combat game would/should evoke

Thirdly, the entire point wasn't that there's an objective measure of improvement for the game between the two games but that my opinion, and those of others, don't find this as enjoyable as it could be. Thus my and others *opinions,* as part of the player base, do in fact matter in an EA all about feedback in order to make a game more fun for all.

Fourth, I dont feel a need to dig into information gathering and pressure created through chance per turn given it's pretty self evident and me, you, and most others can follow that to it's logical conclusion on their own.

Fifth, the assumptions are made to simplify the problem....obviously. It should also be pointed out that we're all making assumptions since known can really test case all of this at the moment unless they can magically get ahold of both future versions of this system and other systems.

@SlamPow
While I don't think things would be as bad as you think, I thank you for agreeing that testing a new system would be worth it.

Also, I have about 300+ hours in D:OS Classic and EE (about 120 in classic and 180 in enhanced). So, I do have experience with the metas and so on. CCing in general was over powered and comboing in general while my mage has 30+ spells brought things down. Always had the perfect tool for the job...which is why I agree that the Memory system has a place here in the sequel. I have about 20 hours in the sequel, but I've never been a fan of pubs so you have me there. I have gone out of my way to see all the different combos and builds people can do though.

As for AI, I actually found what you said highly interesting. I wonder about the Starcraft AI and how much time it mustve took to hand craft it. That's not really adaptive so much as having a plethora of good advice shoved into it's head, it sounds like. Still a beautiful thing though to hear about.

You being up a good point of AI just needing to be challenging. That's doable enough. I just like thinking in extreme use cases cause I'm a minmaxer that'll spend hours in town just making sure I have every advantage before a fight and deliberate between what move to take for maximum efficiency. I'd like it if the AI could still push me to the limit no matter how good or how hard I plan. I'm masochistic like that in video games. My friends think I'm crazy....I think I just love a good challenge. I also don't like the idea of holding myself back to create that challenge...rather than a challenge run, I like putting the difficulty to max on my first run through in order to experience the real "challenge" that game can best provide.

As for me underestimating AI, I am aware how good RTS AI have gotten (though I didn't know about the handcrafted Starcraft one specifically) and I'm generally aware or real time turn based games and strictly turn based games. Everything I've seen and read though points to the limiting factor of an AI to think outside the box and employ high level strategy on the same level as a player when it can't search an entire tree for the best sequence of events. Intuitive and creative decision making are the hallmarks of a human player vs a computer after all; there are some new algorithms to help this along but they're all currently being delved into more. Perfect AI will come one day....I just don't think so for this game. You're Neverwinter Nights 2 was very lovely to read about though....and that may in fact be a good way to do things. I hesitate to equate the two though. Dragon Age type games aren't the same as D:OS after all. I'd say Civ is a closer comparison.

My experience lies more in the realm of either super deterministic games like chess or a text based RPG and GO in one case years ago....and more recently in the Fighting Game ICE environment. Specifically, wanted to create an adaptive AI that geared itself to defeating a specific opponent instead of being more general and had it limit itself based off shown player level -> ideally it would progressively get harder to fight it as it concluded your skill level and what would work best against you. Used MCTS for "ideal" move instead of optimal move, a KNN player model to predict future player moves based off past actions, and some scripts in the same vain as Greedy Portfolio Search (which is what I think the Starcraft AI probably uses since it's very good for RTSs given specific information concerning strategies and counters). We had limited testing time for the project but the results were very promising XD

Again, my issues don't lie with the AI really. The foundation lies with the fact that super deterministic combat such as this in an RPG where combat is limited to a focused group vs group based tactics against a non-human opponent isn't that fun in my opinion. Heck, even Civ incorporates chance with the maps, barbarians, and city-states and some AI behavior and interactions. Elements of chance are good things....Having everything under the sun being deterministic isn't all that great to my eyes and actually detracts from game play. Hell, even in fighting games at least you always know what the human player will do and I go out of my way to be more random for fun. Super deterministic games against other humans is fine cause the human itself creates constructive randomness...not so much with a computer opponent; or overall it just doesn't feel the same when they do manage it to a degree.

We're all allowed to have opinions of course. There's no objective value for fun. Which is why I like the idea of combining systems and tweaking that: give us more options over all and try to satisfy both parties.

