o.O Overkill on the crafting Dumb down?

[syntaxlies]

What exactly do you think these changes will do to make crafting suck less? Or rather what exactly is it about crafting right now that makes you feel that it sucks?

Because with the changes this thread refers to, the only thing that will change are the items you have to make over and over.

The minigame doesn't change and the rate of progression doesn't change... What do you envision these changes are going to do? Because most of us realize the changes brought up in this thread are only going to change the items you grind for skill.

It's really hard not to reach the conclusion that you have no real understanding about what you dislike about crafting and just hear "CHANGE!" and start salivating.

what do you think they'll ruin. I stopped crafting because every time you go to make something there are 100 road blocks in the way. Try to make a rank 20 lth item that requires rank 30 lth parts.you have to buy parts from a higher rank lth crafter to make something you should be able to make alone(example: the use of raptor sinew cords in low lvl recipes).not everyone has 10 mules that they can store a billion mats on. the streamlining is whats gonna make it better for "me", you can argue all you want but that's how i feel and everyone is entitled to their own opinion.

[Amineri]

I would add that this isn't entirely accurate. You only cover one side of the issue. You are describing how +3 mats get down graded into NQ. You do this incredibly well when you discus making ingots directly from ores. But, a problem arises when you talk about moving to the next stage, plates.

When you move from ores to ingots, you mention how most of your +3 mat'l will degrade to NQ ingots. Then you mention how most of your HQ ingots degrade down to NQ plates. I agree with this, but i think that when calculating the odds, you missed half of the equation. A not insignificant number of your NQ ingots will be upgraded from their NQ state to an HQ state. I do not see this anywhere in your calculations. It seems that you remove all of the NQ from the system, only to tack them on as if unsynthed at the very end.

So, the odds are better than your calculations make it seem like they are. Personally, i think the odds are still abysmal. But, if your going to do this calculation, you should not forget to mention both parts of the calculation involved.

You bring up a very good point, and one that I hadn't thought through. One difficulty with calculating this is assessing the probability of generating an HQ result from NQ ingredients.

Assuming a roughly equal level recipe that allows 100 Quality gain during the synth, so starting with NQ ingredients would yield about 100Q. Mooglebox estimates a 0% chance for +2 and 0% chance for +3 with 100 Quality. I don't think that this is entirely accurate, but I do believe that the probability of getting an HQ result using NQ ingredients is nearly negligible, and so would not significantly improve the time needed to create an HQ part.

As a check, I considered the current recipe for making an [item=10300007]Iron Halberd Head[/item]. This is a rank 44 Blacksmith synth, that requires 4 iron ingots to create. Currently, I don't think anyone thinks that there is any substantial chance to get a +2 or +3 result using NQ iron ingots as materials. This is the type of ingredient for the [item=4080301]Iron Halberd[/item] that people leave at HQ or use a +1 result for in the current HQ process.

The current recipe situation had two flaws. First (as people have pointed out), many recipes required parts that were 10 or 20 ranks above the recipe, making the recipe very difficult.

On the other hand, many recipes required parts that were 10 or 20 rank below the recipe. For example, most of the parts used in the [item=7010009]Iron Dolabra[/item] are rank 20 to create, while the final synthesis is rank 43. These low rank parts are much easier to HQ using NQ materials. This leads to the current situation, which devalues the HQ materials collected by gather and battle classes.

It looks like the new recipes will use rank appropriate ingredients. No more ingredients far above the final recipe, and no more ingredients far below. This in turn will result in the de facto requirement that HQ materials are required to make these ingredients.

The exception to this is high rank crafters making finish item that are 10 to 20 ranks below them. In this case, such a crafter would have a reasonable chance to HQ the ingredients, and thus the final synth. Whether these crafters will make these, or focus on the higher rank (and higher value) items, remains to be seen.

[WillRiker]

You bring up a very good point, and one that I hadn't thought through. One difficulty with calculating this is assessing the probability of generating an HQ result from NQ ingredients.

Assuming a roughly equal level recipe that allows 100 Quality gain during the synth, so starting with NQ ingredients would yield about 100Q. Mooglebox estimates a 0% chance for +2 and 0% chance for +3 with 100 Quality. I don't think that this is entirely accurate, but I do believe that the probability of getting an HQ result using NQ ingredients is nearly negligible, and so would not significantly improve the time needed to create an HQ part.

