Thoughts on Crafting 2.0

EDIT: Ignore almost everything in this opening post. On 10MAY16, they revealed that they were mistaken during the 29 April Twitch regarding only getting 1 of the crafted items back. Now, jump down and read post 17.

Last night I watched Friday's Twitch, and I was most disappointed when they talked about how the revert function would work. Up until that point, it had seemed that the only question was if it would be free or cost something, but there had been no hint that you were only going to be able to salvage one of the four items that went into the crafting bench. I know that I shouldn't be surprised that BigPoint would be giving us the short end of the deal, but I had hoped for more.

Well, the revelation that you could only get one item back from a bad craft has made green and blue items with at least one good stat significantly more valuable, and blue items with 2/2 good stats absolutely invaluable. By that same token, legendaries with only one good stat are now garbage. This is most distressing since I had a couple of legendaries with nice stats. Too bad that we can't pay to deconstruct an item. I'd take the legendary and break it down until I got a blue item with the good stat item.

That being said, I can understand why they did it. If they didn't, getting a godlike item or even an uber item could take as little as a couple of weeks. But with the introduction of this new mechanic, it will take significantly longer... "How long?" you ask? Well, lets take a look.

Here is my thoughts on how to best go about crafting a good/godlike item.

TL-DR
It will take a ton of time and dedication to get a godlike item with any dependability... that, or tons of luck! Short of tempting fate and putting items together and hoping that what you need pops out (with probabilities as low as 0.5% for a "good" legendary or 0.05% for a godlike legendary), getting an uber item legendary will take months, if not a year.

Spoiler: Crafting a "Good" Magic

Start with a green item with a good stat. You don't need to waste 4 green items with good stats to make a blue because you'll be able to recover it should you fail to pass it on. So, for each green item, you know that you can get a useful blue if you spend enough gold. Probability says that it will take an average of 4 tries to pass the stat on, but given the low price to craft green items, this isn't a big deal.
---So, given enough gold, you have a 100% chance of getting a good blue.

Spoiler: Crafting a "Good" Extraordinary

Next step is getting good blue items to turn into good purple items. Here is where the stakes go up. Now, if you have a bad craft, you will only get one of the useful input items back. Therefore, it is in your best interests to make your odds as good as possible. Instead of gambling one or two good items in the hopes of a good purple, I think it is better if you put in 4 good blue items. In that case, your odds of getting a purple with 2/3 good lines (from the 4/8 good lines you put in) will be about 21% ((4*3)/(8*7)=0.214). This is not exactly in your favor, but it is better than if you used two items with one good stat each. Then your odds would be 3% of getting an exo with 2/3 good stats.
---So, given 4 good blues with one good stat, you have a 21% chance to get a good exo. To average a "100%" chance of getting a good exo would require attempting this process approximately 5 times and failing 4 times. Since you can recycle one item from each fail, this comes out to 16 (=4+3+3+3+3) good blue items (or good green items + gold) to get one good exo.

Spoiler: Crafting a "Good" Legendary

Now, lets look at getting a good legendary, where good is defined as 3/4 stats are the ones you want. Assuming you've gotten 4 exos with good (2/3) stats, you are looking at passing three of the 8/12 good stats on to the legendary. This has a 26% chance ((8*7*6)/(12*11*10)=0.2545). This is better than trying to get a good exo, but it is still a steep price considering the effort it takes to make those exos.
---So, given 4 good exos, you have a 26% chance to get a good legendary. To average a "100%" chance of getting a good legendary would require attempting this process approximately 4 times and failing 3 times. Since you can recycle one item from each fail, this comes out to 13 exos (=4+3+3+3). Using the average of 16 blues/greens per exo, and to get a good legendary will take 208.

Spoiler: Crafting a Godlike Legendary

The the game doesn't stop here anymore... what about that "Godlike" weapon or piece of armor? Well, it would take 4 good legendaries with 3/4 good stats each. From here, your odds are actually not bad. You are putting in 12/16 good stats and only need 4 good ones to come out. That means you'd be looking at 27% ((12*11*10*9)/(16*15*14*13)=0.272), still roughly 1/4, but your best odds yet. That being said, given that each of those "good" legendaries cost you 208 good green/blue item, this it no small feat.
---So, given 4 good legendaries, you have a 27% chance to get a good legendary. To average a "100%" chance of getting a good legendary would require attempting this process approximately 4 times and failing 3 times. Since you can recycle one item from each fail, this comes out to 13 legendaries (=4+3+3+3). Using the average of 208 blues/greens per legendary, and to get a godlike legendary will take 2704 good blues/greens.

