TH testing: final results

[NynJa]

SE also stated theres no latency issues towards NA players and they're on a level playing field to JP...

Everything SE says =

[Masamune]

3. Do -ga killshots in any way effect droprate?
There isn't going to be a straight yes or no to these, I want opinions because I want my test to have meaning to the people who matter.

1st thank you Thorny for "railing" this thread to the way of "interest" regarding testing TH.

2ndly i may have an anecdote (sadly no parses...) about your question quoted above:

I farm often (as THF of course) @ Davoi for Gold Orcmask. I noticed/felt that pulling carefully the Troopers in such way to not link any other orc around somehow made droprate better than if i killed mindlessly even if i had 5 links...
Same if i get a link or 2 during pull, i would kill links 1st then turn to wait last link body disappear (Trooper unclaimed) to finally engage and kill the trooper.
Again no parses, might be placebo feelings :/

But i brought this anecdote in regards if you -ga a train of mobs, i think droprates might be affected ... but i'm sure of nothing.

P.S.: one thing i would really love to do is doing some Salvage runS with a THFburn setup, like good ol' times @ Lufaise on Abraxas XD.

P.S.2: regarding Thorny 1st question about which mob to test on, having at least 2 drops (one low droprate and other decent or high %), maybe Euvhis @ Sea would be nice test subjects ? They have the rare normal organ, the bit less rare HQ organ, and the common cluster(s).

[Enedin]

Missed a lot of posts here, but this completely blew me away:

1) Thorny posts over 9000 killed colibri with TH2
2) Byrthnoth uses his shiny formula and comes up with 25% and 19% for TH0
3) Thorny posts his 2900ish killed colibri with TH0
4) It's pretty much SPOT ON!

This is pretty amazing, does this apply for most of my results as well? Since he has a huge sample size, this is very accurate, mine will be off at times.

I'm kinda having the feeling we cracked the code to TH potency.

So we can discount the static increase option, as you said. My theory (of a direct drop rate multiplier) can also be discounted, now I'm seeing it on my chart: all the drops that are said to have a higher TH2 'multiplier' have a lower base TH0 drop rate. Check the chart: the higher the TH0 drop rate, the lower the TH2 increase. It can't be a coincidence I only see this now >.>

Man, this is unbelievably exciting. Really. Well, at least for me, since I've been testing potency all along XD

2. Do I need to account for day or moon phase? This type of testing would require active play, so it's not feasible to get a sample size of 10,000 and negate it through law of averages. Does the community believe day or moon phase have an effect or not?

In the original post I killed another 150 Stirges during full moon and that changed pretty much nothing, so I'm confident you can scratch that.

I've thought about the day having an influence (mobs drop crystals, so they have an 'element') but never found anything conclusive. Also a bitch to test.

I'm assuming it has no visible influence, except maybe on mobs like Puks whose behaviour actually changes during windsday. Also Ladybugs night/day, started a test on that night/day effect, but didn't get it done. 50 killed mobs so far give night a huge boost in drop rates.

[Byrthnoth]

Well, as I understand it it's most accurate to regress your TH2 kills back to TH0 results, so:
Bugard Skins: TH0 = ~28%, your number was 32%
Bat Fangs: TH0 = 17.3%, your number was 18%
Bat Wings: TH0 = 17.5%, your number was 17%
Hare Meat: TH0 = 20.1%, your number was 20%
(lower sample size below here)
Black Tiger Fang: TH0 = 19.1%, your number was 23%
Giant Sheep Meat: TH0 = 27%, your number was 23%**
Bird Blood: TH0 = 29% predicted, you didn't test TH0**
Beehive Chip: TH0 = 21% predicted, you got 19%
Cock Meat: TH0 = 16.6% predicted, you got 20%

So, the starred ones are interesting because they shouldn't work as well as the others. Pooling drops from multiple slots should mess up the equation. The TH0 prediction from the formula should be higher than the actual TH0 drop rate.

You got:
40 Doubles - 32%
48 Singles - 38.4%
12 None - 9.6%
Which is frankly weird. Again, only 125 mobs, but it looks like you still got lucky on double drops.

[Mojo]

Those numbers do seem somewhat convincing, although I think it still might be useful to look at the "drops" "no drops" rate if not just to put an end to that theory (if it's possible to parse this out of your logs Thorny.) There would still be some unanswered questions though - How does Treasure Hunter +1 apply to this and are there caps on pools (never seeing six Abjuration drops from Kirin, for instance.)

[Enedin]

Well, as I understand it it's most accurate to regress your TH2 kills back to TH0 results, so:
Bugard Skins: TH0 = ~28%, your number was 32%
Bat Fangs: TH0 = 17.3%, your number was 18%
Bat Wings: TH0 = 17.5%, your number was 17%
Hare Meat: TH0 = 20.1%, your number was 20%
(lower sample size below here)
Black Tiger Fang: TH0 = 19.1%, your number was 23%
Giant Sheep Meat: TH0 = 27%, your number was 23%**
Bird Blood: TH0 = 29% predicted, you didn't test TH0**
Beehive Chip: TH0 = 21% predicted, you got 19%
Cock Meat: TH0 = 16.6% predicted, you got 20%

So, the starred ones are interesting because they shouldn't work as well as the others. Pooling drops from multiple slots should mess up the equation. The TH0 prediction from the formula should be higher than the actual TH0 drop rate.