As a check, I considered the current recipe for making an [item]Iron Halberd Head[/item]. This is a rank 44 Blacksmith synth, that requires 4 iron ingots to create. Currently, I don't think anyone thinks that there is any substantial chance to get a +2 or +3 result using NQ iron ingots as materials. This is the type of ingredient for the [item=4080301]Iron Halberd[/item] that people leave at HQ or use a +1 result for in the current HQ process.

The current recipe situation had two flaws. First (as people have pointed out), many recipes required parts that were 10 or 20 ranks above the recipe, making the recipe very difficult.

On the other hand, many recipes required parts that were 10 or 20 rank below the recipe. For example, most of the parts used in the [item=7010009]Iron Dolabra[/item] are rank 20 to create, while the final synthesis is rank 43. These low rank parts are much easier to HQ using NQ materials. This leads to the current situation, which devalues the HQ materials collected by gather and battle classes.

It looks like the new recipes will use rank appropriate ingredients. No more ingredients far above the final recipe, and no more ingredients far below. This in turn will result in the de facto requirement that HQ materials are required to make these ingredients.

The exception to this is high rank crafters making finish item that are 10 to 20 ranks below them. In this case, such a crafter would have a reasonable chance to HQ the ingredients, and thus the final synth. Whether these crafters will make these, or focus on the higher rank (and higher value) items, remains to be seen.

If you look at the data used to develop the calculator on Mooglebox, gathered by Mogg Fanatic on Figaro, you will see that there is an issue with that calculator when it is used on lower ending quality values input. Mogg Fanatic makes note of this on the Japanese version of his site, but i guess the gentleman on Mooglebox did not check this portion of Mogg's blog out.

Ill link it for you. Check out the lower values from the data on the charts.
http://static2.finalfantasyxiv.com/a...LETXvzTJ50.jpg

You will see that the raw data used to develop that calculator breaks down when lower ending qualities are entered into it. If going, purely by the linear regression developed by Mogg Fanatic, you would have a negative chance at HQ1 when you finish below 60 quality.

I am sure we have all HQ1 and 2'd at below 60 ending quality during our leveling grinds. So, we know that the values for low ending quality on that calculator are to be held question. Again, Mogg Fanatic was aware of this issue, but the guy(s) on Mooglebox did not mention it at all. Well, unless you looked at the y intercept in the formula... it would've clued the user in on the issue. Still, Mooglebox didn't outright mention this issue, so I blame them for your confusion. Instead of making it known, the guy who programmed the calculator, on Mooglebox, decided to return a value of 0 whenever his calculations would spit out a negative number.

The thing is that it is obvious that either the chance to HQ is not purely a linear regression, or the chance to HQ is floored at some value larger than 0.

Long story short, any value below 200 or over 350 entered into that calculator is an extrapolation of the raw data gathered. Only the ranges from 200 through 350 were tested. So, we simply do not have the data to support what you are saying in that test.

And yet all of this fails to mention the largest flaw with Mooglebox's program. All of the values Mogg Fanatic used to generate the Linear regression Mooglebox used are PURELY for the testing done on touch ups. The chance to get HQ when progress reaches 100% seems to operate under a different formula. This is also mentioned on Mogg Fanatic's Blog. It is also supported by Mogg Fanatic's data. But, once again, the verbage used on Mooglebox is very very miss-leading for the user. So, all of those calculations you did were purely for the chances to HQ on touch up. And, therefore, are absolutely useless for calculating the success rate for mat'l/parts.

I gotta say, I dont blame you for thinking that the tool you were using was correct. Heck, if I use a ruler, I would assume it is reasonably accurate. And the general logic behind your arguement was reasonably sound. The only reason why I know about this issue is because I am looking into making my own calculator to do the same thing. Of course... i was going to mention this problems in a known issue blurb below the program instead of taking the slacker approach that Mooglebox did.

I gotta say, it would be pretty darned close to an epic fail if this long of a thread happened because of the program on Mooglebox.