Spoiler: Minimum Number of Magic/Improved Items for a "Godlike" Item

I said above that this was on average. Statistics say that some of the time you will get what you need on your first try, and other times you will go over 4/5 tries to get what you need. That means that it ought to average out to 4 or 5 tries to move from enchantment level to enchantment level. If you were 100% lucky and each craft went your way, using this method would still take a minimum of 64 good green/blue items. Of course, there is no max, because if you luck sucks, you could try for ever and never get a single good legendary, let alone an amazing legendary.

Spoiler: Exceptions

I mentioned above that blue items with two good stat would be invaluable. Here is why. If you are trying to craft an exo with 2 good stats, by including that double item,your odds go from 21% up to 36%. Also, because you would salvage the double in the case of a bad craft, you would only need to attempt 3 times and fail 2 (on average). That means that making a good exo drops from 16 down to 10 (4+3+3). Alternatively, if you get a good stat as your random stat, but one of the stats you were trying to pass on didn't show up, this saves you from having to revert. Now, I'm assuming that these good items would be rare enough (because it involves the random stat being a usable one), that it is unlikely to make a significant impact on the 2700 number, but it may drop it by a hundred or so. Also, if you get a lucky 3/3 good exo, your odds of making a good legendary go up from 26% up to 38%, nearly 2/5. The same holds true for making exos into a legendary, but the difference is that if you get 3/4 good stats from the ingredients and the random stat is a good one, you will have achieved your goal incredibly early, maybe after only 200-600 green items, but this is luck that I wouldn't count on. Even factoring these lucky breaks in, I would still anticipate that it will take 2000 green/blue items to make a godlike legendary.

The other exception is a "safe" gamble. Providing you are not hurting for gold, this could be a method that could significantly shorten you process. If you have one good exo item, say the one you are using, and you find 3 other exos (emphasis on find, failed crafts should still be reverted) that have a couple of good stats on them, but not 2/3 each, there is nothing to stop you from throwing them on the work bench and seeing if the outcome is better than your targeted item that you've been saving. As long as you don't mind spending the gold and losing out on the 6300 glyphs of the items that would be destroyed upon reverting, then you stand very little risk with the possibility of high reward. This is what I'll be attempting early on to make use of this new system before moving on to the more dedicated long term crafting described above. It will let me clear large chunks of my inventory that are currently clogged with exos and legendaries that won't be that useful after the revelation in Twitch.

Spoiler: So it will take months to get a usable item?

To answer the title question, no; it isn't all doom and gloom. The thing is, as you go through this process, it is highly likely that the good exos and legendaries you craft will be a significant improvement over your existing gear. For example, if you craft an extraordinary 2H Axe with 150% weapon damage and 80 damage, you will likely be able to replace your existing 2H legendary (unless your existing one is really impressive). That will only take the estimated of 64 good green/blue items. Then, because it is an exo, glyphing it to 55 will be relatively easy while your search continues. Then, when you get 3 other good exos, remove your glphs from your main weapon and craft the four of them them together. If you get a good legendary, it will take the place of the exo. If you get a dud, you keep the best exo and hope for better luck next time. In this way, you'll just keep getting better and better on your quest for the uber weapon, but instead of a crapshoot in the dark like it is now, it will be a slow and steady progression.

 

Some very insightful number crunching there. I was a bit shocked at the revelation about reverting in the last twitch. I got over that pretty quick though since from the onset it was stated that they did not wanted 2.0 to create uber items over a short time period.

I do I think I see an angle here. Premium players will get a slight advantage with crafting 2.0.

For the past few weeks my game time has focused on the two accounts that I bought 90 day premiums for back in Feb. as those wind down (I already canceled the auto- renew on them.) The reason is that more often than not what I get from a mystic cube is an exo item. My inventories are clogged also as I have saved all the exos (as well as a few 4 of the same legendary) from cubes and drops with the best stat lines.

