The BROKEN 3rd Ench/Mtrwrk percentile

Combat skill, glyph deck arrangement and positioning play a more major part in combat than RNG. There is always something else you can do to try even the odds against you.

I think it's the helplessness high level crafters feel when they come up against the bad side of the RNG gods that's fueling the debate. RL master craftsmen have all these techniques and tricks to avoid a disastrous outcome. There has got to be a better way.

 

I've pretty much said to hell with crafting in this game. Waste of time, and pointedly *un-fun*. I literally have more success and enjoyment crafting in my garage (and go through 100% less tools. )

 

look folks, I appreciate all the points of view but I need to stick to my original point.

The DAMN thing ain't working as the indicators say it should.

Since making this post I did a few more craft attempts & guess what -- more of the same.
There seems to be an internal negative draw from the chance to succeed which is even more potent when an item rolled WELL. Thus, crappy items seem to pass the 3rd round alot more (pretty much as indicated 49% im my case) as do well rolled items which break 75% of the time.

Your overall math calculation on the RNG can just be fine Chris, but there is a secondary factor that you or your team is missing.

 

A much better idea (surely thought if already) would be not to destroy an item on EC/MW fail, but rather end its ability to be further upgraded.

The fail message to be something like:
"You fail to be able to craft this item further"

At this point the item would get a label:
"Upgrade complete" at the bottom to denote it is no longer upgradable.

 

We did indeed suggest it... however it seems to me that the devs are stuck with this very wasteful model to be able to keep high end items rare enough so that the RMT side (which provides alot of funding for SOTA) is "fair and balanced".

This means a big gap between ok items and high end items. Which I am against but this is a story for another thread.

 

And here we get to the real complaint.

 

Is this a double post if I already thought about posting in the other thread?

 

noone is going to hunt you down Tahru

 

Speak for yourself, I'm already outside.

 

haha


Well.. your... logic is impeccable.

 

after testing /roll 1000 with 1000 samples, i've concluded the average and distribution of values this PRNG generated is very random. Since I still believe in the crafting bias, i have 3 theories:

1) I haven't done any analysis on value proximity or "streaks"
2) i did the test during a very quiet time on the server, it could be entropy exhaustion during load
3) perhaps the logic is faulty and e.g. a 48% success chance, is not simply doing (rand(100) < 48).

 

1- you indeed need to make this analysis, and its easier with a roll 100 so you get a proper relational pattern, a sample of 100000 rolls would be ideal
2 - entropy exhaustion, well as we've covered its a pseudo random generator with multiple layers of entropy seeded distribution hubs. It's a nice custom system.
3- The logic is as intended, I'm with those who think we should not even see the chance of success Of course there are many factors, and there will be many more that influence your chance of success and it might not yet be reflected in the crafting UI...

 

I just wish that the
/debug
log would actually log the exact value rolled against.

That way anyone could get some closure by doing 100 enchants and see the distribution in the log.

Instead we are blindfolded by staring at the UI display % and only seeing fail/success.

 

The problem may not even be the RNG, the RNG may be inside a total formula thats flawed. Unfortunately, we have complained about it as long as I have been here and Chris several times just slams us about it what I would consider rudely (for example the last post) and they also make rude remarks about this complaint on the telethons. I dont feel they are listening to their wide complaining customer base on this and taking it seriously (yes, I am screen shooting this Berek in case you delete my simple opinion and observation here. I am expressing a continued concern as well as a concern as to how it has been handled and how people are treated when complaining about it considering I pretty much quit crafting over the issue myself). People have asked for a long time to see the actual data in the game about this with no response at all. I can accept the results if data is provided.

 

I'll let you tackle that one without macros even 1000 samples was a pain. pro-tip: [enter][up key][enter]

however, even at 1,000 samples my average converged at 499.5 so there's really no need for further samples. also /roll 100 and /roll 1000 should have equal distribution. I assumed there was a decimal place under the hood but it should be the same thing.

for more details, i also made 2 types of histograms: one bucketing by integer division by 10, and one by tailing digit. in both cases, the sota random had fewer outliers than ruby's rand()

 

I have noticed something odd very with crafting rolls too. Every time I make a suit of armor, if something breaks, it's always the legs.
The last suit I made five sets of legs in a row blew at 85%, and I don't think I have ever managed to craft a full set of 2 & 2 without loosing at least one set of legs.

