Shadowrun Shadowrun: Dragonfall - Director's Cut

[Ninjerk]

I thought the same thing when I played it.

 

[cvv]

RNG's definitely dodgy.

Every game in the world uses the same goddamn rng. Actually write down every miss and make and what your percentage chance to hit is.

Bullshit. I missed twice with 99 percent within some 30 shots. I missed countless times with 80% chance, sometimes twice in a row. Anything under 60 percent is basically "don't even try". Sure you can always say "random chance is random" and we can argue about it till we're blue in the face. But I've played some games in my lifetime and I can smell sloppy broken RNG like a ratshit behind my sofa.

As for your suggestion that RNG is a standard tool everybody's using - another bullshit. Take nuXCOM. At launch it was an absolute fucking joke. You couldn't hit your own leg with a shotgun while aliens were hitting you regularly from across the map. But then they released Enemy Within and lo and behold, suddenly everything was fine. Everyone stopped complaining. Sure you still miss with 90% chance - a few times in a 100 hour playthrough.

So evidently you can program a fucked up RNG and you can unfuck it.


*****


Btw just finished the game. Absolutely fantastic. Instantly added to my top 10 RPGs of all da timez. Best sci-fi RPG ever? Maybe. I enjoyed it more than Deux Ex, that's for sure.

Spoiler

At the end I butchered my team and let Vauclair kill all the dragons. Aaaand the world ended. So I reloaded and freed Firewing instead. And then they killed the POTUS. Meh, fuck the POTUS.

 

[tuluse]

Lol NuXcom lies about it's percentages, every time you miss it increases chance to hit secretly.

 

[Zetor]

Yeah, nuXCOM uses a streakbreaker (like City of Heroes did, if you ever played that). If you're getting less hits than you 'should' in a certain time window, it nudges the percentages upward until you do.

If you make a 'fair' rng, it will just get a lot of crap from players because of confirmation bias / gambler's fallacy / etc.. and also because (plausible) anomalies like rolling several snake eyes in a row are going to be much more grating when you're only rolling ~1000 dice during your entire playthrough. Consider that if you have a 90% chance to hit, you'll miss on every 10th shot... it's certainly plausible to make that many attacks in a single turn. Missing two 90% shots in a row is completely reasonable from a probability standpoint, but it will cause a lot of rage and smashed keyboards. Also consider that the enemies miss '90%' shots just as often, only they lack the ability to make angry forum posts about it.

 

[Blackguard]

tuluse is probably right because if you only miss so few times with 90% chance then the XCOM RNG is actually fucked. You should be on average missing a 90% chance shot every 1.7 turns considering you got 6 soldiers, and assuming each of them takes one shot per turn. Even 95% chance would amount to a miss on average every 3.33 turns. Point is things that many morons immediately see as proof that the RNG is broken are actually really common.

 

[cvv]

Lol NuXcom lies about it's percentages, every time you miss it increases chance to hit secretly.

Yes, on Easy and Normal, i.e. SuperEasy and Easy. If you play on a Normal or Hard difficulty (Classic and "Impossible" in the modern lingo) the percentages are what they should be. Unlike Shadowrun.

 

[Infinitron]

I don't think Shadowrun lies about its percentages, but what's true is that on Very Hard difficulty, hit chances in general tend to be very low. However, this is critical to giving the game some measure of difficulty.

 

[ArchAngel]

I also noticed RNG in SRR being strange. I missed way too much high % shots.
And I played all SRR games for a total of about 200+ hours.

 

[cvv]

Missing two 90% shots in a row is completely reasonable from a probability standpoint, but it will cause a lot of rage and smashed keyboards.

I missed many, many times with 80% chance. You know how many times I hit with 20 % chance in Dragonfall? I'll tell you exactly - once, that's how many. Despite trying it quite a few times just for lulz, just to prove me right. So don't give me this patronizing bullshit bro. This RNG is broken and that's a fact.

 

[tuluse]

Learn about bias confirmation and I'll stop patronizing you.