Also, if mages are having trouble keeping up wit other "classes" than the answer lies not in their damage or their ability to have more guaranteed CC. Every archetype has a role. If one asks what the role of the wizard is than the answer I'd give is flexibility and battlefield control. If others can do that role just as well or better than we have an issue. Maybe we should encourage warriors to invest more in multiple stats so that they don't have as much utility as straight mages who only need INT an MEM. Maybe we need to change combat abilities to make them more costly to and desirable to specialize (I think this needs to be done already). Maybe we need to decrease the CC abilities of warriors and increase mage CC options. There;s a few things we can do, but I wouldnt label that as anything to do with the armor system itself and more to do with how the "classes" currently work.

Yeah, I suppose we do just have to agree to disagree. Can't wait to see if they do something similar to what you suggested, though! It would be interesting to see how people evolve and adapt to it, and I certainly would love to test it out. For what it's worth, I am a dirty, filthy Drain Willpower lover, so it's almost hypocritical of me to criticize that skill. Also, I'm really glad to have piqued your interest! You've certainly piqued mine. I hope that you find your challenge in Honour Mode, my man. I am looking forward to testing myself as well.

Also, I'd like the point out that the issue of 1 vs 1 combat with an AI does indeed devolve into who can get rid of the others armor first.

Group based combat can and will eventually mitigate this with good AI, but ultimately an opponent focusing on keeping a character alive cause of lack of armor is an opponent playing defensive while I'm playing offense.....and well know how much better offensive playing is vs defensive in D:OS games. Quicker you kill the enemy one by one, better chance you outright win.

There's also the fact that good, patient players using teleport and sneak and so on can turn what would be planned team fights into mini 1v1 or if I'm running a group 4v1-2 fights....in which case it becomes quite easy to focus down and destroy armor much faster than the AI will with a good build.....at which point I just CC them to death.

It essentially turns matches into resources management games where the goal is to make the opponent use up his. When you out number the opponent, this becomes super easy. When the opponent outnumber you, this becomes much harder. When you have the same amount, this becomes a matter of how you approach the problem. But given how creative players are......(ie set up barrels before hand and throw grenades while hidden out of sight before combat).......and the fact that players will always have the advantage it just how much freedom they have and what they can do......I just see lots of issues with balance down the line o.O Many of which relies on specifically targeting the advantages and exploits players can bring to bear. Some of this can be mitigated with "immune" enemies...but I'm not really a fan of that as it just means gear towards what they're not immune to.

Anyway, yeah. SlamPow, I do think we have fundamentally different opinions. I understand where you're coming from with wanting your skill level to be the ultimately determining factor. I just think the game environment itself gives players so many tools and options to leverage skill level that highly skilled players need to be knocked down a peg or two to increase difficulty and force them into "oh shit, what do I do now?" moments.

This is a fair analysis. I just attribute being able to 1v1 things to a flaw in the AI, which, again, I'm hoping they'll fix. But yeah, I would indeed prefer that. I too hate immune enemies, though; they feel cheap. Also, I vastly prefer doing the game with one character; I'll probably keep this trend up even once the game comes out. I too can see reasons why this would be hard to balance, but I believe in Larian. Hopefully, tactician mode will be just hard enough to scare us =D

Welcome back.

You fed me a boatload of white noise that really doesn't further the conversation. You're flat out wrong on a lot of those points and I am not going to write my refutations of all those points unless you want me to (because it bogs down discussion). So, I'll cut to the chase and keep things short so all parties can keep up.

Increase the resources it takes to reach the terminal velocity of any given encounter causes an increase in weight for resource management and allows for more points for information discovery.

By having an ability pool that is seeded per game generation on and a resource pool that is seeded upon an encounter that accepts parameters such as player performance will keep difficulty consistent and diverse without changing combat function.

The short of it: Information hiding/discovery needs to be king.

