• [2018/06/22]
    By using our forums, and our in-game services, you agree to be bound by our Privacy Policy found here:
    skullgirlsmobile.com/privacy

This game just flat out lies about probability

spades111

New Member
Joined
Nov 22, 2020
Messages
12
Reaction score
5
Points
3
Age
55
What we see in the frontend and what the developers have coded in the backend can't be the same.
I wish they would just advertise the actual challenge rather than make people think they're extremely unlucky.


tl;dr "justification"

Take today's dailies and prize fights into consideration.

In the PF both you and the enemy build 5 miasma stacks while on the opponent's side that resets into a fatigue which has 50% chance to stun.
I've gotten stunned 100% of the time in a 15 game streak.
Not sure how many times I got stunned but I decided to counterplay it by avoiding the stun with backdashing and turtling.
Then I noticed the a.i was never getting stunned after their fatigue ran out.
Using lower FS character to make matches last longer I was getting 5-10 per match ad gave up after a 7 match streak.
Not a single enemy stun.
Assume 5 fatigues per match, that's 35 fatigues total (5 x 7). The probability of not getting a stun is...
0.5^35*100=0.000000003% chance of that happening.
Yet here I am

In the painwheel daily...
+30% chance to crit, crits reflect damage. me using a painwheel that has 5% crit chance stat. so 35% chance total.
Ignoring some hits in multi-hit attacks, I was consistently having 7-10 hits of crit happen in a row.
The odds of that, let's go with 10 hits.
0.35^10 *100=0.003% chance.
That's of it happening once.

It's the same nonsense with any chance based stat. They proc way more often for enemies than they do for you

Final thoughts

I don't care that the game is the way it is as playing around these things isn't impossible. But stop lying about it. Let me know that I'm in a situation where I have maybe a 80% chance of getting debuffed or proccing an enemy buff. At least then I'll feel like I overcame a challenge or something
 
Last edited:
  • Like
Reactions: AnonJustAnon
I personally haven't experienced anything like this and I think it might be confirmation bias.
If you really do think there's an issue, your best bet would be to record some videos of this happening, or to gather some community members and do testing together.
 
Sounds like bad luck. The rare times the enemy is winning against me and gaining fatigue, they get stunned lots of time, saving my bacon.

One must remember a 50% chance will always be a coin toss regardless of previous outcomes. You could get a thousand-streak of stuns and that would still be fair
 
Yeah, it's all experimental vs. theoretical probability at work here. Devs can only present the numbers. How it works in any given situation is up to the RNG.
I've had battles where I'm not stunned once and man, it is great when that happens.
It's frustrating when random forces undermine skill. At least you know the odds going in.

Better luck next time and stay frosty.
 
I say this as someone who works as a developper, not in video games but still, and as someone who has worked in videogame QA for several years while completing my education.
The probability of being so unlucky so consistently isn't most likely not bad luck, it's most likely by design, or unintended design, or a bug that breaks your random function.
It's not like we can see the source code for the RNG or even if someone posted a snippet, we can't be sure that's what exists in out builds.

However I do assume there are other factors such as resistance stats. Who knows maybe I played against 7 teams that invested heavily into resistance which affects how often I'll see the enemy get stunned

Like I said I have no issue with the game being the way it is. I'm going to believe what I believe. The only thing that could convince me otherwise is if there was legal ramifications in having misleading stats and if it is simple to get discovered. In that case, yeah I wouldn't think any company would take such a pointless risk

In any case I'm just venting for the most part. I have zero interest in setting up my device for recordings. (I don't have an android version with easy access to screen recordings)
If the devs studio wants to test it themselves they can. I played like a QA tester would. Just backdashed to my corner and let the enemy do it's thing while never getting stunned

I had a similar issue with a Big Band daily. Where ever hit would trigger the armor, unflinching and then invicile. And waiting for the buffs to expire just didn't leave me with enough time to win matches. I think the odds of it happening were supposed to be 10 and 15%

The solution is simply to keep trying until the stats behave "normally."
It is what it is.
 
As a person who does biological and medical statistics on a routine basis, I can assure you that with enough time and huge sample sizes you are going to see all kinds of weird things. I have seen a set of 200 patients, 18 of them were diagnosed with cancer, and 16 of them were right next to each other in the alphabetical list.

As a person who observed any strange probability deviations through my whole life out of scientific curiosity, I can assure you that crappy things are tend to happen altogether. It never rains, but it pours. I've tossed 12 tails in a row. I've had my important blood sample given wrong reference number four times in a row by four different people. In last 3 weeks I have seen my bus leaving the bus stop earlier than the scheduled time, when I was just across the street, 15 times out of 15 (and I always leave my workplace at different times). I have seen Untouchable ignoring every hit of Val's ECG Flatliner.

