Need Tips for Performing a Test on Direction Myths

[urat]

So I've decided after the interesting info my fishing parses have found, I'm going to do a similar test on testing once and for all the whole directions myth.

As a woodworker large sample synthing is very easy, I'll probably start by doing a test on a very high T synth, most likely stone or wooden arrows, which I can pump out easily 100-200 synths a single day in sandy. If anything my biggest block will be wind crystals, lol.

Anywho, I'm posting to ask you guys for some tips on my methods. I'll most likely start punching all this info into a spreadsheet for now (Btw... I need a good free spreadsheet software, anyone got one?)

What would you guys say is a large enough sample size to be able to pretty well state whether or not the direction synthing is true or not?

Here's my method, similiar to what I did in fishing via testing the effect of fishing apron. As fishing is for a fact affected by moon phase, as per the ye ole brady guide, I couldnt simply catch 100 fish with it on, then 1000 fish with it off, as moon phases would seriously fuck up the numbers.

Instead, I alternated on and off apron every 20-30 fish, swapping my parse back and forth. I'll do the same for synthing.

This way, even as I skill up my woodworking thes elast 2 levels, even if I do, the bonuses will cancel each other out and we may even see a slight increase in both chart's HQ rates from the 1~2 skill I might get.

Wooden Arrow: Woodworking skill 4

Wind crystal

Arrowood lumber

Chocobo Fletching

Stone Arrowhead

Directions: ( Days and Directions of Synthing - FFXIclopedia, the Final Fantasy XI wiki - Characters, items, jobs, and more )

Southeast: Better success

South: Greater success

East: HQ rate, lower success

Others:

Earthsday: Higher success rate

Iceday: Higher skillups/HQ rate?

Moon: I'm not even sure, what is the general thought on moon phases on success vs hq? Full = success new = HQ? Or other way around?

Anywho, I'll log every days current moonphase / day combo when synthing, so for example if its iceday new moon and the graph shows a dramatic HQ peak, we can assume then that there may be more merit to the moon phase synthesis then we think?

If that makes sense? Thoughts? Sample size too small? A better synthesis?

I suppose since I'll most likely be synthing my own mats for the arrows I'll log those synths as well to take into account T0s and T1s

[cdgreg]

I really think this guy said it best

http://www.bluegartr.com/forum/81290...-2009-a-3.html

You are never EVER going to be able to prove conclusively anything with regards to resist/day/direction etc. You have next to zero initial information about what affects HQ chance. Therefore you must first ASSUME that it is exclusively modified by a few choice variables (who's to say you'll guess right). You then have to collect the data for each variable in order to determine its effects while either keeping all other variables constant or taking a huge sample in order to reach an appropriate confidence interval.

Since keeping all variables but 1 constant (some or none of which may ACTUALLY be variables) is quite difficult (especially where day/weather/time is concerned) you're probably just gonna need a huge sample size. So, in order to meet this huge sample size you're going to have to get other players' data since doing it all on your own is no fun.

However, then you have to contend with these people that go 5/20 on dusk gloves or 8/100 on sheikh manteels. These are statistical anomalies but they're still statistics so discounting them would skew your results. But what if one of them is lying about his results? How do you know if he is or not? He could just make up his HQs and the conditions to screw with you. In short, can you even trust the data other people give you? You can probably trust most, but on something like a T0 those HQ rates have a very real potential to skew results.

Oh, as an extra fuck you: what about the vanadiel month/year? maybe days behave differently year by year

To put it simply there's potential for so many different variables that you might as well just treat it like a random number generation. If you want to make money crafting, always craft things that in the long run will make money (i.e. not haubergeon, scorpion harness etc.). If you live by the law of averages you can't lose.

You can go on from there to say maybe being from a starter city that has that guild may help your overall HQ rate, elemental armor, etc.

[Weeks]

There are not that many reasonable variables, and doing this with cheap consumable synths would enable you to get the thousands of necessary data points to filter their noise. We crafters are just lazier than the people who figured out how every other aspect of this game functions, so we lie to ourselves and pretend it's an insurmountable task.