 

OP's original item is still broken, FYI. There's no point crafting stuff beyond 2nd MW/enchant unless you can stomach the massive amount of item breakage on the 3rd MW/enchant (i.e. that you should not be getting until the 4th MW/enchant). Should be around 50% success rate on 3rd attempt but it's more like 15%.

My very simple math head says that means instead of a roughly 25% total chance of getting a 3rd MW and enchant on an item (i.e. .5 * .5 = .25), you are getting about a 2.25% total chance (i.e. .15 *.15 = .0225).

So, make 200 items.
Enchant once, lose 10 of them (190 left, 95% chance).
Masterwork once, lose 9 of them (181 left, 95% chance).
Enchant 2nd time, lose 29 more (152 left, 84% chance).
Masterwork 2nd time, lose 24 more (128 left, 84% chance).

From there it should be enchant a 3rd time, lose 68 more (60 left, 47% chance), then masterwork a 3rd time, lose 32 more (28 left, 46% chance). So out of 200 pairs of epic leather boots, you now have 28 that have 3 MW/enchants on them, of varying usefulness, and you have maybe 5 or 6 that have pretty choice stats on them, the others have a hodgepodge due to the randomness of the enchant/masterwork selection process.

Except instead on that third round, you lose 85% of them each time. So that 128 you had left turns into 19, and that 19 turns into 3. Probably none of which has anything near any coherent groupings of stats like you were hoping to hit on.

I'm starting to believe maybe it's intentional to keep the number of very high stat crafted items intentionally low until after release, or something. Because it keeps getting mentioned all over the place that the 3rd attempt is using the % of the 4th and no one seems to be able to take 5 minutes to check that.

P.S. For those of you not deeply familiar with crafting who say "hey, that actually seems too high, ending up with 28 pairs of 3 MW/3 ENCH boots that probably will sell to somebody ffor a pretty penny," keep in mind that in order to get there you burned (if % chance was actually working correctly):

- base mats and fuel for the 200 items
- 1000 gold ingots and 2000 mandrake root on first enchant
- 950 silver ingots on first masterwork
- 905 gold ingots and 1810 mandrake root on 2nd enchant
- 760 silver ingots on 2nd masterwork
- 640 gold ingots and 1280 mandrake root on third enchant
- 300 silver ingots on 3rd masterwork

That's 2,545 gold ingots (10,180 gold ore + 2,545 coal) and 2,545 mandrake root and 2,010 silver ingots (8,040 silver ore and 2010 coal).

At current prices that's around 1.1 million gold in mats. For three pairs of pretty nice boots...well except you wanted mainly dex and DA...well, you got some of that anyway. Good on ya!

P.S.S. 44% chance I messed up my math somewhere.

 

if your going to have percentages they should be right !!!!!

 

My guess on this is masterwork/enchanting percentages work diferrent then crafting percentages.

For example you can craft 100 items with say 29% exceptional chance and find afterwards you got 25-29 exceptional items wich of course is an accurate reflection of your crafting procentage.

The enchanting /masterwork works different basically with 49% on third try doest mean half of your items will succeed third enchant it means that for every item you have an almost 50/50 chance for it to succeed. So basically you can roll the dice on the negative side and lose 5-10-20 times .What still puzzles me to this day is why is it we can literally fail 10-20 items in a row at 49 % but you never get 10-20 items to succeed at 49% in a row...Highest streak i got at 49 % was maybe 7 items in a row to succeed. I also agree with the op that junk items seem to have a higher chance to pass third enchant while exceptional or good enchanted items seem to blow up more often then the rest.

Anyway my ''conspiracy theory'' is :

1.Theres definetly a math involved that we do not understand because of lack of information provided. And we're not being told because it will discourage people to waste theyre hard earned resources.
2.Crafting process has been deliberately made this way to help clear the huge amounts of ore in the game and keep the economy going.
or
3.Theres and error in the coding which if ever will be adressed we are not gonna be told about it because imagine the rage of all those people who lost alot of time invested in harvesting those resources. My hope is this will eventually be adressed and we gonna wake up one day battling the 15% enchant and forget all about the way things were.