Larian Studios Forums Divinity: Original Sin 2 Divinity - Original Sin 2 - Suggestions and Feedback VIDEO: COMBAT SYSTEM IS RIGGED WITH SAVED SEEDS

This. is. GARBAGE.

Watch this:
https://www.youtube.com/watch?v=24giMb-n5u4

I am pretty ticked off to discover this. I first noticed this issue when I loaded a quicksave I made in the arena.

I don't remember why I made the original save, but I noticed that Sebille's Ricochet skill ALWAYS bounced to 4 targets but only The Avid One was hit, everyone else was missed. In fact, I couldn't hit anyone except the Avid One with Marksman's Fang (two pairs of the enemies were lined up) or normal attacks either.

At first I thought it was because I was a level below the enemies and all except the Avid One had Encouraged. But on the next turn I was able to hit Rex the dog at a closer range even though his buffs were still in effect.

That prompted me to investigate further. I loaded my first save and moved Sebille close to Rex so distance wasn't a factor and made a new quicksave. Nope, still couldn't hit him. I tried Ricochet and Marksmans's Fang, and I still couldn't hit anyone except The Avid One.


THIS IS GARBAGE.


What is the point of having a stupid ridiculous "73% chance to hit" percentage? Why not just change it to "YES" or "NO", because that's what it really boils down to.

I was okay with seeds to predetermine loot from chests, that makes it easier to balance. But Larian stepping between the player and the combat system? That's BULLSHIT and is unacceptable.

If Larian doesn't want people to save-scum in combat, then DISALLOW SAVING IN COMBAT. Not the stupid half-measure crap that this is.

Personally, I would much, much, MUCH rather allow people who want to save-scum in combat to do so than to interfere and micromanage combat on this level for the player's own protection".

This is NOT a good system, and it is NOT a fun system. CHANGE IT.

Tbh, knowing how this type of game has been working, I had a feeling the "hit/miss" is actually calculated at the start of the turn to make it less heavy for when the attack has to be carried out and also prevent save scumming to begin with. Some games take it further where if they have Fire-emblem like leveling, the random stat is handed as well for that turn so if you want to abuse it, you have to save scum a turn ahead.

It actually has more positive side than negative.

THe only real flaw it has is why sometime, combat doesn't work properly in town where you have like, 20 neutral targets you can hit around and the game has to calculate the chance of hitting everyone in that area ahead of time and save it, which causes some of the skill effect to not work and shoot anything because the game is busy.

Preventing save scumming to guarantee 100% hit rate is actually a good thing and if it is coded the way I used to see, then it should not affect what happens the turn after so if the players still want to save scum the combat, they can, but it's a lot more painful to do so because they have to do it a turn ahead.

WRONG!

This is far worse than save-scumming. This is interference and meddling with combat. It's tinkering with actual chances to hit to "even out" combat so everyone has the same chance to hit or miss. It makes those garbage "73%" chance numbers lies because it's really "YES"/"NO". The numbers are supposed to be there to help you assess your chances, but they don't do that anymore.

Imagine a soccer(football) player doing a kick from the corner near the opposing team's goal. Let's say that the player has a 10% chance to score a goal. Under the rules of DOS 2, if his seed says "you score", then that player will score no matter where he kicks the ball. Kick it at the side, kick it down the field, kick it into the stands, anywhere, it'll bounce into the net.

If Larian REALLY wants to remove combat save-scumming, then they should do it by forbidding saving if a party member is in combat. This chance-to-hit meddiung method is bad.

AND Who gives a fuck if someone else save-scums? Give me save-scumming a million billion times before you give me random seed combat once. THIS IS SHIT. And I am someone who NEVER save-scums in combat, I play all fights start-to-finish unless I'm specifically testing something.

If you want to test out if it's a hard coded to miss or not, try reentering the combat from the start and shoot the same target instead of save to shoot. If you always miss no matter how many time you reinitiate the fight, then the system is flawed because the actual calculation function is not working the way it should be (Should only calculate one turn instead of one fight ahead).

However, if the result is changed, then it is as I said, it only calculate the next hit ahead of time the moment the combat start and will not really affect people who just play normally because if they go to the same fight at different time, they will experience different result. Only when they save scum will they experience the same hit/miss.

This method of coding has been in gaming since Fire Emblem ages. This game is just being nice and allow you to quick save so people try to abuse and get to see its anti-cheat in effect.

