This topic contains 34 replies, has 9 voices, and was last updated by 
Gray 5 years, 6 months ago.
-
Posts
-
Then yes – I was using Makers Mark too! Otherwise IMHO you have to be mad to use 62-key (!!!) rotations just because you can’t optimize a short and nice Muscle Memory rotation…
62 key or 42 key rotation doesnt make a lot of difference. Youre crafting with the bot anyway. If you want micromanagement then yes macros are the way to go. But this means youre still enjoying the crafting system while to me and many players it has become a grind months ago. So i prefer to click start and come back in a few hours to a full Rakshasa HQ set in my inventory. I dont babysit it and i dont care if it takes a few minutes longer than necessary.
So once again, that’s not me “chopping wings”, but rather someone else uses heavy cannon where a simple rocket is more than enough. The tricky question is, though: how the hell can one teach Miqo to use this rocket…
Miqo team already explained several times that number of steps is not part of the algorithm because otherwise it would blast solving time x100. Quality bar is also removed from the equation and assumed infinite otherwise it would have to solve x10000. What youre left with is indeed a heavy cannon. Its just simple math. Optimizing for a moderate solution would probably require another year of development. And just recently you asked Miqo devs to focus on minor QoL improvements before starting another big system. So what will it be?
So you are going to say that you pre-set some kind of start sequence, then let Miqo calculate the rest, which means you can actually use Innovation AND synthesis in the end, without waiting for 22 mins?
No i simply follow the guide from Stormblood Stage 2 thread. I dont use innovation and keep my solving time as low as possible – 2 minutes for rakshasa tier and 20 seconds for everything else. Innovation is only +4.4% quality and costs 15 minutes. Makers Mark and Patient Touch are +26.6% and +41.5% and come at zero time cost.
62 key or 42 key rotation doesnt make a lot of difference. Youre crafting with the bot anyway. If you want micromanagement then yes macros are the way to go. But this means youre still enjoying the crafting system while to me and many players it has become a grind months ago. So i prefer to click start and come back in a few hours to a full Rakshasa HQ set in my inventory. I dont babysit it and i dont care if it takes a few minutes longer than necessary.
Then there is the difference between you and me: I care about every minute.
Though, this was not the main reason why I questioned Crafter Solver’s performance. Yoshida just said about “extremely difficult recipes” and so on (see startpost). This made me wonder: if Miqo is 20% less effective at crafting simple 70*** mats than script, then it may just fail you once ShB is released. This is why I didn’t quite understand Arc’s joy…
Miqo team already explained several times that number of steps is not part of the algorithm because otherwise it would blast solving time x100. Quality bar is also removed from the equation and assumed infinite otherwise it would have to solve x10000. What youre left with is indeed a heavy cannon. Its just simple math. Optimizing for a moderate solution would probably require another year of development.
No, it won’t. Everyone who has ever been working on optimisation of something as complicated knows: you just simply can’t check all possible combinations, so you have to concentrate on the most promising ones, and chop away the rest. This is why I say: use pre-sets. For example, I’ve got 4 simple endgame pre-sets which I combine in all possible ways, receiving next to best results. These are:
1) Ingenuity, Steady Hand II, Prudent Touch x4, Piece by Piece
2) Steady Hand II, Ingenuity, Prudent Touch x4, Careful Synthesis II (/III)
3) Steady Hand II, Ingenuity, Innovation, Prudent Touch, Great Strides, Byregot’s Blessing, Careful Synthesis II (/III)
4) Maker’s Mark, Steady Hand, Piece by Piece x2, lots of Flawless Synthesis combined with Comfort ZoneThis way every skill receives most useful buffs, and no buffs are wasted. The latter one, as I’ve said already, is the last resort (Onishi & Rakshasa only), while materials have to use Muscle Memory, to save time. What Miqo has to do is a simple math which shouldn’t take lots of time: insert 2 Manipulation II and 1-2 Comfort Zone somewhere. Yes, it might be still not optimal. For example, maybe Innovation could be placed somewhere else. But I’d say: NFC. It works, it’s fast, and nothing else matters (C).
And just recently you asked Miqo devs to focus on minor QoL improvements before starting another big system. So what will it be?