You got:
40 Doubles - 32%
48 Singles - 38.4%
12 None - 9.6%
Which is frankly weird. Again, only 125 mobs, but it looks like you still got lucky on double drops.

Well, this is all very convincing to me. It seems that my testing has paid off, and that it is nearing conclusion. Does this also work for TH1 btw?

Oh yeah, don't forget that, when testing mobs that drop multiples of the same item, I multiplied the number of mobs killed by the maximum amount of multiple drops possible. Since 3 G. Sheep Meats is possible, every mob I killed accounted for 3, really. Maybe that's why? I did this because otherwise, the percentages would be over 100%, and I wanted the individual drop rate, not the combined.

[Masamune]

I have also a lot of drop data from my pDIF testing parses on Lesser Colibris and Colibris, but would have to do some work gathering them all to display results here.

Which format would you guys want me to display them ?

[Enedin]

I have also a lot of drop data from my pDIF testing parses on Lesser Colibris and Colibris, but would have to do some work gathering them all to display results here.

Which format would you guys want me to display them ?

Like Thorny's is all we need. Number of drops and number of mobs killed, TH level, TH+ items worn (if not, also mention). Any relevant information that's static across the test.

I started all my testing as TH2, and you can calculate...

TH0 = 1-cube_root(1-TH2) ---> most accurate
TH0 = 1-square_root(1-TH1)
TH1 = 1-(1-TH0)²

It's easy to change the formulas for when you only have TH1 or TH0 samples.

I'm inputting all these formulas in my chart (since I can rule out my assumption of a direct drop rate multiplier) and it's close alright. Will post later.

[Enedin]

Alright, chart done. It's probably unreadable, like most of my charts, but I'll try and make it clear.

http://enedin.be/RandomPics/THTEST2.png

The first chart is my OP chart with my theory scratched, and the calculations added in. I based TH0 and TH1 calculations off of TH2, but you can do it otherwise (see chart 3 for the formulas). In chart 1, "offset" is how far off the actual tested values are, in comparision to the calculations.

The second charts (3 of them) were made to quickly see the effects of TH0/TH1/TH2 when you're already farming with the given trait. Conclusions from that chart are in the 4th chart, the formulas used in the 3rd one. TH1+, for instance, displays the % increase in drop rates you'd get, compared to TH0. TH0- in the TH2 chart, displays how much drop rate you'd lose if you wouldn't have TH.

If these conclusions hold for Dynamis drop rates, that would mean that a THF pretty much triples drop rates on AF.

[Byrthnoth]

If these conclusions hold for Dynamis drop rates, that would mean that a THF pretty much triples drop rates on AF.

Yep! Did anyone ever find a Drop rate/mob for Dynamis AF? Honestly, Dynamis may be the undoing of this TH theory.

So, say there's a 1% drop rate for the 15 starter jobs (all in one slot, which we know from drop rates) off every mob with TH0. If you kill 100 mobs, you'd expect to get 15 AF. With TH2, you'd expect (1-.85^3) = ~38%. I know I kill more than 100 mobs per Dynamis - Sandy with TH4 (let alone TH2), and I know I don't get any 38 AF. If you break it down and say 1/128 instead of 1%, it's still about a ~30% predicted TH2 drop rate.

There could be a variety of reasons for this, and I'll go in order of biggest effect to smallest:
1) Currency could share the AF drop spot. (Does anyone have a screenshot of 4 Currency + RotZ AF?)
2) Crafting materials could share the AF drop spot. (Does anyone have a screenshot of Crafting Material + RotZ AF?)
3) Starter Relic weapons could share the same AF drop spot. (How about screenshots of D1 weapon + RotZ AF?)

Toss a 38/128 (~30%) drop rate of other crap into the AF slot (1/128 for each AF) and you end up with an 80% drop rate in that slot, but only a 15% drop rate of AF. 15 AF per 100 mobs? Yeah, that's getting closer.

Also, we know that new job AF (BLU, PUP, COR, DNC, SCH) all share a different drop slot. So we would need 1, 2, and 3 things for those job's AF as well.


Edit: Another thing to consider is how this theory accommodates TH4. If it's actually +1% (honestly, I'm not too convinced by Divisortheory's tests), then how would it work with multiple items per drop slot, like in Dynamis? It likely would not add +1% per item, because that would have a huge effect.

[Tortalius]

I would figure dyna is more of a "Does an AF drop?" if yes "choose from one of the 15 at random."

