Margin of Error calculation

[Taj]

Yeah but how does that solve the inherent evil involved with both women and cats?

It cures inherent evilness, too. It just doesn't say it on the box. It's a hidden effect.

[Saga]

what the fuck maths

[Ryko]

Yeah but how does that solve the inherent evil involved with both women and cats?

It cures inherent evilness, too. It just doesn't say it on the box. It's a hidden effect.

Omgah use mistmelts

[Taj]

Ashira Spike Flails when you pull her tail.

[Ryko]

I like the drama involved.

[Ashira]

oO

[Croaker]

There's really no accurate way to get the drop rate of a monster, considering all the bizarre variables that are factored in.

It depends on how accurate you want it to be.

If you flip a coin once there are a lot of bizarre variables that can be factored in too. How many times did it flip, did you catch it or did it hit the ground, how tall are you, what surface did it land on, ect.

If you flip that coin enough times all of those variables average out and you will find that you get 50% heads and 50% tails.

If a drop is coded to be 25% and you kill it enough times, you will also find that those bizarre variables average out and you get a drop 25% of the time.


If you want to know that drop to 0.00001% certainty then it will take forever, but if all you want to know is a droprate within a margin of 5%, then it doesnt take that long at all. TH effect >>>> bizarre variable effect.

** A t-test takes into accout both variability AND difference between the means.

Considering that there is a large difference between TH4 and TH0 (>5%), it doesnt take very long to determine a lot of things.

What is effect of TH on Rare/EX drops? (OP origional question)
What is the effect of THF kill?
Is TH alliance or party?
Does TH stack?
Does a THF need hate?

Find a freqently poping mob with a decent drop rate, bring a THF with TH4, and you can know that and more within a day.

[Tordall]

I think any serious discussion in this thread has long since gone away, but would like to make a comment that the OP probably needs to pay attention to in order to get any meaningful results.

Moon phase has a big effect on drop rates, if you want to get meaningful results you'll need to do a FFT and remove the 'day of week' and 'moon phase' parts.

For those that don't know FFT = fast fourier transform. It basically breaks down time-dependant data into the magintudes of the different time dependancies. For things with cycles (like the real world, and FFXI), you should use them on time-dependent data in order to remove things like being able to see U-boats at night or during the winter vs. during the day or during the summer.

[epigram]

I would just shoot for a margin of error that makes sense to you. 95% at 5% should be fine for what you need. This should at least be able to give you an idea of the average rate for each TH. Personally a 1% to 5% increase in drop rate would be near-worthless to me. Yay for 5 more ridills for every 100 that dropped!

As far as the day/time/moon thing, either test to see if there is a significant difference before hand, or try to hold constant what you think will affect the results.

FFT, or more corectly DFTs in this situation are hardly necessary. Simplicity would be much more apt, as would stratification.

[Lordwafik]

Besides, mithras are ugly.

This man speaks troof.

[Naver]

Wow that purple dalmatica looks sweet. Taj did you create that? Would be nice if it was detailed to match Adaman Hauberk! :D

[Nystul]

Wow that purple dalmatica looks sweet. Taj did you create that? Would be nice if it was detailed to match Adaman Hauberk! :D

by making it blue?

[Naver]

Wow that purple dalmatica looks sweet. Taj did you create that? Would be nice if it was detailed to match Adaman Hauberk! :D

by making it blue?

yes and no, mostly the gold chain around the neck, and maybe the pattern embroidered on the body to be blue, or vice versa have the embroider purple and the main body blue. with the cuffs slightly white.

no woosh sorry

[Taj]

Reverend Hauberk.

[LinktheDeme]

Wasn't there something implemented a while ago if you kill the same mob in within the same zone for a period of time the chance of getting drop rate decreases

[Icywave]

Drops in this game are definitely programed according to the game clock.

To get down to that 1% margin of error... according wiki graph you will need 4802 kills, yet I am not sure.

I agree you should run those tests under all conditions, which is easier for you in a long rum because you can normalize them at the end at the same time.
However, I also suggest to record the "Days"
of which noticable drops.

Every mob have different tier of treasure pools,
and those tier's drop % changes constantly under the game clock and many other factors.
However, I suspect "Days" have quite heavy influence on mob drops what.
(thank you good old millions of Kirin)