Tempering is good and bad , love the flexibility , hate the anxiety inducing rng , if I hit smoke grenade size one more time , I will scream , wait already did that! 
Hey are game is so shallow and empty that we artificially have to increase player time when finding and tempering gear, this will keep the player playing longer and possibly buy more mtx and in no way will people be frustrated or quit because when blizz implements a new feature they are always right and its the best thing to ever happen to the game for sure a player issue /s
Did you get the extended warranty bud? Those things are generally a scam but occasionally worth it.
And yes, yippee for āDamage to Distantā on bash barb 17x in a row.
Without empirical data, though, this wouldnāt tell us much. Again, we get a lot of reports on the forums about times that the system behaved āoddlyā from the userās point of view, but those behaviors are biased by typical patterns of play and by the tendency to report surprising negative results but not report surprising positive ones.
So the first thing to do would be to choose the manuals you think arenāt working, and before you salvage your next inventory full of trash, temper those manuals onto a whole batch of items and record the results.
I think this is probably just superstition, but you can also do this by just rolling on some random other item if you want to just break things up.
Literally all items systems are systems to āartificially increase player timeā. They could just let you select all the affixes and rolls and never farm for one. But they donāt because farming for items is the whole fun of the game.
That sucks but rng is rng. Many times itās cruel but when it gives you what you wantā¦ Joy!
1 Like
I agree. (just getting back).
This is what I will start doing. I wanted to whip-up a a quick tracking app so I can have a written record (just easier data entry and pattern recognition). I might do a preliminary full inventory without it, just to get some quick data.
Maybe my feelings are an amalgamation of only remembering the bad rolls and isnāt representation of the actual numbers. I can admit to that. Iāll run some controlled temperings this evening and see what the preliminary data shows.
Iām interested to see what you find. I havenāt done a systematic test yet myself because every time I start one the results just seem so random it feels pointless.
I think itās highly unlikely Blizzard coded some weird mechanism to specific screw people over and make their experience with tempering worse. But itās certainly possible that some of the individual manuals are bugged or that affixes with different roll ranges are somehow treated differently.
I think bringing proper statistics into it is like frosting a moldy cake. Systems in D4 are notoriously prone to ābugsā and unforeseen interactions. Also, dev communication is such that itās unreasonable to expect a problem be publicized & addressed even months after itās identified informally. So, at best, an informal pattern can be countered with āit could be working as intendedā. Thatās the result of bad coding/ poor QA + poor communication.
Setting that aside, gear overhaul since S4 was to replace 300 randomized bloat affixes with 300 new affixes (ātempersā) which the player has a hand in selecting. The sheer time investment makes it a wash (when the goal of the S4 revamp was supposedly to help cut down on wasted time staring at garbage loot). If this is working as intended (unannounced ābugsā and all) then we were expected to not only equip a bunch of 4/5 affix items - where the 5th affix is a dud - but also expected to invest Masterworking loop to improving said gear.
Otherwise - if those MW mats & time investment are considered wasted on 4/5 gear - then we were expected to continually farm de facto garbage until 5/5, before MWā¦ meaning the S4 loot revamp wasnāt just a wash (in time investment) it was designed as a time extender, perhaps to counter the steady shortening of the 1-100 journey.
Either scenario fits this.
Yeah Iām thinking a quark in the code over malicious. Iām thinking back on how the Grandfather had one of itās affixes āhiddenā but could still be rolled. I know this is a different system but just an example.
When I get some prelim data Iāll post it here.
Here is the post I am referring to. As you said, the calculation is easy so I am assuming the numbers that are given are accurate.
Feel free to provide your own Chi^2 test results to support your statement.
The actual chance of this happening is 1/(.33% ^ 14) = 1 in 5.5 million.
You have 7 rolls with a GA ring. The 6 rerolls and your first roll. This is 14 rolls in a row on a three option roll with 2 rings.
This is unreal bad luck.
You sure about not hitting anything else?
I know this is a laughably low sample (10) but itās a tedious task, one id only continue with a tracking application setup on my laptop (so I dont have to tab out for each entry/result).
The setup: I went with the idea these were going to be true upgrades for my current build, I waned DD Damage and Berserking Damage, without getting both, the item would brick. This was also chosen as each manual has three tempers. A lot of manual have 4,5 and 6 tempers, meaning things would brick faster; that said a 1 of 3 manual should be the easiest to succeed.