Unless you list here every shot you took, chance to hit, and result, I don't believe you.

I'm sure SRR is just using built in C# random functions and they work perfectly fine.

 

[DeepOcean]

If enemies could hit you 100% of the time and you were dependent on RNG I would understand the reason for raging but you miss alot and the enemies too. On XCOM, you can kill an enemy or be killed with one or two shots (this can make RNG infuriating) what isn't the case on shadowrun unless you have 20 of life or something. Besides, accuracy is a key component for the character progression on the game, only strengh improves the damage of melee weapons, guns and spell skills mostly improve accuracy and critical chance, if you started with way too high accuracy you would have 99% accuracy all the time.

Anyway, I remember stunning whole rooms with critical attacks from that special melee weapons attack that consume enemies 4 AP per hit, I depended on hitting the target or I was fucked because my character was exposed. Remember that at 90% accuracy you could hit with some reliability. This discussion of RNG always end on something subjective, after they fixed the bug where the game applied huge debuffs on your accuracy, there is nothing concrete to say the system "cheats".

 

[Infinitron]

Note that old versions had an issue with the difficulty level changing around on save/load. Have you ever noticed a situation where you had a high chance to hit an enemy, then loaded a save, and then the percentage was lower/higher?

 

[cvv]

So combing through a couple of forums, it seems to be a frequent complaint.

I'll just quote two posts that caught my eye and then I'll be forever silent since this game is awesome and I don't wanna bash on it too much.

https://steamcommunity.com/app/300550/discussions/0/610573751146375631/?insideModal=1#p3

Yeah there is no way the displayed hit % is the actual hit percentage. To call "gambler's fallacy" on questioning missing five 99% chances in a row shows a gross misundestanding of the term and the subject. It's mistakingly thinking that past events will influence the odds of future independent events. It has nothing to do with looking back at the results of a series as a whole.

Anyone who's taken a statistics and probability class knows that there's essentially a rule that you get to call B.S. when results start falling past a certain unlikely threshold. The odds of getting a royal flush in poker is less than 0.001%. It's possible to get one, but when someone gets five of them in a row, you get to call B.S. because they're definitely cheating. You don't defend them by shouting "gambler's fallacy!". So Tuluse is full of shit.

It seems like the actual chance to hit is 25-30% (or more!) less than what the game tells you. Taking a 89% shot is practically a coin flip, and anything less then 50% is almost a guaranteed miss. The theroy that selective memory means you only remember the misses at high percentages goes out the window when you realize can't remember any moments where you hit with a low percentage. I never made a single hit all game when the odds were under 40%.

The same thing happened in Dead Man's Switch. There's got to be some kind of negative modifier the display doesn't show you.

It's just bellyfeels, sure, but it's exactly how I felt too.

The game's code for chance-to-hit is a mess. It calculates a to-hit estimate for the UI using one set of functions, and then uses a completely different set of functions when you actually perform the action.

What they SHOULD have done is precompute the chance based on all factors, show that chance in the UI *and* keep the value for use later during the final random roll. The way they handle it at the moment is wasteful, and much easier to introduce bugs.

TL;DR: Ignore the number shown on the UI - it has little to do with the actual chance calculated when you perform an action, and will just make you rage.

Also, they DO have code which tracks the number of lucky/unlucky shots you've made/missed, but I didn't bother investigating how it works in any detail. The % ranges were 30% and 90%. Perhaps their modifiers are backwards and they're penalizing you for missing 90% shots instead of "giving it to you" if you miss too much. That would explain what people are saying.

Yeah, most likely explanation - there are some calculations going on behind the scenes and not reflected in the HUD percentage numbers. Fuck. Whatever.

 

[tuluse]

I have yet to see any data.

People complain about the RNG in literally every game ever, so no surprise there.

 

[Jaesun]

If you REALLY want to complain about RNG, just go play any mod in FRUA. Seriously, what in the fuck.

 

[Ninjerk]

You guys talk about this shit like gambler's fallacy, etc. are some kind of new phenomenon in video games.

