At least you recognize the limitations of your sample set.
The next step is reject conclusions based off of biased small sample sizes.
You'd like me to reject a conclusion based off a biased small sample size? Sure: I reject yours. It fits that criterion, after all. ^.^
Tongue-in-cheek quips aside... The conclusion was that players are feeling this way, based on the testimony of players who are feeling that way. I make no definitive statement as to the magnitude of the effect, just that it is occurring - Larian are literally the only ones with the data to verify how pronounced the effect is in real terms of actual player choices, but it is what it is nevertheless - the vast majority of all who have spoken here and elsewhere publicly to the topic have felt this way, with precious few speaking against it. That is the only conclusion being drawn, and it's perfectly valid.
...it is not my experience that saves are made at an unreasonable rate nor has it caused me in any way to change my spell selection in game. [...]
I, however, am just a single data point amongst many and do not speak for anyone's experience but my own.
You
talk with a wording and tone that acts as though your isolated data point of experience is somehow more valuable and more potent a case to make that it should somehow negate or counter the reported data points of
several others, and should quiet them. No. You have your experience, and that is all. You are also an unreliable narrator, just as everyone else is, and you have your own non-objective bias about your results, as everyone else does. "It hasn't been your experience that..." if you can poo-poo a dozen other people for saying that, and write them off, dismissing them as non-objective sources that
shouldn't be listened to, then you
MUST do the same thing for yourself. Perhaps it wasn't your intention to project yourself in that manner - if not, then what I said here doesn't affect you, naturally - but please do take it under advice that this was how your posts came off, to me at least.
This forum is designed for feedback, so please keep giving your personal feedback.
Larian will balance that feedback with actual data and make their decisions accordingly.
That's what folk here have been doing - You are are the one who came into this thread just now and attempted to shut them down, dismiss and discredit that feedback and say that it shouldn't be listened to - please don't do that. Give your feedback (You've done so, thank you for participating ^.^); please do not jump into threads to tell other people that theirs is not valid.
What I am curious about, and something you've not voiced an opinion on yet, to Alodar, is... As you point out, mechanics in 5e have a degree of favouring towards landing attack roll spells, over fully landing ST spells, and this is intentional in the design based around other inter-woven elements of the system and how it is built. It's been pointed out that the majority of Larina's homebrew mechanics and implementations push this balance even further in the direction of favouring attack roll spells - the ways in which they do so have been listed and explained multiple times, and they're not opinions, but legitimate facts of mechanics... what's your stance on that? If you're aware that the success leaning already favours attack roll spells, and Larina's homebrew and implementation choices pushes that favouring even further, without any kick back to ST spells - and indeed adds things which hamper a large section of ST spells further than usual - do you feel that these additions are a good decision?
==
The only thing it demonstrated was the woeful lack of understanding most of the folks involved had about statistics.
Fortunately for us, the ones aggregating the results and illustrating them visually were not amongst those people ^.^
The data was a large enough sample to show a distinct flaw in the rng and the rate at which it sourced chaos to diffuse the algorithm it uses. Because it was based on a mathematical algorithm, it still produced the suggested averages in the correct degree, even over a large test set (this was the initial reason why Larian originally dismissed the feedback; they later
revised this position, which should tell you something), however, it did so without sufficient chaos or reseeding and created a visible pattern in doing so. This normally would
not be a problem, as long as the overall averages and results were correct - and they were! - but within the context of a video game that relies on successive dice rolls to adjudicate events, it does become a problem because it increases (marginally) the likelihood of successive poor, or good rolls over something with a better randomness emulation - and in the case of a system that uses elements like advantage and disadvantage, or contested checks, this has a legitimate impact on gameplay outcomes.
As other have mentioned, the related systems have been changed since that time, and I have not performed another in depth test since; it may have been addressed. If I did another test now, however, and found that the algorithm did not appear in the results, and that it did, indeed, appear to be sourcing and seeding better now, I would still maintain that it would be dishonest and disingenuous for anyone to insist that 'nothing had been changed' between the two data sets, then and now; it would be an irrational claim, given the data shown.