I guess, I can assume that the whole world is rigged to make me suffer. But then I find $100 bill on the sidewalk. I suddenly win a free copy of a very expensive software product. I pull 2 Shiny Diamonds in less than 3 months. That's just statistics. Any outcome is possible, just the probablilities are different. But remember that the brain tends to only notice bad coincidences, while good ones are taken as a normal state of things.

I don't believe that the game is rigged. I have never seen anything that would be too unusual for me, and I pay extreme attention to such stuff. But even if the RNG is malfunctioning somehow, I don't think that it is in any way intentional. Like @Fel said, the best way would be to record a video or try to reproduce all conditions associated with the thing that bothers you and see if it reappears.
 
meh if its bad luck then its bad luck. i do acknowledge the possibility.

if no one else is observing exponentially improbable things happening consistently then it must be me.
but if that's the case, then boy am I consistently unlucky.

bad luck can be hype. in fact if the 0.000000003% situations i described happened once in awhile I would have been making a topic being excited about it.
Lol the way I proc unlikely stats you, you would think I could pull 11 diamonds in a row or something. But nope, just bad luck XD

I say ^ that with regard to post #6. Yes given enough time you will see plenty statistical anomalies. But if they keep consistently happening then the issue no longer has anything to do with "yeah you'll eventually see it"
But obviously in the case of post #6 they aren't experiencing what I have been.
 
  • Like
Reactions: AnonJustAnon
I knew i was a little frustrated with this PF modifier too, but i've gotten no less lucky than the AI personally.

Some tips on how to deal with this the next time it comes up

1. Equip and use moves that give you immunity or use supports like Surgeon General that give it to your teammates

2. If you're too lazy to equip such moves, like myself, another good way to avoid the stun is to just tag out shortly before the fatigue runs out or queueing up a BB3, becuase those are not interruptable by stun
 
but if that's the case, then boy am I consistently unlucky.
I can feel your pain, believe me.

Lol the way I proc unlikely stats you, you would think I could pull 11 diamonds in a row or something. But nope, just bad luck XD
That's just the way it happens. I have accepted that several years ago and stopped worrying too much, It's now a default mode for me. People call me "pessimistic", I call myself "prepared".
 
  • Like
Reactions: AnonJustAnon
As far as I know we don't intentionally misrepresent stats in any situation. For example, "let's just make the AI proc this ability more often so players lose more" is not something that has ever been considered. The stats you see in game on descriptions are populated dynamically from the same place where the activation chance is set so that we're not setting it in two spots.

Since Skullgirls is PvP and the AI works by controlling characters the same way you do (via a virtual controller) it's really important that we have no differences between P1 and P2, since that's more headaches to deal with when dealing with determinism in our netcode which uses save load states.

This weekend I wrote a script to run the Fatigue stun activation section one million times and recorded the results, and then ran that 9 times, and the result was 50% across the board, give or take an extremely small deviation as expected.

Note that if you are not a player who invests in resistance or accuracy, and you're playing against opponents who do, you'll see your own Debuffs fail to apply more often - although, it should print "RESIST" above their head in grey text, but it can be easy to miss.
 
For counter playing I think I'm fine for the most part, I typically complete whatever events without Bruteforce because... I can't yet for the most part, at least not as my go to strategy.

It's just that mathematically the probabilities are messed up.

This weeks accursed experimentation... the boss node. 5% chance to lose 25% hp when hitting the opponent.
I decided to test it out.

About 5 rounds without ever getting hit by the opponent, I've KO'd myself consistently. Often with 75% damage done within a single ~10 hit attack string.
Even if accuracy and resistance applies to event buffs/debuffs/modifiers (which imo is dirty) how much accuracy does the a.i have? You need 900% accuracy to get that proc up to even 50% chance

Again it's unfortunate but I don't have a device that supports newer android with the easy screen record and I'm not going out of my way to provide footage for something we can all easily check if our characters are strong enough to get there
 
Even if accuracy and resistance applies to event buffs/debuffs/modifiers (which imo is dirty)
They are not supposed to.

This weeks accursed experimentation... the boss node. 5% chance to lose 25% hp when hitting the opponent.
something we can all easily check if our characters are strong enough to get there

As requested: Last Hope Valentine, AE Boss node, no Res/Acc bonuses, avreage combo length is 30 hits, hitcounts are based on combo hit numbers dislplayed in-game.

Round 1: 93 hits total, 4 times HP lost;
Round 2: 127 hits total, 4 times HP lost;
Round 3: 96 hits total, 5 times HP lost;
Round 4: 106 hits total, 6 times HP lost;

This results in average chance of HP loss being equal to 4.57%, standard deviation of +/- 1.13%, with probability of statistically significant difference from the stated number (5.00% chance) being less than 1% (Mann-Whitney U-test for small sample sizes).

That's the pure math, nothing subjective.