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

Not here. That doesn't work. Hit or miss chance doesn't care about your actions. That hit or miss is pre-determined, it doesn't matter if you use a normal attack, Ricochet, Marksman's Fang, or first try to move and then use one of them. I tried.

I got a suggestion on the Steam forums about whether accuracy-increasing skills would help.

In short, using Mark and increasing my chance to hit from 73/74 to 88/89 changed one of the 74->89% misses into a hit, but the other 74->89% miss was still a miss (as was the 73->88).




Excellent question! I did a quick test, I loaded my pre-arena fight save, took my guy to buy Mark and walked back and saved again. This did not seem to affect the initial pre-fight seed.

It'll take me some time to make a video, as I have something like 13 scenarios to demonstrate, and I should do at least 5 attempts rounds with each scenario, multiply that by 4 targets... it'll take a while.

I tested Mark. There is one enemy with a 73% chance to hit (Rex) and three with a 74% chance to hit. Mark boosts the chance to hit to 88/89%. Rex with the 88% chance to hit is still a miss. The Avid one is still a hit. The Cultured One at 89% is still a miss, but The Tenacious One at 89% changes to a hit. So I still don't know what's going on.

I reaffirmed that moving close to the targets does not change who I actually hit, and I also tried delaying my turn and having my mage teleport me to high ground. That also did not affect who I hit at all. That's right, even with two people moving before Sebille and moving her to a different location, it didn't affect my chance to hit at all.

I don't care what you have to say about XCOM - This is OBVIOUSLY not working properly.

The thing is, you don't fully understand how it works. The seed is generated EVERY TURN. This is why Fort Joy Ghetto is such a mess to be in until you kill half the town with all the random dialogue/item freezing or massive slow response. If your game suddenly freezes, it's because the game keep recalculating your "On next hit" odd and saving it for everything, including the neutral, on the screen because they can participate in the combat at any moment.

Remember that in this game, your accuracy is pretty much constant across all target because it only looks at your level and your weapon level. Dodging part is not shown yet but from the look of it, it's a "After hit chance" calculation type of thing. They don't need to look at different target. They just save your roll for the next hit based on your current chance and saving it.

So what you did, was you went into a fight, where you already rolled a "Fail". If you want to show that the system is flawed and is cheating you, you have to make 100 different characters, go to that place and try to hit and make sure it misses on all 100 hits.

Edit: Based on your new found information with mark, this means the whole thing is calculated per target individually. They should change it to just work with certain % and save to reduce the amount of strain it put on the game. Getting real tired of having to wait for ages in Ghetto only to see it freezes completely. Either that or Mark is one of the few skill that can be used to recalculate the hit chance, or the game calculate both normal and boosted hit chance, doubling the work.

But that's not the case. I'm on my 5th run where I abuse everything and on this run, I had 0 misses on my crossbow in the Arena and just 2 shot stuffs with my Houndmaster crossbow + haste on 2nd turn with high ground power. This is why "Stealth Snipe" produces different result if you wait a bit before you shoot before you go into combat. Sometime you crit, sometime you don't. The game will reroll if there's no flag to trigger combat alert. Only people who try to save scum will feel punish by this. If you play normally you won't be affected by this whole "My first hit always miss" because the game will still calculate hit/miss properly until you get into combat. This only affect people who purposefully save before they attack.

I doubt it. Nothing in your video suggests that.

First, it's worth stressing that changing your action probably doesn't change your seed, and if that's, say, a "1 out 100" roll, there are good chances it will always result in failure of that next action. Otherwise it should be influenced by your chances. But if it's a "73 out of 100", then doing a higher-chance-action should remedy the miss/failure.

Second, there IS a sure way to alter what seems like a pre-determined failure: using a "sure hit skill" that doesn't rely on chance, for example.

Now, what should be tested at this point is if that "failure roll" will just be skipped using the "sure hit skill" or just delayed for the NEXT action, since the previous one didn't "roll a number".

You could also, you know, just deal with the occasional failure, which is bound to happen at some point, instead of presenting this as some sort of ethical bankrupt, frankly.

It's too late to reply to everything right now, but I've done some more testing. I believe I've determined in part how the hit calculation works.

Distance to the target is irrelevant. Height difference is irrelevant. Amount of time passed is irrelevant - the game is not generating new seeds every turn (every 6 seconds). Amount of combats engaged in is irrelevant (as long as you have not leveled up).

I loaded several different saves, from just after killing Silence (but before fighting the crocodiles), from just reaching level 2, to before reaching level 2, and after reaching level 3. The only thing which seems to affect the seed is leveling up. All the fights I engaged in at level 2 had the same hit pattern. I'll do another test seeing if leveling from 1 to 2 again produces a different seed or the same one.