[nitsuj]

I will say that while working in the game industry, I've noticed some interesting things. I work closely with gameplay algorithms in regards to RPG stats and once I worked on a game that shall remain nameless on a certain RPG related algorithm.

On the top, much like FFXI crafting, there was a very simple input/output to the player. You do X, Y happens, sometimes Z happens.

The algorithm, though, to the programmer showed so many variables being plugged in that it took me 3 entire days to decypher *exactly* what was going on enough to determine what was happening to call myself an "expert" on the system and report back to the designers.

Some interesting things were learned, like that no matter how simple the input/output to the user is, we have no idea if it's "DO X, flip a coin and heads Y happens, or Z happens" or if it's "Do X, take into account A, B, C, D, E, and F, add in and multiple 10-12 game settings that the designers can tweak a little bit till the system feels right to them and then do some more algorithmic tweaks to get the final outcome. It's really crazy.

I also learned, in this particular equation, that after all the crazy factors that went into this gameplay system, all the multiplicative and additive multipliers the designers used to finely tune the system... the system (which was designed to determine a result on a scale between 0.0 and 1.0) was CLAMPED between 0.0 and 1.0 and then applied some MORE math/gameplay elements into the equation.

So even if the moon phase, direction or whatever applies into the equation (which personally, I think some of it does) the number could be affected by so many things that are *actually* impossible for us to figure out.

tl;dr version:

No matter how many tests we do, and what results we find there are two issues. 1) The *is* a random number generator involved. And 2) It's impossible for us to take into account clamps/gamesetting "magic numbers" and other things because most likely it's not a simple "X + Y * Z etc etc". No matter what, our results are just an educated guess and will come to down to what your personal experience and HQ results have shown.

I know it's a lot of random shit, but I've been thinking about it a lot this week and this post kinda made it all come out at once

[urat]

This is what I'm trying to take into effect, in other words, if I dont see at least a 10~20% conclusive increase in HQ rate by facing HQ direction through the whole graph, not just spikes here and there, I think we can state that if the HQ rate does improve, their are too many other factors to notice, or the HQ rate increase is infinitesimal and not worthwhile.

The main reason I want to do this is #1 I'm honestly curious, and #2 I think if there is a HQ rate it will be extremely noticeable (I'm hoping direction is increasing the rater by 30 40%) so if I don't notice any change then I will assume there is no connection between direction and HQ rate / success.

Who knows, I might see a jump in 3x the break rate on synths and 50% HQ rate when I face proper direction.

On another note, I might also need to face a 4th "non effective" direction for further comparison, which means Im looking at about a sample size of 4000....

[Mansin]

Here's a good tip: Don't waste your time

[cdgreg]

If you were to find anything, I doubt it'd be anywhere near the type of increases you are hoping for. If they were that drastic, or even a 5% increase, it would have been discovered and proven a long long time ago.

I stink at math and may be interpreting your writing incorrectly, so if I am nm, but I'm reading it as you're hoping to see the standard 1/16ish T1 jump to something like 20-50% HQ rate because of direction.

[urat]

If you were to find anything, I doubt it'd be anywhere near the type of increases you are hoping for. If they were that drastic, or even a 5% increase, it would have been discovered and proven a long long time ago.

I stink at math and may be interpreting your writing incorrectly, so if I am nm, but I'm reading it as you're hoping to see the standard 1/16ish T1 jump to something like 20-50% HQ rate because of direction.

If it exists, and thats on a T...4? 5?

If it doesn't exist it wont be a noticeable increase at all, this is what Im hoping.

[tsaroth816]

If you were to find anything, I doubt it'd be anywhere near the type of increases you are hoping for. If they were that drastic, or even a 5% increase, it would have been discovered and proven a long long time ago.

I stink at math and may be interpreting your writing incorrectly, so if I am nm, but I'm reading it as you're hoping to see the standard 1/16ish T1 jump to something like 20-50% HQ rate because of direction.

How many years did it take us to figure out the DEX/crit formula?