 

[almondblight]

Bullshit. I missed twice with 99 percent within some 30 shots.

There's around a 1 in 23 percent chance of missing twice with 99% accuracy in 30 shots. If it was more like 35 shots, it would be around 1 in 17. Running into that once during the entire game doesn't exactly scream broken RNG.

Also not sure of the accurate the numbers are. I can't think of any situation in the game where I'd be firing off 30 shots in a row with 99% accuracy. I only played on very hard - do the easier difficulties give you a 99% chance to hit most enemies or something?

 

[Ninjerk]

Bullshit. I missed twice with 99 percent within some 30 shots.

There's around a 1 in 23 percent chance of missing twice with 99% accuracy in 30 shots. If it was more like 35 shots, it would be around 1 in 17. Running into that once during the entire game doesn't exactly scream broken RNG.

Also not sure of the accurate the numbers are. I can't think of any situation in the game where I'd be firing off 30 shots in a row with 99% accuracy. I only played on very hard - do the easier difficulties give you a 99% chance to hit most enemies or something?

I'm playing again and not seeing the issue this time. Did someone say something about a patch earlier?

 

[agris]

These complaints are as old as prostitution. The whiners never stop, and the mathematically inclined can explain until they're blue in the face, but there's a reason its called perception bias.

 

[razvedchiki]

ust do not build a close combat character in SRR OC

i have seen a lot people saying the same,and i must say its bullshit.

i made a human melee street samurai both on OC and on dragonfall ON VERY HARD and i can say that the build is definitely playble,more so on OC cause for some reason its easier than dragonfall.

troll is the best race for this build not only for the +1 on strenght and body but because also it allows you to hit the cap on body close combat and strenght.

statwise you need to focus in body/strenght/close combat and whatever leftovers should be spent on charisma if you want to talk your way through instead of butchering everything.
also its higly advisable to invest on cyberware that boost hp and strenght.

on my dragonfall playthrough im now 1 mission away from paying alice and i have 11 strenght 10 body 6 charisma 3 quickness 3 dodge 7 close combat 120 hp and 6 armor.
still i dont rush on the enemy,but instead i run from cover to cover and usually throw a flashbag or 2.


edit
just to clarify,due to the crazy hit chances on very hard i still died a lot at the begining but later i stocked on medikits and flashbangs and things went smoother.
most difficult mission for me was paradoxicaly loose ends,i assumed that because of the corridors battlefield i had the upper hand but those knight errand fucks just didnt fucking DIE ALREADY.
i had to concetrate fire from my whole squad to just one pesky knight to have any chance to kill him in 1 round.
and that final battle with the basilisks and the reinforcments.....

also
https://www.youtube.com/watch?v=nt44yteqhhk

 

[Ninjerk]

These complaints are as old as prostitution. The whiners never stop, and the mathematically inclined can explain until they're blue in the face, but there's a reason its called perception bias.

You can't see it, but my eyes are in the back of my head right now.

 

[eXalted]

I'm sure SRR is just using built in C# random functions and they work perfectly fine.

The C# RNG is kind of fucked up. Unity has its own RNG function which is better than the vanilla one. But which one they use... hmmm

 

[thesheeep]

Default RNGs in any language are usually not suited for games.
Their distribution is simply not very even.

 

[Doktor Best]

Default RNGs in any language are usually not suited for games.
Their distribution is simply not very even.

Well aint that kinda the point of it?

 

[thesheeep]

Doktor Best

No, it isn't.
In fact, not at all.

In most cases, you want the numbers to be random but evenly distributed. For example, you want numbers between 0 and 100, but each number should be equally likely to come up. Like a dice roll.
Most standard RNGs in languages have a RNG that is not that evenly distributed. It varies from implementation to implementation, but it is not unheard of that the standard RNGs have some kind of focus in the middle, beginning or end of their spectrum. Like rigged dice.

If you rely heavily on RNG, you should make sure you use an implementation that suits your needs (which usually means not using the default one).