Now, since it will be folly to craft the the ones with good stat lines together, I will have to rely on future drops alone or renew my premium to get more crappy exos to craft with the good ones I have saved to accelerate the process.

For me the decision is a no brainer. I returned to this game with no real goals and the only reason I did the premiums was to help me adjust to the changes in the game and progress my favorite old level 45s. At this point I have a decent enough start on endgame play and I'm not in any real hurry.

Luck be with ye,
Mikey,
Tegan

 

Almost got a headache from all your numbers, almost, but when crafting does arrive I will post the results here from my legendary craft.
For testing purposes I prepared a lot of legs: 4x leg rings cca 1500 HP, 4x leg rings that hav 2x res 2x HP (mid range, meaning 600-700-800 each), 4x leg gloves varying from 17-24% speed, 4x leg shield (almost perfect shields), 4xleg boots 17% speed, 4x adornments with crit 830-850, 4x armor with 15-16% incr. dmg, 4x pink gloves with also 17-19% speed, for now I have 3 pink hammers all based on strong incr. dmg % (waiting on 4th).

After it arrives we will see exactly what god/junk items come out. At least I'll get the clutter from my inventory cleared.

 

So if i got 4 exo hammers, let's say:

Hammer A : 45% dam, 25% dam, 50 dam
Hammer B : 30 dam, 40 dam, 30 dam
Hammer C : 30 dam, 20 dam, 60% dam
Hammer D : 42% dam, 48 % dam, 90 dam

With craft 2.0 does % crit dam or crit line may still occur as result?
or the result only limited only for % dam & dam as the source Hammer only provide those 2 kind of lines?

 

Three of the four stats on the legendary will come from your exos and will unquestionably be damage or % damage. The fourth will be random and could be any of the four kinds of stats that can appear on a 2H weapon. Crit and crit damage are very possible for the last stat, and actually equally likely as damage and % damage.

 

As always thx Baragain

 

Sounds like too much work to me, and not much different to what we have now hoping for good stat lines.

Time will tell I suppose.

 

I've been farming as much as the error 23 will allow me to find better than average weapons in anticipation of crafting 2.0, only to find out that 3 of the 4 will be lost. Even though I knew it in my bones, I so wanted to believe that they might just listen to the players on this one. I kind of feel like someone should at least buy me dinner and tell me that I'm pretty.

I know that there was some fear on people being able to create a god like weapon, but you still had to find 4 good items with good lines. That was enough to limit the chances. Punishing us for trying to craft is a bit much, even for DSO.

I'm curious to see how they "adjust" the price of crafting.

 

I have a feeling like we were expecting too much ... and they trolled us again.
What is the point of "revert craft" if I loose 3 items.
This way nothing changes significantly ... green/blue/purple items are still the way to go.

 

Don't expect much gents or you will be deffo dissapointed.

 

Well, don't expect to throw 4 random items in that happen to have a couple good lines and expect a god like weapon. You do what I describe, and you can have reasonable expectations over time. Patience is the key.

 

Proper implementation is key, to which the DSO team has one messed up vision on how that key is supposed to look.

 

I, for one, think that the crafting 2.0 system is completely fair and balanced, because after all, you are not going to be the only one crafting.
Think about it: random drops would have no use, so bosses would be useless, and people would just farm in high dungeons and pw and get perfect stats.
If they didn't, they'd just try again. The only limit would be gold.... right?

wait, WHAT? Devs are actually listening to players: buyers/people with tons of ander do not get a real advantage!!!

this way, people still have value for all tiers from green to leg, and people don't run out of gold!

anyways, this is FAR better than what DSO has right now in terms of crafting...

 

Hmm, I was also surprised to read that. That Twitch statement back then also lead to quite some fuss and discussion in the German feedback thread.
But as I interpret it, it was not the devs changing their mind (again)... actually many people said it seemed reasonable to restrict the revert function so as to make it not too OP... @Spieletrend (from my p.o.v. correctly) stated that the restriction would move crafting 2.0 from "Pay2Win" to "Farm2Win".