This. What I always said is: don’t go for complicated math in attempt to find absolutely best result for absolutely any purpose. Instead, use simple solutions, and then work on things which take minimal time, and bring lots of effect. Geez, I’d praise a simple .wav (replaceable) played when Miqo completed crafts or finished grinding!!!
No i simply follow the guide from Stormblood Stage 2 thread. I dont use innovation and keep my solving time as low as possible – 2 minutes for rakshasa tier and 20 seconds for everything else. Innovation is only +4.4% quality and costs 15 minutes. Makers Mark and Patient Touch are +26.6% and +41.5% and come at zero time cost.
Correction: they come at 2 mins cost per craft of each material used in every Rakshasa. And average Rakshasa requires what… 6-8 materials each? Then you waste 7*13*2 = 3 hours every time per each full set crafted…
Then there is the difference between you and me: I care about every minute.
Yes and nobody else cares. You don’t seem to understand, over and over again, that this product is not aimed at you or your expectations. Its a tool for people who enjoy the game, to do grindy stuff while they do other stuff like watch shows or work out.
>The path of minimal effort beats the fastest for me.
>As already said by several people: It is not fast, it is easy and it is done automatically with zero effort.
And it does that perfectly. I crafted alot of endgame HQs without having to think. I finished the Moogle Event. I got all my crafters/gatherers up, I leveled alot of jobs to 70.
Was it always the most efficient and fastest? No… ofc not. Did I have to do anything myself? Absolutely not and thats the point. I could have collected the moogle tomes using the combat assist instead of running squad dungeons or went for the BLU tech. But I much rather do it over night, having a GM alarm setup and have to do nothing at all.
Correction: they come at 2 mins cost per craft of each material used in every Rakshasa.
2 mins for rakshasa only and 20 seconds for each material. And my scenario generator groups materials together so it solves only once per stack of crafts. Are you sure youre not baiting me on purpose?
don’t go for complicated math in attempt to find absolutely best result for absolutely any purpose. Instead, use simple solutions, and then work on things which take minimal time, and bring lots of effect.
This is exactly what devs did. They took an algorithm they created two years ago and added one formula for 380 recipes. Minimal time, lots of effect.
Everyone who has ever been working on optimisation of something as complicated knows: you just simply can’t check all possible combinations, so you have to concentrate on the most promising ones, and chop away the rest.
Which means developing a new algorithm that will check combinations to start with. At the moment crafting solver uses hyperdimensional solution maps.
if Miqo is 20% less effective at crafting simple 70*** mats than script
If you use correct skills its 60% more effective. But like i said this holy war is unnecessary. You have found your preferred crafting method and ive found mine. Miqo supports both. Everyones happy.
Then there is the difference between you and me: I care about every minute.
Nice. I prefer to spend my minutes on something more fun than micromanagement. But im glad youre enjoying it.
Then there is the difference between you and me: I care about every minute.
Was it always the most efficient and fastest? No… ofc not. Did I have to do anything myself? Absolutely not and thats the point. I could have collected the moogle tomes using the combat assist instead of running squad dungeons or went for the BLU tech. But I much rather do it over night, having a GM alarm setup and have to do nothing at all.
How can i set up an GM alert?
Yes and nobody else cares.
I hope Yoshida will make you care once ShB is released, with new crafts which are impossible to HQ unless you’ve got 100% correct sequences (plus some luck with conditions)…
2 mins for rakshasa only and 20 seconds for each material. And my scenario generator groups materials together so it solves only once per stack of crafts. Are you sure youre not baiting me on purpose?
Duh! First you said “You removed Makers Mark … Its a user error.”, then you say “20 secs for materials”? Well, “20 secs for materials” obviously mean you are not using Makers Mark while crafting materials (because it’s followed by 2 mins of Flawless Synthesis spam). So why do you want ME to use Makers Mark?? “Are you sure youre not baiting me on purpose?” (C)
This is exactly what devs did. They took an algorithm they created two years ago and added one formula for 380 recipes. Minimal time, lots of effect.
Yea, I just wonder how much time did it took to implement that algorithm back then. Some months or so, much like with combat assist now? And it could have been done way faster with presets…
Which means developing a new algorithm that will check combinations to start with. At the moment crafting solver uses hyperdimensional solution maps.