Thanks for the welcome

If you'd like to point out how I'm "flat out wrong" you can put it in minimized spoiler tag and I'll respond to those points in the same manner. I agree that bogging down discussion isn't good and those points aren't really directly relevant anyway, as you said. But hey, you did ask for said "white noise" when you asked for why I brought in chess, so no apologies ;p

But onto the meat of it: So seed the resource pool and abilities based on player performance? Thus making information discovery/hiding king? Since we won't know the abilities of a given opponent? And greater resources mean bigger challenge and places greater importance on information and allows the player to discover more in a given encounter cause it lasts longer?

This certainly has some merit, maybe. And increasing the weight of information is also a very good suggestion I fully endorse.

The problem is that once we reach terminal velocity though. It basically signifies the end of an encounter once that velocity is reached. The problem I have with this is that resource isn't even health related though. Why have HP on an enemy if it ultimately matters very little in the end?

The only good argument I can think of for that is that a player is then spending time finishing one enemy off while the others get a chance to attack. But this is then mitigated by my above points on player leverage to turn group based fights into smaller encounters.

There's also the fact that the only way to counter player leverage when being creative (ie barrels) is to then massively increase resources to negate said leverage later on. Would then the AI seed such resources even in encounters that don't allow for as much leverage? That'd create problems of it's own and now you're discouraging creativity or enforcing it later on.

More information weight would also be better though. Hiding how much magic/physical armor an enemy has would be a step further towards this.

EDIT: Oh, and there's still the fact that I can divide and conquer by choosing how many enemies to fight at once with patience; effectively limiting opponent access to how much of a resource they can have. That doesn't always work out, obviously. But it works a lot more than I'd think they ever intended.

You can express that in an exponential curve; creative solutions are weighed as consuming a lot of resources so the AI won't be granted any extra resources but repeat exploitation leads to a sudden shift until resources normalized.

Furthermore, things like sneak need to be looked at as resources for the AI as well, the AI might have one stealth patroller this time but next time it might have 3 or other ways of detecting stealth.

But really the AI should just go hostile the second you start moving extra barrels around the vicinity.



Hiding information such as how much armor and health they have in higher difficulties but displaying how much total damage that unit has suffered would be great.

Recall that the question being posed through all of this is: Given you have no moving parts how do you make a game challenging through multiple play-throughs and guides? My answer to you is that information entropy must be king.

The implementation details matter less, it's more of me challenging you on the idea that a game cannot be rich and complex through multiple runs without system built entropy such as 'saving throws' etc.

In the end, all systems have their weaknesses, you make compromises based on what you desire out of it. The answer to some exploits might just be to ignore it if there is a certain experience you are aiming for.

I'll also answer the other points you made earlier at later time tomorrow or the day after since I really need to get some coding done.

I guess we didn't play the same game.
In my game exploitation is practically a central game mechanic.

There's lists of exploits already but the one where you surround the enemies with barrels/crates before combat generally suffices :P

This was the case in DOS classic and EE as well. I really hope they change this; stealth, barrels, stealth grenades, Far Out Man cheese, high ground abuse, it's all too prevalent, like you say. But, like someone else suggested in another thread, I think it would be pretty cool if these were not eliminated, but given to the enemy to be used against us. Imagine walking into a room, and starting a conversation, then as soon as you press end, a flashbang comes out of nowhere and blinds you! Or, as you are approaching a distant encounter, shots start raining down from a crazy distance, chipping away at you a couple of times before combat starts. Or if enemies actually used barrels. Or had stealth of their own. May not be fun, except in tactician mode, but boy oh boy would I like to see it happen =D

Yeah, stealth is also pretty broken since you can sneak past encounters etc and actually ambush enemies which are supposed to be ambushing you..
Which I suppose is fine, except it works out VERY "not as intended" since invisible enemies spawn in right next to you and get slaughtered.

Also sneak/snipe = invincible
Etc etc.


As interesting as it would be to have the AI use the same exploits as you, I think it would feel "not as intended"

Fair enough. The strategies as they work right now would certainly need some tweaking on the numbers involved. But I would like to see the AI learn how to press "c". And use snipe from it. That just sounds so basic to me. It literally pains me when I see an AI use an unstealthed Snipe.