Rather I think the CMs had, in that Twitch session, chosen an example no one was thinking of any more, i.e. crafting with different inputs... (which had been excluded in the published draft design and probably later added again).
I do like Haruki and her usually well researched replies, but sometimes I get the impression relying on Twitch information too much isn't wise.

Thanks for your math, btw. I had done some similar calculations in the German thread BEFORE that twitch. We agree on the basic calculation mechanisms, so I suppose they are correct .
My probability posts can be found here and here, though I suppose the German text may not be so interesting for most people here.
(I ended up with an estimated, expected cost of 3500 gold for crafting 1 legendary with 4 desired input lines from 256 green items, or 137k gold for crafting it from 4 legendary inputs with 1 desired enchantment each (Chance of 1:1820))
Sorry, but I am too lazy to translate it now, given the recent U-turns in the "published" game design.
We can discuss the stuff when the final design and costs have been revealed .

 

I still do not understand how reverting makes anyone believe it will lead to p2w of farm2win.

You still have to find the 4 pieces with the right kind of enchantments.
You still pay the gold to craft.
You still deal with the random number generating gods with the stats and enchantments.

No amount of money helps with the drops. As someone who has spent years still trying to find what most would call a "good" 2h weapon, something needed to be changed.

Of course some will end up with god-like weapons, some already do. Some of us however are going to need more than a few times to get something worthy of posting in the show off threads.

I'm happy to see that we won't lose pieces if we craft 4 like items, but before I get all excited, I'm waiting to see what the "gotcha" will be.

 

Well, if you get all input items back by reverting, you could just take 4 legendary items with 1 great stat line (not difficult in PW fatal) and re-roll until exactly those 4 stat lines are taken over to the end result (chance 1:1820). Just needs an unlimited amount of gold/andermant and very little farming.
Finding about 2000 useful input items, as @_Baragain_ calculated in the opening post, needs a substantial amount of farming.

 

OK folks, here we go...

As seen in the CM thread and some of you have already commented, the revelation that you will get all 4 items back means that getting a good item goes from a herculean task to something possible for us mere mortals. What I will do first is look at what the odds are for promoting good stats at various levels. Because you will get all 4 back, it is now less about playing the statistical long game and more about sequentially concentrating good stats no matter what level of enchantment it is found on. So, for this reason, I'll look at the odds of passing a given amount of stats from one enchantment level to another with the most in depth given for the crafting of 4 legendaries.

I'm also going to be using a notation that I've figured out to better explain good stat items vs crap stat items that is similar to the way we talk about shields or swords. Right now, we talk about 3/1 or 2/2 for a shield and we (probably) mean block rate/block strength. In my system, I'll use good stats/total stats. Using that, a 1/1 would be an Improved item with a good stat. 1/3 would be an Extraordinary with one good stat. 2/4 would be a Legendary with 2 good stats. I hope this makes sense because I'll be using it a lot in the next couple paragraphs.

Improved to Magic:
Because you can only pass along 1 stat from a green item to a blue item, and the odds of passing 1 stat on is 25%, and green to blue crafting is cheap, I'd still recommend doing the craft with 1x 1/1 item and 3x 0/1 throw away items (all of the same type of course). Because of the 100% return on revert, that means that you can try this 4+ times in rapid succession without having to find a ton of 0/1 items. Your odds can go up to 50%, 75%, or even 100% if you use 2, 3, or 4 1/1 green items, but in my opinion, that is a waste of good items even though it could save you gold. In the end, the benefit is outweighed by the cost.
3x 0/1 + 1x 1/1--->1x 1/2
25% Chance.

Spoiler: Read this if you want, but some inconsistencies are bugging me and I need to double and triple check.

Magic to Extraordinary:
Here is where you need to start considering using more than one item with at least 1 good stat. To begin with, you can pass along two stats from the magic to the exo, so it is a good idea to have at least 2 good stats in the mix. Here are the odds for a sampling of the possible options, including 2/2, if you were that lucky.

