Larian Studios Forums Divinity: Original Sin 2 Divinity - Original Sin 2 - Suggestions and Feedback On the Topic of Initiative...

Have all the mobs have a variance in their initiative that is determined at the very beginning of a chapter / zone. Though meh.

There is an argument to be made that randomness doesn't exist in the universe. Though I'm sure quantum physicists pretty strongly disagree.

Which only brings us back to the original thought that "save-scumming" is therefore not an applicable argument against RNG mechanics.

Hell, and the seed mechanics just discourage, they wouldn't even stop it.

I offered one viable solution and immediately dropped it in favor of seed saving. Why have people felt the need to focus on the forum?

As for RNG not being random....well, yeah. But the RNG in D:OS felt plenty healthy enough to me so I really don't see why someone's ranting about it or what it is in programming (as a prgrommer, I appreciate the trouble of explaining it for others)

Now then, no one has offered a viable reason why the RNG systems of D:OS would be bad, per se, aside from "save-scumming" but everyone doesn't want to stop that either.

You can't complain that save scumming can break a mechanic and in the same breath complain about not using a mechanic because of save scumming when you refuse to give up the former.

RNG mechanics were perfectly fine for
The most part in game one. Only problem was that players abused the shit out of it. Thus seeding can be used to discourage it. Devs don't have to account for a determined player save-scumming, just the average one.

Now, you can say that a mechanic can be changed without touching save scumming. They did: the new armor system and the competitive initiative.......both of which fail to create a truly comprehensively enjoyable experience. Which is why these discussions take place anyway

Now I'm not even advocating turning everything into RNG, but there are a few things that could benefit from it:

Bodybuilding
Willpower
Accuracy and chance to hit
Dodge chance
Initiative roll
CC chance

People loved D:OS and the RNG there worked very damn well; the only complaints I saw were those who hated it when they got unlucky in a roll or those who oppose all things having any element of chance. Neither of which is really a legitament complaint.

Now, we're seeing for a fact that an ultra deterministic system has its own gaping flaws.

Solution? Balance of the two.

As it is, I think the randomness in D:OS2 is overall balanced, but that's because it is usually restricted to attacks, chances to dodge/parry and criticals and most importantly, they don't rely on seed saving (which, as said, can be circumvented). Making initiative based on randomness?

Character 1: Initiative 9, roll d20, result: 2. Total: 11.
Character 2: Initiative 2, roll d20, result: 14. Total: 16.

Character 2 acts first without having ever bothered with spending points in initiative, while character 1 is headbutting the computer screen for having spent points in initiative. I'm using "character" in a very loose sense of the word here. During fights, the player's party is usually at a numeric disadvantage, which means that enemies have more chances to score higher initiative rolls, thus acting first, despite the player having bothered with spending points in initiative. Doesn't seem 100% fair to me, as much as having skeletons and zombies with 14 initiative does.

Edit: this argument of mine only stands with the game in its current iteration; I'd have nothing against returning to the (sadly gone) old days of bodybuilding, willpower and CC chances; in fact, I'm kinda wondering why were those mechanics changed in the first place.

Seems fair to me.

Laws of mathematics dictate that as long as the character with 9 initiative is getting a much better "average" over time than he's being rewarded for his input.

Highs and lows are expected in such a system. Even an almighty God Dragon can roll a 1 in D&D and fall on its face for "reasons" And that's part of the appeal

Averages rewarding player stat choice are what counts here overall. The mathematical averages should be what's argued, not outliers. And that just comes down to which algorithm is used.

You said yourself that the RNG works very well for some things. We don't have to make it the standard for everything but I vote that most of all D:OS EE was healthy for the most part and beneficial to the experience

Hell, if you want to be more forgiving as a dev? Alter the weights on rolls for a character with higher initiative such that his minimum rolls feel significantly better than his counter part. It's smoke and mirrors to give the illusion that someone's initiative of 2 simply has a very very very low chance of ever beating 9 cause we make it so, but that all comes down to how you decide implement things

I can see such an implementation making examples like the one you have extreme outliers and thus naturally moot. Hell, if there's a one percent chance of that happening relatively and it does, that's a unique and fun experience in its own way

Edit:
I too miss the old mechanics. Which is where I'm coming at this from.

As for seed saving....well, I'm not particularly a fan of it, but if people feel the need to absolutely discourage breaking RNG mechanics somehow and maintain save scumming, the seed saving mechanics are the only viable middle ground I can think of

We're kinda running in circles here, friend. I'm not against randomness in initiative rolls OR initiative rolls for that matter.

My main issues are that
A) initiative rolls still are disadvantageous to the player in a PVE setting where there usually are more foes than friends, especially if the initiative of mobs stays the same (once the initiative of foes is toned down, though, it would be alright, and initiative rolls would be very good in PvP) and
B) I'm against seed saving because it kinda kills randomness, especially if it's followed by restrictions on save games to reduce save scummings, and most importantly, because seed saving can be circumvented.

Were initiative rolls to be implemented with a reduced mob initiative, and without resorting to seed saving and measures to restrict save scumming, I'd be absolutely okay with it.

*Yawn* Switching to actual laptop and waking up fully

*stretch* One sec...Rebooting.....3....2.....1..And we're good! \o/

Well, going back to the algorithm that does the rolls: Can't we just increase player weight for better average number vs AI?