I've loaded my pre-fight save a couple more times and it definitely seems to be rigged with the same miss chance. Round 2 does not seem to always be the same.

But I disagree: the system is not flawed because of a calculation problem.

The system is flawed because it exists in the first place!!!

I don't care what other, completely unrelated games in different genres created by different developers do. This is not Fire Emblem. I am talking about D:OS 2. This is VERY BAD DESIGN.

Allowing free use of save scumming to get different results in combat is far, far better an idea than this terrible garbage deciding in advance, before a fight even BEGINS whether or not your attack on an enemy will hit.

It is INDEFENSIBLE. PERIOD.

Hm, seems to be the same rage-inducing thing that was in XCOM 2 as well. That being, no matter how you move during a turn, you'll ALWAYS miss the shot if the seed says so. Was pretty infuriating to get squadwiped because I couldn't shoot that one guy who'd then blow up everyone.

I hope that system isn't in DOS2.

It's not in defensible. It's one of the fairest system that has been used for ages in games like this.

If anything, it's more like the system is bugged and calculate everything way too ahead of time because if it is truly rigged with the same miss chance even if you reload the file from pre-fight, then it needs to be fixed to start calculating at the start of the fight. That is the only flaw it has at the moment because you did state that Round 2 is not always the same, which means that the thing is working the way it should be. Only Round 1 is bugged and if they fix that, then no one should experience the forced miss on first hit anymore and only when they start save scumming will they notice that they game isn't allowing them to cheat the fight and crit on every hit.

This bug isn't hard to fix either. Simply add the "Calculate Hit" function to the "Trigger fight" function.

And if free use of save scumming is allowed, you don't ignore it, you design game with that choice in mind. This is what a good game design is like, fully support legit options while also preventing cheating at the same time.

Modern strat game that are similiar to this game with same RNG element (like Blackguard) allow saving because sometime, combat can be longer than planned and you need to go somewhere and certain fight in this game do last ages for players that don't know how to min/max. It also exists as undo option for some which follows the golden rule of design where the user should feel like they're in charge on normal ground. Real designers don't reduce the game quality of life simply because the cheating option exist. They come up with way to make it not rewarding to cheat, and this is one of them and why it has been used for ages.

Do you really want to see people come ask for help on fights and the only advice they get is "Reload until it crit on every hit"? That is no different from making a mockery of a game that is supposed to be about strategy.

I just thought of yet another flaw:

If this move is supposed to REDUCE save-scumming, well, I can tell you at least for me, it's going to have the OPPOSITE effect. I basically NEVER save-scum before fights, I always play them straight through from start-to-finish. The only exception is if I am testing something.

But now, with the chance of success pre-rigged in advance, It's actually making me more inclined to save at the start of every character's turn so I don't do things like waste a Ricochet which was preordained in advance to fail, putting it onto cooldown for 3 turns.

Why shouldn't I save-scum now, to let me go back and try something else just in case my tactic was destined to fail in advance?

CONGRATULATIONS, LARIAN!

There's a huge difference between save scumming to guarantee hit + crit and saving to undo mistakes a few turn back.

The first one is the only one that is truly frowned upon because it directly affects how much damage you can dish out in a turn and make fights way easier.

On the other hand, repositioning yourself or trying another method is actually a legit strat because the AI will also change the way it reacts.

Wow, this is horrible!
I'm definitely not pleased to discover this and I really hope this will be changed to a functioning system. There's no point in pretending to have a dice roll or percentage based system if everything already is pre-calculated. The original game had a functioning system and I don't see why it should be necessary to change alter that. I certainly hope this system only is present during the early access version of the game.

Seeding the numbers generator is a completely normal practice.

Just because it prevents save-scumming, doesn't mean it's somehow flawed or that it should go away.

A 'completely normal practice' that wasn't present in the original game if I'm not mistaken.

Again, other people save-scumming is hardly my concern but if the measures that are put in to combat them suddenly affects me it becomes my problem as well. Once again, there is no point in including the percentage for a skill to hit if it doesn't actually work.

It does work though.

It's still the same X% chance to be set to yes or no, just that this chance isn't re-rolled every time you load the game.

You want a different result, you take a different action, not just repeat the same one until it works, otherwise all actions might as well have a 100% chance of success.