The hit calculation is tied to each specific enemy, rather than to the player. So it's not a case of "first shot to any target misses, second shot to any target hits". It's "first shot to enemy X hits, second shot to enemy X crits".

***

So if you think that distance to the target should not be a factor in ranged weapon hit chance (because getting closer does not improve your chances), this is fine.

If you think that leveling up should be the sole determining factor in your chance to hit an enemy in possibly several hours from now, this is fine.


EDIT: I'm pretty sure that I'm wrong again! I loaded a game from when I was level 1, I reached level 2 and it still didn't change the pattern!

That means your chances of hitting are determined when you start the game. TOTAL GARBAGE.

Distance from the target has NEVER been a factor in this game or the previous one. You always only had to care about being in range of the skill/weapon or not.
If anything, now they even introduce high grounds etc.

I'm not sure why that's suddenly so surprising to you.

Also, you keep talking about all the extensive tests you made (which you are not detailing too much here, so it's hard to judge) but I'm still not sure how any of what you are pointing should suggest unfairness.

I am a simple man, the game says i have 90% chance to hit, then i expect to hit about 9/10 attacks, if i do that i am satisfied, if it turns out it was destiny for me to always miss that one shot that one turn no matter how much i try reload it then so be it, i dont mind.

Yeah, sorry to say but that video proves nothing.
The seed is indeed fixed, that is proved, but what would be the issue about that?? Yeah you can't savescum, which ... is annoying??
The way a RNG works is taking a value from an equation, which will, given the same parameters give the same result.
The fact of the matter is : there's no need for new seeds during a game, like EVER.
The only need for a new seed is for a new game, from there, the whole game will be different seeing as values will popped here and there and the algorithm will just be popping random numbers here and there. The only way to prove your theory would be to show on two different games the same patterns going on and on.

The funny part about that is : What do you think Larian did? Create a rng? No. They just took one already existing, probably Mersenne-Twister and got done with it. It's not like they intentionnaly created pattern, and if they fixed the seed, again, how does that constitute a problem? Cause I'm not seeing it. A fixed seed does not mean that a 73% chance of hitting is a yes/no situation, it means that the algorithm has 73% of giving a yes, and given a specific seed does not or does not do so. But whenever a second number is asked you rng will pop another number.

It goes something like this :
Imagine for the beginning of your game 4 numbers will be the first to pop on a scale of 1 to 100 :
25
87
92
12

Those number may be fixed but not the situation where you will use them. So it could go :
25 => Missed a shot
87 => Failed to persuade someone to comply
92 => Shot success
12 => Failed to do something
Or:
25 => Managed to persuade someone
87 => Shot success
92 => Shot success
12 => Shot missed

There are a lot of numbers generated in a game, so it's pretty much invisible to a player what goes on behind the curtain, it's not like you could really predict the RNG, the only way to do so would be to reload a game, manage to not trigger the value you want to apply it to something else, which is :
1) Pretty hard to do
2) Something a dev should not worry about.

A RNG is only good as long as it generated a series of number. If you recreate a seed everytime, your RNG is not as good...

You're making way to big a deal of this.

Watching the video and reading what has been said here I'm more like... I like how it is. I'm the type of player that play Hardcore mode, I'm not save scamming and I like the fact that that what we saw happening in the video, is how it is because I'm very much against save scamming. I'm of the opinion that player that save scam should either play Hardcore just to learn why we play games, or just don't play at all.

Isn't that a bit harsh? I don't do so either, but I don't really care how other players play: it's their game and they can play it how they like. Maybe it's a different story in MP, but in single-player games, I don't see why it matters.

It is not only that. There is something broken in the chance to hit.

For example (and please don't anyone say there is a 0.001µ chance) I made a 2 handed axe for the red prince. For 2 levels it missed every single time. For 2 levels. Not mostly but EVERY TIME. It had a 95% hit rate but missed every single time for 2 levels.

I switch back to the standard 1 handed Braccus Necro-Fire thing and it hits 95% (or most) of the time as it should.

Something is not right there. Could be crafted 2 handers don't work. Could be 2 handed generally - I'm not sure

Am I the only one always missing my first attack as Sebille against Magister Dayva (Mage on the docks, with silent monks) ?

Every time I tried to engage the fight from afar, wether stealthed or not, sniping or not, I always miss. In every playthrough since the new patch.

I don't have a too much of a problem with the system as it is, but this one particular case is worrying me.

So, if I understand it correctly, each character in the game is assigned a random number queue which is set-up at some very early stage. If the first number in the queue is, say, 70, then the first attack against this character will fail if it has a hit chance 60% but it will succeed if it has a hit chance 80%. Just as Linio and Tuco described.
If this is the case, doesn't feel unfair.