[Enedin]

I would figure dyna is more of a "Does an AF drop?" if yes "choose from one of the 15 at random."

Well yeah, only one RotZ AF can drop per mob.

EDIT: Byrthnoth's example says 1% drop per AF in Dyna, but I rather think it's something like 1% for ALL RotZ AF together, and that it chooses after that. Bringing TH2 at least boosts that to 3%.

[Byrthnoth]

Hypothetical City Dynamis Mob (For COP mobs, remove the items that obviously don't drop):
Slot 1: 15% RotZ AF, 30% Currency, 20% Relic weapons
Slot 2: 5% ToAU/WotG AF, 30% Currency
Slot 3: 30% Currency
Slot 4: 30% Currency
Slot 5: 10% Random Items (Wootz, etc)
Slot 6-10: Unused


Well, lets assume: 30% Currency drop rate, 15% AF drop rate, 20% Weapon drop rate (5% each, city).
65% drop rate total TH0, 35% no drop rate.
1-.35^3 = TH2 = 95% drop rate, 22% on AF, 30% on weapons, 44% on Currency

Now, lets assume: 30% Currency drop rate, 15% AF drop rate, no Weapon drop rate (0%, CoP).
45% drop rate total TH0, 55% no drop rate.
1-.55^3 = TH2 = ~83% drop rate, 28% on AF, 55% on currency

I would be very interested in seeing these three/six pictures:
1) A screenshot of 4 Currency + RotZ AF off one mob.
2) A screenshot of City Item + RotZ AF off one mob.
3) A screenshot of D1 weapon + RotZ AF off one mob.
4-6) The same pictures, but for ToAU/WotG jobs.

Think of it like trying to complete an FFXI quest. It potentially hours to do and has almost no reward beyond the self-satisfaction of having done it. If you're on Lakshmi, I'll send you 1400 gil so it really feels like an FFXI quest.

[Enedin]

To the people who haven't been following this thread, it seems that TH1 and TH2 work like this, example:

- Base drop rate (no TH) of let's say, 25%.
- TH0 (no TH trait) = 25% chance of obtaining item.
- TH1 adds another chance. So now you have 25% chance, and if that doesn't get you the item, you have another 25% chance on your second try. Some math shows us this chance is now 44%.
- TH2 adds ANOTHER chance! So if the first roll fails, and the second (with TH1) fails, you get another chance with TH2, this brings your total chances up to 58%.

In other words, having a TH trait is like having an extra character to lot with.

[Mojo]

After looking through some of my data a bit more thoroughly, I would say this theory is still be no means conclusive. It only works well if you completely disregard cases where it doesn't work. Dhalmel Meat, which can double drop, dropped at 100/1000 for TH0 and 358/700 for TH2. Looking at the TH2 data should produce the best results as the drop rate would be checked 3 times instead of one.

There are two slots for Dhalmel Meat, so 179/700 could be attributed to one slot.

1-(1-D)^3=179/700=179/700 => D = 0.093753

Applying this to two slots of the TH0 set.

2*1000*D=X/1000 => X = 188

Only 100 Dhalmel Meat were obtained during that set, so the prediction deviates heavily from what was observed.

[Byrthnoth]

I definitely agree. Every double-dropping thing posted so far has looked suspicious.

I tried to go out and got the TH0 sample for Bird Blood (with Tirgen). There are a few stipulations attached to this data set. First, I was on Trial 13 so I synced to level 30-32 so I could get xp. Second, I didn't think to do it until I was more than half done the trial. And finally, I had an AFK 30-32 WAR/THF in party the whole time. Finally, I play on PS2, so everything was done by hand. That said, I had the magian thing counting down at the same rate my sample size was counting up, so I didn't miss any.

TH0 Sample Size: 221
No drop - 123 Leeches - 56%
Single drop - 80 Leeches - 36%
Double drop - 18 Leeches - 8%

Enedin's TH1 Sample Size: 150
No drop - 55 Leeches - 36.6%
Single drop - 56 Leeches - 37.3%
Double drop - 39 Leeches - 26%

Enedin's TH2 Samples Size: 125
No drop - 50 - 40%
Single drop - 60 - 48%
Double drop - 15 - 12%

I think the sample sizes are a little too low to say anything conclusive though. If you look, the drop rate went down going from TH1 to TH2.

Edit:
Also, keep in mind that nothing forces SE to make double-dropper drop rates the same in both slots where it can drop. The maximum that we could predict would be a three-fold increase in drop rate though, which is exceeded by your dhalmel meat example. When you say "700" do you mean 350 (but they can double drop so you doubled it), or do you mean you actually killed 700 Dhalmels?

[Enedin]

My drop rates on TH1 on Leeches are assumed to be near the high end of the error margin, and the TH2 a bit low. I noticed this before, that's why I didn't use them, even though I could have (even w/o knowing TH0 results).

On multiple drop mobs, something is indeed fishy. But I'll say the results are conclusive enough for items that can't drop multiple times from the same mob.