Sorry if the tone is a bit amiguous, or inconsistent. I was going off on one, not
going off on one . It wasn't personal or intended to be talking down or something. Mostly talking
past you tbh. I tend to ramble, cut down because it was too long, and then post the nonsensical result.
"If I had more time I would have written a shorter comment" If it helps I was going for friendly devil's advocate, not authoritative here's-why-you're-WRONG source.
For all I know you could be right. Maybe it displays 90 but really only goes to 70 internally? Or wrong? I don't know. You don't seem to know. Between us we're probably not going to aquire the teraflopyears needed to fill and process a spreadsheet large enough to definitively conclude either way.
I was offering an explanation (mostly for myself who frequently runs into similar-feeling situations in other RNG-heavy games, and by exetension anyone else feeling similar).
Html games doing rng is pretty bread-and-butter standard. As is savescumming them to get the result you want. As is sitting there clicking 2 buttons for 40 mins straight getting frustrated at failing until you win and then immediately clicking the button again because you're stuck in muscle memory and getting even more annoy-
Maybe each one collectively copied a bad RNG implementation? Maybe I'm cursed to get bad rolls? Maybe the more times you roll a dice across your lifetime, the more likely you are to have seen (among all collective rolls) a memorably-long string of 1s?
Just because you know a fallacy doesn't make you immune to it.
You are not immune to propaganda. Mate, I will never emotionally recover from gambler's fallacying
the bioshock infinite coin flip. Contextually it's hilarious: By doing that, I managed to make the only
objectively wrong choice (drawing that specific conclusion from that information). Doesn't make anyone a "moron" for not being eternally vigilant. We're wired for pattern recognition to save brainpower, not for statistics. Interpreting statistics accurately is primarily fighting against our instinct to draw patterns where there aren't, or to draw incorrect conclusions by accidentally omitting vital situational context.
At least that's what I tell myself to feel less thick when I forget something obvious that I already know.
On this version and earlier versions I
subjectively feel that I had more success after peforming a similar ritual (loading a different save and then reloading) but it could equally have been placebo. Especially since I only do that after scumming "normally" for a while without success, and I stop when I get the result I want. This skews the result by several orders of magnitude. Which is further amplified by basing the success rate on what I remember, which comes with a free negativity bias (you naturally remember more of when the RNG robbed you of a "free" win more than when it gave you a "hopeless" win). Also when I win, I'm going to continue playing not bother writing it down or commit the event to long-term memory. Knowing this objectively doesn't change how it feels subjectively.
And I can't prove that ritual did anything practical like clearing a variable / resetting a stuck loop or something, bar actual variable checking. Which I am far too lazy to do on a savescummable html game (one of my many failings).
Your original comment is about feeling some kind of bias, which is subjective. It never feels fair when you fail a roll for an event you want, especially if the success chance looks achievable. What I was trying to say was that that same sequence of observed results could have been obtained from a real or a bad RNG implementation. There is no way of knowing for sure solely by looking at the results.
But this game is savescummable, and as far as I know the outcome of a roll isn't baked in to the savestate :. I trust it to be doing
some kind of mathsy numbo jumbo behind the scenes. But equally that trust may be misplaced, html games can lie (wasteland lewdness has a stat check when you take a pill at the start, bit it's unpassable and probably a completely fake check).