[Amineri]

If you look at the data used to develop the calculator on Mooglebox, gathered by Mogg Fanatic on Figaro, you will see that there is an issue with that calculator when it is used on lower ending quality values input. Mogg Fanatic makes note of this on the Japanese version of his site, but i guess the gentleman on Mooglebox did not check this portion of Mogg's blog out.

Ill link it for you. Check out the lower values from the data on the charts.
http://static2.finalfantasyxiv.com/a...LETXvzTJ50.jpg

...
I gotta say, it would be pretty darned close to an epic fail if this long of a thread happened because of the program on Mooglebox.

Oh, I recognized (though didn't put it explicitly into my original post) that the dataset was only valid for the bulk of the statistical distribution, and was less valid in the tails. This is pretty much always true when trying to gather data for the purposes of estimating a distribution curve, because you inherently get less data on the tails of the curve, making the estimate much less reliable.

I also figured that the data was primarily derived for touch-ups (since it is much easier to collect a lot of data from touch-ups), rather then the actual chance of HQing a single ingredient when touch-ups aren't available. Unfortunately, this is the best data set that I've seen, so I chose to use it, rather then trying to "guestimate" a number on my own. This is also why I made sure to reference the source of my estimate!

In part this is how I justified using such a 'back of then envelope' approach to analyzing the new system. There are clearly a fair number of nonlinear transformations, and with a non-linear transformation, the transform of the mean ~= the mean of the transform. I would hope that an SE analyst, with the exact distribution in hand, has a model that allows for a Monte Carlo style testing approach that would allow the dev team to tweak the HQ setting to just the desired value.

The primary goal of my post was to show that SE can easily set parameters to make items still difficult to HQ in the proposed HQ system, and with the revised (simplified) recipes. It all hinges on their Quality-to-HQ conversion formula. They can make it very high, thus making it quite easy (e.g. 50% HQ with 100 quality), or absurdly challenging (e.g. 3% HQ with 400 Quality). They could even make it ridiculously easy by making HQ ingredients be 100% given HQ raw materials.

However, the HQ system and the simplified recipe design itself does not inherently have the flaw that it is unavoidable to make HQing items easy. The design itself is quite elegant, and gives the recipe designers an easy method to tune the difficult to HQ various items.

Having asserted that the system is sound from a game mechanic perspective, I do feel that it has lost something in terms of artistic quality. However, in most cases sacrificing a bit of the artistic aesthetic to improve gameplay is a worthwhile trade.

Finally, thank you a bunch for making clear the weaknesses and assumptions built into my model. I didn't really consider that some people might derive the other conclusion, that SE was going to be making the new system impossibly difficult to HQ items.

[WillRiker]

I still dont see why you are sticking to this 100 quality number. You are making parts, not finished products. this means you arent saving durability for touch ups. It also means that the mooglebox calculator simply cannot be used.

You would need to go back to Mogg Fanatics blog, he has the raw and compiled data there, to look up the HQ% numbers before he tried to touch up. I don't want to make it seem like I am actively offending you, as it is obvious that you have a good head on your shoulders. But, not only are you are using the wrong statistical data in your calculations; you are also forgetting about the other side of the equation.

I am not sure how much more I can help you. I pointed all this out above and on the previous page.

Don't get me wrong, the numbers are still far too low for my liking. And, I know that removing parts will increase the chances for HQ. We are both agreeing on the same general points. But, I just want people to know that there are fundamental flaws in your mathematical model. The current odds are better than what you make them out to be.

If you are looking for better data regarding the finishing quality vs HQ rates for parts i would use this data (Also from Mogg Fanatic):
http://www.bluegartr.com/threads/104911-Test-Resuluts-On-Stats

Youll have to scroll down on the first page, but a bar chart is there. And you will see that the effect of moving from NQ to HQ result is not insignificant... even at lower numbers of ending quality. Just note that this bar chart combines HQ HQ2 and HQ3 into one nebulous category. But this isn't messed up with results only good for trying to touch-up.

[Amineri]

Oh, well ...

>.>

<.<

In part, I'm figuring that the dev team will be adjusting the HQ rate for converting raw materials into ingredients, or ingredients into other ingredients. I suppose it would be a worthwhile exercise to crunch through using that data (and the implicit assumption that +1/2/3 would map to HQ in the 1.20 HQ system), just to see how easy THAT would be.