Long story short, it is easy to pass 1 stat on. It ought to take about 8 tries, but it will give you a 1/3 instead of a 2/3 that will be very useful in crafting a 3/4. Personally, 3.5% for 2x 1/2 is not good enough for me. I don't want to take a 1:28 chance and have to do that an average of 28 times. If each craft takes 11 gold, that is 308 gold just to get a good item on the second tier. Compare that to using 3x 1/2. The odds are almost triple. Trying 10 times at 10 gold for a cost of 110 gold would be much more reasonable. This of course is dependent on if there are any price changes and what the prices will be, so if it is dropped low enough, maybe the 3.5% chance wouldn't be too bad. This also shows how valuable a 2/2 item is if you find one, or get a really lucky craft on your Improved to Magic. This doesn't cover the mixing of 1/2 with 2/2, because that would square the size of the table. While that isn't too bad in this case, it gets unwieldy with exos or legendaries. The legendary table could take your entire screen

Extraordinary to Legendary
Here is where it starts getting important to concentrate stats by including multiple good ones in an effort to get legendaries with 2/4 or 3/4 We'll just launch straight into the table. Just like the previous chart, this doesn't cover the mixes of 3/3 with 2/3 or 1/3, but it gives you a good feel of what you are looking at.


Legendary to Legendary
Here is the big one.

Down at the bottom, you see the 0.05% chance (or 1:1820) for crafting a 4/4 out of 4x 1/4, but your odds go up if you work with 2/4s, you actually have chances like 10% for a 3/4 or 3.85% for a 4/4.

More analysis coming, and maybe more in depth charts with duplicates removed by looking at just the total number of good stats going into the bench vs the total good stats out. This would remove things like 2x 4/4 and 4x 2/4 both yielding 10% chance, because in terms of total stats in and out, they are identical. Also, the duplicate error values would be reduced.

 

Okay Baragain, while your working out the math, how do we factor in the random?

I go out and farm (or craft) 4 exo swords.

Each sword has 2 lines of % dmg and 1 line of dmg.

I strut my stuff over to the workbench and throw down my 118 gold.

Sure enough I get a leg with 2 lines of % dmg, 1 line of dmg, and 1 random enchantment (doesn't really matter which).

For whatever reason, I don't like it.

Yes, I can roll the dice until I get that magical 4th line that I want, but their is still random in the mix.

My base (white number of top) can vary.

The % dmg line and dmg lines still fall into the random number generator to at least some degree.

Just because I can get the 4 lines I want, doesn't necessarily make those lines great.

It seems that too many are thinking that is going to be so easy. Yes it should level the playing field quite a bit in the long run as those of us who don't have the luck to have a nice weapon fall out of the sky.

Once we can test and actually see just how random things are we all may change our minds on whether this is a good or bad thing.

 

Never count on random, but be pleasantly surprised if it ever works in your favor. We have no data on how often good stats show up from a random number generator generated stat, but I'd wager that it is rather low. Better that you work on condensing your good stats until such a time that you have a high probability of getting what you want just from the tables above.

You lost me here for a second until I realized that you were talking about a different version of crafting 2.0 than I was envisioning. It is my understanding from how the system has been described thus far that it isn't just the type of enchantment that is inherited, but the stat and value. You seem to be describing that the type is passed on, but not the value.

I am basing my assumption on this statement from the original feedback thread,

The post then goes on to explain how stats are passed on:

Unless you interpret this differently than me, or reality turns out to be different than the announced plan, the entire second half of your post becomes irrelevant.

 

Yeah well, the formula is a bit more complicated if we calculate probabilities for a mix of good and bad enchantments in the output - you need to take combination possibilities (binomial coefficient, "n choose k") into account.
Also, we need to be precise in whether we speak about the probability of getting AT LEAST or EXACTLY n "good" enchantments.
(BTW, I had the same problems in figuring out the correct approach as you did, @_Baragain_ ... )
I made a new spreadsheet for the probability of getting EXACTLY n good enchantments from a given number of good input enchantments.
As you can easily see, the probabilities add up to 100% per column, which ought to prove the calculation correct ...



This does "of course" NOT account for the chance of getting a "desired" enchantment from the additional, random enchantment added by crafting an item of a higher rarity level.

EDIT: My calculations probably include a small error at least for legi-to-legi crafting, see posting #68 in this thread.