Not to tone down enemy initiative but such that we favor the player in initiative rolls. This is because, as you said, we're being biased since the AI normally has a numerical advantage or some such

I only bring up seed saving cause I feel that if people want to keep save 'anytime' feature and also discourage the breaking of RNG...well, I can't think of anything else.

I would've preferred more restrictive saving in order to leave RNG alone and cause it'd make players put more thoughts in each choice as well, but you saw how quick people wanted to kill that with fire.

Since people want to save scum, we don't need a method to prevent breaking RNG....just discourage it from being really simple as quick save and quick load, such that casual players who aren't determined to break RNG wont.

As for seeding breaking randomness, well that just depends on how we do it. I don't think seeding one turn ahead for every possible action would kill randomness; the roll is still taking place per normal....just before every possible action that could involve RNG. Whether the roll takes place directly after the forst grenade is thrown or directly before the player chooses to throw it.....RNG hasn't been killed really in my eyes.

Seeding ahead multiple turns is what I personally have a problem with

There's also the option of seeding in case of save-load, rolling per normal if game not freshly loaded, use seed for one turn ahead if player loads. That just requires a check for when a game is a loaded to lock a seed in place.

Those are actually good points. Perhaps we could use an "Ironman Mode" like in XCOM 1 and 2 for higher difficulties, only 1 save per turn with this turn's seed saved automatically, while lower difficulties still get no seed saved at all. Or maybe make ironman an available option that it is available to all difficulties.

I feel like the save one turn ahead seed would work best on tactician and higher. No need to be restrictive on classic or explorer.

Or as you said, it could just be an option at each new game:
To seed or not to seed, that is the question!

As far as I know, X-COM does it the other way round. The harder the difficulty the more the rolls are rolling in favor of the AI, only on the lowest difficulty I think the rolls are in favor or the player.

As long there is no real artificial intelligence, the AI will always be dumber at the end, so they most needs an advantage to remain a challenge.

I'm not so sure about the enemy ai being dumber, considering that their pathfinding has no troubles at circling over elemental fields while my characters' pathfinding always dives right into them

This is also very true.

It's hard to balance out roll bias, especially how that bias is effected by difficulty.

I'm unsure where I stand on that issue completely, honestly.

EDIT:
In the end, a slight bias in favor of player in classic but changes to slightly in favor of AI in tactician? Mmm, we want to keep AI challenging by giving them as many advantages as possible without destroying the illusion of fairness

Just adding fuel to the fire, ini needs to be fix and saved scumming is a horrible, horrible thing. A lot of save scumming is due to horrible ini and broken CC.

Because you are giving the order to got straight to the destination. If you want to avoid those field, you must tell the correct path yourself.

The main issue with AI dumbness is anyway not really pathfinding, more going for the most fitting target.

An for more examples here:
http://larian.com/forums/ubbthreads.php?ubb=showflat&Number=593058#Post593058

I feel like this has turned into two separate discussions that have almost nothing to do with each other.

1) Initiative is currently massively in favor of the AI to the point of being unmanageable through conventional means. This will almost certainly be handled at some point either by itself or when they (hopefully) get around to reworking the stat system.

2) Whether or not rng should be involved in initiative rolls. This is literally a matter of opinion, the best solution in my eyes is to make it a toggled feature so everyone can decide for themselves. Easy to implement, everyone is happy. Will they do it? Probably not because the entire population of people who care one way or the other is likely currently involved in this forum thread.

/thread

That's a fair summary of the fact that we are discussing two topics.

But the only reason the first was brought up is that it handily solved the first and other things at the same time

I'm doing a run where we all have +5 initiative from Leadership and +2 from 1 piece of gear each and we all act before the AI, with 0 points invested in Wits.
The AI's initiative isn't insurmountable, countering it efficiently just isn't very intuitive.

I mean, that's legitimately just straight up false, it just adds randomness to it and does nothing to level out the imbalance between the player and npc initiative. What fixes the problem would be actually intelligently assigning initiative values to the mob based on how agile they seem like they should be. (ex: bandits are probably medium initiative whereas zombies would be slow) The real way to solve the problem is to put 3 tracks for enemies: fast, normal, and slow which automatically assigns an initiative based on what level the monster is. Ideally something that someone with intentionally high initiative could feasibly overcome. Such a system would be easy to design, solve the initiative entirely, and once created would allow the devs to propagate it simply by marking any new monster they create appropriately in the code. It would take a very small amount of effort now and would make the game much better overall.

An initiative roll would be just as easy. And it solves the issue through the simple fact that bias rolls are a thing. The current issue is that initiative on AI is just so high that putting points into iniative doesn't feel rewarding unless you really put some effort into it.

This effect than extends over time as enemy iniative grows over the course of the game with level and gear.

Whether someone puts 10 points into initiative or 5, there should be a noticeable change in game play. Things don't currently work that way.

Law of averages dictates that initiative roll would present that change presentably. It would change how often one goes last or first or land in-between.

Also, there's no such thing as truly random. It'd be pseudo random at best technically. We were discussing how bias the rolls should be and to whom.

That's one algorithm function call to assign initiative roll value with one call from combat start passing in the initiative stat fir each character on the screen. Can't get much more simple than that

Not that your way couldn't work but i;m not a fan of it.

EDIT:
Hell, adding the algorithm would change game experience without needing to touch intiative values anywhere else

This is in line with my estimates. One must have there level initiative + 5/7 to really be going before the NPCs in the current EA