And, you raise a good point that without the worry of saving durability for touch ups, really dedicated crafter might well push the delta-quality to 150Q. I'm all for trying to stress the proposed system to see how easy we can make it. With all the possible breakage from people busting their HQ gear trying to stuff in five materia, the market could pretty easily handle a larger influx of HQ gear. The 16.5 hours for a single pair of chain boots is kind of steep, but the 2000+ HQ raw mats needed seems like the more critical roadblock. Not to mention that coloring the gear (to customize it / improve the stats) is another 5+ hours and 700 or so HQ mats.

Anyhow, I'll scribble on my envelope some more and post results in a bit.

[BlackRoach]

I say leave the crafting system as it is supposed to be* with the exception that you need to keep it worth it to lvl. Meaning you dont go throwing out the best gear in quests and leves, only maybe a piece of what a crafter would need to make this good item.
You made crafting such a chore it needs great rewards, but I really enjoy myself when there available.

If you want to change it up, add more items in there that can be crafted, and I liked that color thing, be able to use any dye as long as it meets the crafting lvl. Meaning leave the dyes at the lvls they are but be able to use them and any lower lvl dye as well once you hit the mark.

ONE MORE THING!> YOU ALL RUINED THE GUILD LEVES!!!! Why you took them all away and only left us with 2 is beyond me. that was definitly a step back. Crafting is harder than lvling to begin with, and the previous reward system was barely worth it if existant to keep you going. At this rate this will be another game with useless crafters made worse in that you made them impossible to keep motivated to lvl.

[Amineri]

This is a refinement of my original estimation here.

This estimated that to make a pair of HQ Uncolored Cobalt Sabatons would take about 16.5 hours of crafting and 2000+ HQ mats. Following a suggestion by WillRiker (thank you, Number One! ), I'm reworking this with an even MILDER set of assumptions, by using some raw gathering data located here. This data is specifically for materials, and gives the percent chance of achieving any +1/2/3 result. So, this estimation will work off the most relaxed assumption, that SE will make an HQ result in 1.20 equivalent to any form of HQ result in the current system.

Even better, I'm going to assume that for an equal rank synth, a really assiduous crafter could push the quality gained up to 150 quality for HQ mats, or 125 quality for NQ mats (with less starting dura), as no durability has to be saved for touch ups.

NQ ~> 125 quality ~> 10% HQ rate ~> 10 synths per HQ
HQ ~> 450 quality ~> 25% HQ rate ~> 4 synths per HQ (I estimate generously here)

Using all HQ materials

Looking first at the easier ingredients: the boar leather, cobalt rings, and electrum ingot. Assuming HQ materials, only 4 synths are needed per ingredient, or 8 minutes per ingredient.

Next, the plates. Assuming all HQ ores, each ingot synth generates 0.75 NQ ingot results and 0.25 HQ ingot result. Using the easier assumption that each ingot synth yields two ingots, this means 1.5 NQ ingots and 0.5 HQ ingots.

Two NQ ingots yield 0.1 HQ plates, and 0.9 NQ plates (expected value), while 2 HQ ingots yield 0.25 HQ plates. Normalizing per ingot, each NQ ingot yield 0.05 HQ plates, and each HQ ingot yields 0.125 HQ plates.

Putting these two together, each set of four HQ ores yields 1.5*0.05 + 0.5*0.125 = 0.075 + 0.0625 = 0.1375 HQ cobalt plates. Roughly half comes from the NQ path, and half from the HQ path. To get 3 plates then requires 3/0.1375 = 21.8 of the original synths making ingots from HQ ores. Rounding this to 22 ore ~> ingot synth nets an estimated time of 44 minutes synthing ingots, and another 44 minutes spent synthing plates, with 44 HQ specular iron ore, and 44 HQ iron ore used.

Totals:
112 minutes of synthing to yield all HQ ingredients
50 HQ specular iron ore, 50 HQ iron ore, 16 HQ electrum ore, 4 HQ boar hide, 4 HQ alumen

-------------------------------------------------------------------------

On a side topic, I'm interested in how valuable the HQ raw materials are in this scenario. For comparison, I'm going to go through the same calculation, but assume that all NQ materials are used instead.