^^ SO MUCH THIS ^^

+10

No one wants to remove the new armor system. Most of us like it. The problem lies in why an HP bar, if it doesn't really matter. And that the new defensive combat abilities (magi armor, physical armor, and vitality) don't really appeal much at all in comparison more useful abilities. These appear in conjunction with other problems (such as the sword and shield issue) and thus some of us have discussed the interest of combing systems. Armor and armor skills would be their but the actual abilities would be changed to the old ones. This way CC would not be 100% (if you level up the ability) when there's no armor and blocking would make a come back to make shields relevant again. CC is too strong in some of our opinions.

Limz has brought some interesting points to improve the current system as it stands. It leaves some holes but is workable solutions in a way.

SlamPow is interested in trying the new system but has issues with systems that include chance in such a way that it can be used to hard counter skilled tactics.

I obviously, along with a few others, want a combined system or to get the chance to try one at the very least.

Ressing this more or less dead thread in the hopes that Qioux and other people show up so I can have the pleasure of denigrating their existences while offering legitimate solutions all the while failing to get any work done because I enjoy being a forum warrior more than I do coding at times.

And we have unfinished business here.



Depends on how abstract you want to get, but I think it's more agreeable that we focus on the more immediately relevant examples or just clarify what 'elements of chance' actually means. There's a difference between external entropy and internal/system entropy. For example, a game like Starcraft has no internal entropy but fuckloads of external ones, same goes for Street Fighter V, but not Chess. That lame ass game is mapped the fuck out.

So you basically need to qualify it and not just say 'chance' because you may be conflating it with something else.



Elements of chance can also reduce tactical combat to the point where it becomes unenjoyable because your actions are largely meaningless. Thus, my entire point is that neither system is necessarily/inherently favorable to the outcome of 'rich blah blah depth'.



Not all feed back is racially superior. Some are born with genetic gifts that must be harvested. My feedback is such. Mainly, there are diminishing returns in the quantity and quality of feedback in relation to actionables. The game is also deliberately selecting against certain audiences (noobz who want this to be all real time).



The reason you explore the system is to ensure that your suggestions are in line with the intended design goals while being able to accurately extrapolate and understand the relationship between effort vs change vs whatever. Tweaking an axiom, as a general rule of the thumb, is much harder to predict than working with the context.

Since Limz wants me to I'll bite: *ill get to the above points later, though since I'm at the gym and only have time for this real quick*

So someone else gave feedback on armor system and I liked his take:




Threw in stuff on the stat system as a bonus to catch more fish ;P

Ah got the racial thing too....ah well...double bonus!

Dude, you don't even lift. You know all the forum warrior abilities scale off of strength not finesse right?

Bah, I enjoy staying attractive for my girlfriend lol and she likes swimmer types

Also! I lift my fingers to type and some times I even have to move for food and drinks!

Argueing about the current stat system seems to be pretty pointless at this time, because the current system is bugged and not working as intended?

I want to thank SlamPow for offering a pretty insightful bit on game AI.

Now, while I can openly admit to being intellectually inferior, I do have one perspective about the game that I would share:

If we're talking about the single player, the campaign itself, then the goal of the game is not to offer the player a strong set of mechanics to toy around with. The game has to deliver on several aspects, mostly focused on the story it presents via the medium it uses.
So, I would argue that the difficulty of each fight should be in fact tailored to the context of the fight.
So, for instance, we're fighting a pretty militarized organization. The context here would be that they should be good at team tactics, not necessarily great individual combatants.
Another case is the voidwoken. Unless we're talking about a rather smart void commander, they should be largely disorganized, but individually strong.

I can agree that in the arena mode, things should be balanced around a proportional result to the understanding and use of systems of each player.

My idea would be to make a good balance in arena and then translate that into single player with contextual adjustments. We already have an easy contextual adjustment via stats, but the adjustment I am most interested in is in the AI that runs for each separate encounter. But I am by no means anything close to a game designer, so I may be spouting worst practices for all I know.


Separate from that, I want to note that even if you use elevated language, calling someone stupid is still an insult, particularly when the target is not necessarily stupid, but misinformed/misguided. Insults do not make one a better person.
Yes, I do prefer it when there's less hostility in the air. Not saying that as part of the green name to the left, as I do not moderate, I clean spam.

Well, actually it's working mostly as intended so you can in fact judge it.

The reception was just so bad larian already says they plan to change it but ultimately it's still gonna be the same basic system just given fresh paint