Is it that hard to add new presets to solver in the beginning of new content? Took me just some 2-3 hours once I made it into HW from AAR… and there was even nothing new at the start of SB.
If you use correct skills its 60% more effective.
More like 25% more effective, but I get your point. Though, it’s still 20% less effective than maximum possible – that’s the problem. Or rather, it MIGHT be a problem in ShB. We will see.
-
This reply was modified 5 years, 6 months ago by Gray.
/facepalm
Is it that hard to add new presets to solver in the beginning of new content?
Switch Miqo to macro crafting. And use your presets. Whats the problem?
And it could have been done way faster with presets…
Presets were implemented first, crafting solver second.
Yea, I just wonder how much time did it took to implement that algorithm back then. Some months or so, much like with combat assist now?
Ask Miqo team. I believe they have more pressing matters atm but if you think your opinion requires an urgent answer you can contact them: https://miqobot.com/#support
Anyway im getting a little tired of this meaningless conversation. Let me know how it goes.Though, this was not the main reason why I questioned Crafter Solver’s performance. Yoshida just said about “extremely difficult recipes” and so on (see startpost). This made me wonder: if Miqo is 20% less effective at crafting simple 70*** mats than script, then it may just fail you once ShB is released.
And this is what I am denying. You are chopping off Miqo’s full potential by not using the recommended cross-class skills and then asking why it doesn’t work as perfect as I am claiming. Miqo themself said, which are the optimal croff skills:
1. Ingenuity I or II: 38.9% to average HQ rate 2. Byregot's Blessing: 33.3% 3. Maker's Mark: 26.6% 4. Flawless Synthesis: - (required for Maker's Mark) 5. Comfort Zone: 20.2% 6. Piece by Piece: 13.0% 7. Tricks of the Trade: 9.8% 8. Steady Hand II: 6.5% 9. Innovation: 4.4% 10. Reclaim: - Hasty Touch: 2.0% Careful Synthesis II: 1.2% Rapid Synthesis: 1.3% Rumination: 0.0014% Waste Not I: 0.00072% Waste Not II: 0% Manipulation I: 0% Careful Synthesis I: 0% Muscle Memory: 0% (3.7% when no Maker's Mark)If you select these Top10 skills Miqobot is in fact a ship-mounted railgun, that can fire once every 2-8 minutes. If you do not select these skills, you shouldn’t wonder, why you’re not getting perfect results. With these skills and melds even a 100% human crafter would consider “decent” (which is budget pentamelds – for reference, these are the “broke beggar melds” according to the Balance, while these are considered midcore budget stuff and these are absolutely rich Best in Slot melds) Miqobot will never fail your crafts, no matter how hard the recipe is.
Also something you should keep in mind is, that Miqo works mostly off CP. So the more CP you have, the more effective she is, even if your other stats are rather crappy.
I hope Yoshida will make you care once ShB is released, with new crafts which are impossible to HQ unless you’ve got 100% correct sequences (plus some luck with conditions)…
Well, then you will care, because your scripts won’t work anymore since macros do not take different quality states into the equasion and just spam a preset rotation, while Miqobot with the skills mentioned will be able to do the 100% correct sequence, since it works like a human on steroids (who doesn’t care about solving time, but only about solving reliability).
Switch Miqo to macro crafting. And use your presets. Whats the problem?
Hmmmmmmm… gladly! but how? FWIK, macro can be added in the beginning only, while my whole sequence consists of presets. Of course, I can type the whole my sequence into Miqo, but then again how is it going to be different from script (other than occupying Miqo licence)?
Miqo themself said, which are the optimal croff skills:
I’m going to discuss this matter too, just a bit later (have to set up perfect grinder first, so I have more free time on my hands). To be short: if Miqo themselves say something, it does not mean it’s 100% correct. Well it MIGHT be correct, while I may be mistaken… everything is possible, that’s why it needs discussion.
Miqobot with the skills mentioned will be able to do the 100% correct sequence, since it works like a human on steroids (who doesn’t care about solving time, but only about solving reliability).