NQ ~> 125 quality ~> 10% HQ rate ~> 10 synths per HQ

Using all NQ materials

Looking first at the easier ingredients: the boar leather, cobalt rings, and electrum ingot. Assuming NQ materials, 10 synths are needed per ingredient, or 20 minutes per ingredient.

Next, the plates. Assuming all NQ ores, each ingot synth generates 0.9 NQ ingot results and 0.1 HQ ingot result. Using the easier assumption that each ingot synth yields two ingots, this means 1.8 NQ ingots and 0.2 HQ ingots.

As before, two NQ ingots yield 0.1 HQ plates, and 0.9 NQ plates (expected value), while 2 HQ ingots yield 0.25 HQ plates. Normalizing per ingot, each NQ ingot yield 0.05 HQ plates, and each HQ ingot yields 0.125 HQ plates.

Putting these two together, each set of four NQ ores yields 1.8*0.05 + 0.2*0.125 = 0.09 + 0.025 = 0.115 HQ cobalt plates. To get 3 plates then requires 3/0.115 = 26.1 of the original synths making ingots from HQ ores. Rounding this to ore ~> ingot 26 synth nets an estimated time of 52 minutes synthing ingots, and 52 minutes synthing plates, with 52 NQ specular iron ore, and 52 NQ iron ore used.


Totals:
164 minutes
72 NQ specular iron ore, 72 NQ iron ore, 40 NQ electrum ore, 10 NQ boar hide, 10 NQ alumen

--------------------------------------------------------------------

Comparing the HQ and NQ materials

HQ mats: 112 minutes, NQ mats: 164 minutes

The NQ synth take about 50% longer, so the time factor means that HQ materials would command a 50% premium over NQ material,.

HQ mats: 50 HQ specular iron ore, NQ mats: 72 NQ specular iron ore
The NQ synth requires 44% more ores, so the material factor means that HQ materials would command a 44% premium over NQ materials.

Putting these together, HQ specular iron ore would sell for approximately 116% more then NQ specular iron ore.

-------------------------------------------------------------------------

Conclusion

• Using the easiest of all assumptions, and using all HQ materials, a pair of HQ uncolored sabatons would take approximately 2 hours to complete.
• Using all NQ materials, the same item would take approximately 3 hours to complete.
• Market value for HQ materials would be a bit over twice that of NQ materials.

[Waaltar]

This is madness.

They are seriously junking the complex crafting system to save HD space in save files??? They need to get rid of nuggets to save HD space?????????

What the heck are the save files made of? Maybe .wav audio files 'encrypted' by Navajo code talkers?

It doesn't make any sense. Save files are server side so they don't need any encryption, all they need are backups. They should be small .txt files with a string of numbers and letters. The entire game population should fit on a single ~$100 terabyte hard drive with plenty of room to spare. Actually, it should only be a few gigabytes of data, smaller than a comparable number of pictures.

Even if in programming terms this is the unfixable result of horrible programming/design, there is a fix by another route.

For the love of moogles and masamunes SE, don't be stupid, let people throw money at you.

Let people buy more storage.
Let people buy cosmetic items.
Let guilds buy a guild vault and guild bank etc.
Let people buy early access to features that will eventually be put into the game as part of the basic package for everyone to help fund the game development.

Think in terms of one time fees rather than monthly fees.

You can see I'm not talking about the dreaded pay to play or pay to win stuff. I'm just saying if SE is under such extreme cost pressure than do not solve it by cutting stuff out of the game, solve it by building up the game and do that by letting people throw a $5er at you here and there.

And if SE isn't going to do anything about Gil sellers, and needs money to help fund the development of the game, then SE should cut them out and sell gil. Or take a middle path and mostly cut them out and sell a limited amount of gil to players on a one time basis at a low rate that gil sellers can't compete with.

Whittling down the game because it is free and has a low player base is not a path that leads to it becoming a big beautiful complex successful game. It is such a fail strategy that they might as well just give up now if they won't try something else.

[Arcell]

This is madness.

They are seriously junking the complex crafting system to save HD space in save files??? They need to get rid of nuggets to save HD space?????????

Where did they say that? They said they can't give us unlimited/much greater storage space due to HD restrictions.