As soon as the required temper was imprinted, I moved onto the next.
Sanstorm Augments: (1) Dust Devil Damage, (2) Dust Devil Cast Twice, (3) Dust Devil Size
Berserking Finesse: (1) Berserker Damage, (2) Wrath Damage, (3) War Cry Damage
Goal Results on both manuals = 1
Item 1
Sandstorm 1
Berserking 2,3,3,3,1
Item 2
Sandstorm 3,3,1
Berserking 3,1
Item 3
Sandstorm 2,1
Berserking 1
Item 4
Sandstorm 1
Berserking 2,3,2,1
Item 5
Sandstorm 2,1
Berserking 1
Item 6
Sandstorm 1
Berserking 2,3,2,3,1
Item 7 Brick
Sandstorm 1
Berserking 3,2,3,3,2,3
Item 8 Brick
Sandstorm 3,3,3,2,3,1
Berserking 2
Item 9
Sandstorm 1
Berserking 1
Item 10 Brick
Sandstorm 3,3,1
Berserking 3,3,3,2
7 successful, 3 bricks.
Given the small sample here, we can see that 3 is favored, but due to the small sample size this can still easily fall within the RNG. There are a few streaks of the same number in a row; again that can fall within the RNG. I did not bump into one of the unlucky streaks where the result was 5+ in a row. A larger sample size is needed. I used all weapons for this sample, and noting with a GA. Rolls also include the free, first, rolls.
This is why people are asking if its bugged.
I was rolling for bash cleave and landed upheaval size 7 times in a row.
Iāve bricked plenty of items but seeing the same roll come up that many times in a row just makes it feel like something MUST be wrong. There needs to be a streak breaker or something.
Sounds like a ātemperā tantrumā¦
The Murphyās Law of Tempering:
Trash or āAcceptableā Temp Gear = Top Rolls or desired rolls no sweat, no muss no fuss.
GA Items = Overpower, Overpower, Overpower, Overpower, Overpower, Overpowerā¦ 
%Chance to Heal ā¦
Iāve bricked every single GA item Iāve found.
I know it will brick and accept it.
What I DO NOT ACCEPT IS the lack of the simplest feature:
KEEP CURRENT ROLL
Itās shameful that a system was designed to make players totally reluctant to flat out HELL NAW! when it comes to Player Agency and improving their gear.
Get desired Affixes 1st 2 rolls and theyāre minimal values?
Stop. Done. Over. Finished.
I donāt mind bricking at all.
Not being able to IMPROVE my item when I was lucky to hit the Affix without going back to blind roll risk it all?
Bad decision.
1 Like
I like it and it jives with how enchanting handles the same situation.
It is a small sample, but it looks quite random. The berserking one had 14 3s, 8 2s, 7 1s, and the other had 10 1s, 8 3s, 3 2s. If those numbers continued it would of course indicate that some weighting was involved, but one thing to note is that the variance of the not-desired affixes is always going to be higher in a fixed-size sample.
That is, if you temper 10 items, the most instances of the desired affix you can ever get is 10. But the floor is higher too because you will never stop early unless you hit the desired affix or ran out of rerolls. So if you look at just the rolls where one or the other slot stopped early (with rerolls left), you should find a higher percentage of those contain a 1. So the number of 1s is going to vary within a tighter band (in this case 7 and 10) than the number of 2s or 3s (which vary from 3 to 8 and 8 to 14 in this sample). The second slot will also have more variance on the 1s because it will end in bricking far more often.
Anyway. It does seem like it could be worth collecting more data on the ācast twiceā affixes - thatās the one outlier here.
1 Like
Yeah, I agree with this - they need to give some way to spend the extra rerolls that people are willing to actually do, either by reducing the cost of failure or by making it impossible to lose what youāve already rolled.
1 Like
I think what Im going to do moving forward is to record upgrades along my characters path to 100, current level is 81 (second character) and only has a few 925 pieces, so lots of upgrades to come.
Right after I completed the above test I went to upgrade a GA weapon, and it came really close to bricking. I was thinking to myself, ohh the irony! hah.
Yeah it feels real stupid to finally find an item with the affixes on it, especially if itās got a GA, only to brick it and have it be completely useless.
I have no doubt theyāll change this, but itāll probably be changed after far too many people are burned out by it.