It sure seems to be! I tested changing position before attacking and it didn't affect anything. Distance from target is not a factor for the chance to hit, it's just YES/NO your shot/special attack will hit.




Yes, you DO ignore it becuase it is NONE OF YOUR BUSINESS. I didn't mind it so much when loot was fixed in order to make a better loot table for distribution, but I didn't particularly mind save-scumming for loot to begin with. I don't mind that loot save-scumming is gone, but the only times I did that in D:OS 1 was for those certain chests with random skillbooks when I already had a party member with that skill.

The game should absolutely, under NO CIRCUMSTANCES be deciding ahead of time to enforce a certain number of hits and misses in order to conform to some ridiculous made up equilibrium. If the chance-to-hit meter is wrong, it shouldn't be in the game.




That is a bullshit strawman which has nothing to do with save-scumming at all.




That's wrong too. All targeted attacks (I am excluding surface creation) either hit or miss. If a normal attack will miss, so will Ricochet and Marksman's Fang.

And I have a high level of skepticism that the percentage chance is accurate when I have a 73% chance (or more) to hit when aiming Ricochet at any of 4 targets and yet only one of them gets hit. What are the chances that three 73%+ chances to hit all miss? About 2 percent.

I agree with the OP. I had tons of fights I retried for various reasons, and no matter the attack I opened with I would always get a miss. For hard fights, this often required me to save before actions so I wouldn't waste an attack pre-rolled as a miss and instead use healing or buffs.
Something is wrong with the hit chance system, and it's not just the seeded rolls, the % listed isn't accurate.

It might very well be that you have rolled those 2 percent of a chance to fail, and since the result is seeded that's what will always happen.

That's not a problem with the system, it's a problem with your expectations.

Welcome to XCOM.
Errr... wrong game.

Possibly. Or possibly it is a bug with the miss chance calculations. I don't know, I don't have access to the code calculations. But there's no harm in telling Larian about that so they can decide if this is working as intended.

It's like you and me are discussing two different things.

I say the hit rate is supposed to be calculated individually at the start of turn for each character when a fight start, not a forced hit or miss to begin with. If the system already calculates the whole thing before the fight start and forces you to stick with it to the point the Arena fight has a predetermined miss no matter what you do, then the calculation portion is likely bugged. Saving accuracy for that one single turn is fine but saving accuracy before the first turn isn't.

But one thing to mention, if the hit/miss is truly forced, why can I start off nearly every fight with a Stealth snipe? If it says I should miss, then I should miss. But the thing is, I don't and start off the fight with hundreds of damage.

Another possibility is that the calculation is broken like how Blind breaks the game until reload because the player is stuck at 0% hit rate. The first hit actually starts off at 0% accuracy because they really forgot to make it calculate the hit rate the moment the fight begins until an action is taken, which is why Stealth-snipe works and has varying crit result. Stealth ticks the accuracy check to actually work before Snipe is used.

That's NOT how pre-generated seeds work (usually. Can't really say for this game).

It's not a list of "YES/NO" values. It's typically a feed of pre-generated numbers (i.e. In a system taking variables from 1 to 100 it will be a list like "23, 71, 56, 12, 88, 1, 44, 99, 17" etc... And then the game just uses these values as "dice rolls").

At that point, success/failure is not predetermined yet.
For the sake of an example, let's take the third value, 56.
It doesn't act as a YES or NO, it act like a dice roll of 56 on a 1-to-100 scale. That means that the success of this action is entirely dependent on WHAT DO YOU DO:
- Do you attempt a very easy action where a low/average roll is enough to win? Then it's a success.
- Do you attempt a complex action, when you need to roll a 75-or-above to land a hit? Then it's a failure.

Nothing really fishy or dishonest about it, really.

P.S. People get outraged over the most trivial bullshit when they realize the trick, but "pre-generated seeds" are not a XCOM 2 invention either. A classic like Jagged Alliance 2 used them years ago.
It should also be noted that while they are introduced for several reasons (and prevent save-scumming could be one of them) there are often ways to "cheat" a pre-generated seed.
For instance you have just to change the order of your actions to get different outcomes.

I do remember problems with trying to seed pseudo-random number generators which is that for the same seed they would always come out with the same sequence of numbers: I imagine not much has changed except that the scope of seeding them has improved somewhat over time. But if seeding is pre-defined then, yeah, the outcomes will always be the same.