Plz… I know you are fond of Miqo (and I am too, just I praise its gathering rather than crafting), but still… gimme a break, skip such a big words. Miqo fails to find 100% correct sequence right now, even for a lvl20 crafter, let alone once you come as far as lvl70*** crafting. You don’t notice it because of some combination of three simple reasons: you fail to grasp the fact that time costs money (which, for example, can be put into gathering and thus bring gil), 2) maybe you’ve got crapload of gil invested into melding, and 3) current recipes are just too easy. Imagine crafts which require 20% more quality than Rakshasa, and you get my point…
you fail to grasp the fact that time costs money (which, for example, can be put into gathering and thus bring gil)
Nobody cares, not the point of miqo – invalid
2) maybe you’ve got crapload of gil invested into melding
Endgame gear needs a certain amount of melds, and for everything else if you have appropiate equipment for your level nothing fails.
3) current recipes are just too easy. Imagine crafts which require 20% more quality than Rakshasa, and you get my point…
I fail to see the point. Miqo literally aims to get Quality as high as possible, miqo does optimize the rotation to get to 100%, miqo optimizes the rotation to get as high as possible.
if Miqo themselves say something, it does not mean it’s 100% correct
It is accurate since they used the bot themselfs to log quality improvements per ability. Also all the pentameldede mastercrafters I talk to suggest roughly the same cross class abilities.
Endgame gear needs a certain amount of melds, and for everything else if you have appropiate equipment for your level nothing fails.
No, it does NOT. I craft Rakshasa out of NQ base materials having about 99.8% success rate, while having just a single overmeld on each piece of eq (and weapons w/o overmelds at all). Boo.
I fail to see the point. Miqo literally aims to get Quality as high as possible, miqo does optimize the rotation to get to 100%, miqo optimizes the rotation to get as high as possible.
No, it does NOT. Once again for the most perceptive ones: re-attaching 3 examples of the same craft with the same stats – the best possible sequence, sequence with “progress first and no Innovation” and sequence used by Miqo. As you can see, the second one is 9% weaker than first one, while the latter is 24% (!!!!!) weaker than first one, and 16% weaker than second one.
Once again, we would like to kindly remind everyone to keep the conversation friendly and constructive.
We are fully aware that Miqobot is not perfect. And we also know that it never will be.
We are working non-stop to release new features as fast as possible, however the rate at which we receive feature requests far exceeds our capability of implementing them. Therefore there will always be more work to do.We always welcome constructive criticism and new ideas for improvement.
But please understand that emotional discussions do not speed up the development process.
They only slow us down because we have to spend more time on community management.Thank you for your understanding very much!
We are fully aware that Miqobot is not perfect. And we also know that it never will be.
We are working non-stop to release new features as fast as possible, however the rate at which we receive feature requests far exceeds our capability of implementing them. Therefore there will always be more work to do.Sorry if I’ve offended you! it wasn’t my intent. My point was stated in my first message here: I just wanted to say that it’s a bit wrong time to cheer about incoming crafts, as we don’t know yet how hard are they / if Miqo will be able to fail-safe HQ them or no.
Now that you’ve answered personally, I just have to add (after all my bitching about crafter solver): keep up the good work! Looks like you are really working on the thing most people want (even though I’m not among them); I just hope you start implementing “quality of life” minor additions once your work on combat assist is finished. Oh, and while are are at it: now that I understood beacon system, I have to say that gatherer is just GREAT! Add vendor support on top of it (to buy things in exchange for scrips), and that would be everything I need (well, plus minor stuff like tell notifications)!
My point was stated in my first message here: I just wanted to say that it’s a bit wrong time to cheer about incoming crafts, as we don’t know yet how hard are they / if Miqo will be able to fail-safe HQ them or no.
And my point was, that as a genuine crafter who first aims for a really strong crafter gearset (i.e. pentamelds everything at least to mid-tier – I for instance do it to BiS tier, which is why it takes me really long) and then starts with the whole crafting process automated via Miqobot is going to be pretty happy about hard incoming crafts.
And just fyi, as you said that a few times: I didn’t spend millions and millions of gil on materia for overmelding. Actually I didn’t spend a single gil in general. I simply just did my weekly custom deliveries, did some basic scrip farming and bought all the materia I’ve pentamelded my gear with from the scrip dealer NPCs over a course of about 1-2 month (for almost complete BiS melds that is!).
-
This reply was modified 5 years, 6 months ago by
You must be logged in to reply to this topic.