I’m beginning to lose count of the number of time I had 70% chance to upgrade and ended up getting 0 out of 4. Or 1 out of 4. I just went through my 3rd straight where I got 0 out of 4.
Are those percentages sorta like attack power? What I mean is are they supposed to be a vague approximation of the actual chance but affected by other factors? Because if not I am just one of the unluckiest guys playing this game.
For me it was closer to 50%. Doing level 100 pits when that % hit 70, most of the time I would fail twice and hit twice per run. There were occasional times I’d hit 3 and even 4 fails in a single run. Also there were occasional runs were I hit only 1 or no fail. The majority were 2 and 2 though.
All of Blizzard’s RNG percentages are fake. They can deny it if they want but anyone who actually plays the game will see otherwise for themselves.
It’s the same for me.
3 or 4 fails at 70% 4 or 5 times in a row. Then maybe 2 once, and back to 3 or 4 fails.
It feels like a 70% chance to fail.
I’ve had it fail repeatedly at 90%
I’ve also upgraded some at 5%
Generally speaking, it feels ‘real’ enough; the caveat being that, as we all know, there is no true random (even if it isn’t deliberately weighted!)
Random is (mostly) random 
I’ll say it like this. It’s real like Santa Clause and The Easter Bunny are real. Technically a bunch of guys do dress up as them in suits…
Yes, for me it also feels more like 90% is more like 70% and 70% more kind of 50%.
It is frustrating. I had 8 55% fail in a row. Last week only hit 2 70s out of 12 in three runs.
But those are too small of a sample size. I bet had to tracked every single roll for all your glyphs from 1-100 the numbers would align.
It’s easy to get frustrated and introduce some sort of malice to the system when this happens. But people tend to ignore the times they got 4/4. I once got 3 10% and a 25% all in one go. I didn’t assume something was off just that I got lucky. The inverse is true too. Nothing is off it’s just unlucky.
In fairness, I wasn’t attempting to introduce malice so much as other factors that I could not be aware of (and that a simple game UI would not permit them to make me aware of. I understand that higher attack-power does not necessarily equal better damage. I was wondering if this is the same way.
But point taken.
You leave the QA team alone. 
I took my own advice and started writing down upgrade results instead of basing judgement on “feels”. My only reliable data is on 70% chance upgrades, since I did those almost exclusively.
Earlier today I was at 34/57 = 60% succesful upgrades. I though that was already off but kept going on.
Now I’m at 51/87 = 59%
This is not a small sample. The expected value is 61 successes and the chance of getting 51 or less is only 1.6%.
You can see the probability distribution here, bars for “51 or less” succeses are marked in pink:
https://ibb.co/h27GyYS
What I believe might be the case, is that UI displays only rounded “chance tier” values, while the actual chance differs under the hood depending on precise glyph-pit level difference.
I for the llife of me do not know why leveling glyphs to 100 is important , specifically speaking there is no tool tip that says “you get this power at glyph 100” it just appears the glyphs just rank up in power for each time you level it past 46 right ? or am I just not seeing the bigger picture?
On my cringeborn it was easy but not so much on my fire sorc. 
It’s a percentage for each roll if it’s 70% you have that same 70% every time you click the upgrade button it’s not a total 70% over all of the rolls put together.
everything in this game is weighted random, even if they would have you believe it isn’t
Probability of failing four times in a row at 70% = 0.3^4 =0.0081 = 0.81%
Yeah, I had plenty of 4 fails in a row at 70%. Just as I had plenty of 4 successes in a row. My purely unscientific feel is that 60% on D3 succeeded more often then 70% on D4. But someone would have to sit down and track it over 1000 tries or something to have decent test. People tend to notice the streaks particularly when they fail.