[cdgreg]

How many years did it take us to figure out the DEX/crit formula?

Maybe you didn't read all the replies here. Urat said if he doesn't see a huge increase in hq rate like 10-30% he said he's going to consider the theory debunked. My point was if there is an increase it's going to be very small. if it were something ridiculously large like boosting yor hq rate 20% facing the "right" direction it would had been discovered a long long time ago.

I don't know much of anything about the dex crit rate % which is why I'm not posting in a thread related to that.

[Massimo]

If it exists, and thats on a T...4? 5?

If it doesn't exist it wont be a noticeable increase at all, this is what Im hoping.

lol

After all his water resistance for fishing bullshit, why do you even attempt to still take this guy seriously?

[Mistress Stowastiq]

To OP,
There is a free software called Open Office, openoffice.org.

Sample Size: a minimum of 1000 reps.

Wind crystals from starting nations with low level mobs who drop those. Especially on weekends.

Also Norvallen for the npc vendor of arrowwood log in town (and guild when you catch the low price on lumber), as well as Tavnazian Safehold standard merchant.
Also lots of Arrowwood lumber drops from logging Ronfaure and Gheselba Outpost and chocobo digging if you happen to be doing it anyway.

Pad yourself with armour for future Sath-ings to come from this thread. :D

While people are telling him not to waste his time, surely he took/takes time into consideration before attempting this so advice goes both ways.

Good luck and I am interested in seeing your results.

[BerenTebogo]

I had a long diatribe typed up. The guy is applying science and math to an experiment not shouting theories from a rock with his head up his ass.

Whether he follows through or not it's farther than 99% of the grotesque excuses for human beings that frequent this forum get on average. I assault stupidity, not progress and experimentation.

If it needs to be said again then I'll say it again, coming here seeking some retarded sense of appreciation for offering up what you consider to be a BRAND NEW THEORY will land you with my fist up your ass. I don't want to hear it, nobody else wants to hear it, go the fuck away. You can test on your own, and honestly, anybody who has passed a high school science class should KNOW HOW TO CONDUCT AN EXPERIMENT without or help but for the sake of not ripping on anyone I've let this guy get advice.

Pick one variable, test. Test said variable in all different surroundings. Pick second variable, test second variable same way. Compare results, adjust tests accordingly and retest.

So hard, so hard.

That being said, I wish the guy luck if only to shut the mindless masses up. There have been literally over 1,000,000,000 synths completed in all of FFXI. It's reasonable to assume that at the very, very least, 1,000,000 Tier 1 synths have been completed. If a pattern exists it will rise up out of the chaos no matter what, it is how the world works. Through communication and shared experience we've been able to examine a huge number of these synths on an even bigger scale so that we, as a community, are given the opportunity to see if a pattern emerges.

Easily 50% of the people who post their HQs here follow every day/direction/fingerupass theory that exists, I know this for a fact. The rest of us synth facing the auction house the moment we buy materials. Somehow, strangely enough, everything seems to come out even on the large scale. Isn't that just the darndest thing?

I support the effort, but the ignorance spewed by some people is stifling. We, as humans, were able to discern evolutionary patterns in animals before we could light a fucking lightbulb. You really, REALLY, think that we can't discern a pattern in a fucking video game on a scale that would make darwin shit his pants?

Really people, honestly.

[Zouri]

Even with a 1 billion sample base, you would never be able to prove anything. It could be as simple as "we allow this many hqs per (#) in tier (#) over all servers." Since you do not use every single crafting character across all servers, its not possible to prove even the simplest of possibilities. I think it took sath 6 paragraphs to say that.

Test away, but from as much as i synthed when i played... its unpredictable. Nothing i ever tried made a difference. ... ... well, maybe synthing through every single material until i got a hq helped some, but that was more stubborness than anything.

sath, hurry up and post some cerb. +1s or dusk gloves +1s.. You have the power! ^.~ (at stage 4 yet?)

[BerenTebogo]

@1700, supply at non-inflated price is dumb and I refuse to synth til I have stage 4 in my hands.