Treasure Hunter testing: data gathering

[Enedin]

In part I we found out that TH1 and TH2 both add another roll. I'm currently testing to see whether the trend continues with TH3, 4 and so on.

I'd like a bit of help from BG on this one. I've chosen a mob that's easy to kill, easy to count, easy to link, easy to reach and most of all: it drops a shitload of stuff.

Labyrinth Lizards in Crawler's Nest. You can find them here:

http://enedin.be/www/images/cnth.png

- They always drop one tail. (which makes counting kills easy)
- They link by sound. (which makes afk killing easy)
- They repop fast. (5 minutes)
- They can drop up to 6 Lizard Eggs.
- They can drop up to 5 Lizard Skins.

We're not in the the theories part yet, so don't bring up your view on how/why TH works, unless it's backed by statistical observation. For now, we're just killing stuff and gathering data.

The data I have so far:

http://enedin.be/www/images/cnthtable.png

The drop rates are calculated using the max amount of drops per item. For eggs, that's 6, for skins, that's 5. So killing 1 monster gives us 6 possible egg drops, which means 6 tests. Killing 50 brings the number of egg tests to 300 already, and the number of skin tests to 250. So the good thing about this mob is that we reach a low enough error ratio faster because of the multi-drops.

If you want to contribute to testing, please watch out for TH Kupower (it adds a TH tier), AND for TH tier increasing when on THF. Not paying attention to this skews results. I did it by focusing on TH4 kills when I was going for TH3, and writing the TH4 kills/drops down on a piece of paper. When I had 50 TH3 kills, I'd just count my inventory, and subtract the drops I got on TH4.

Any data is welcome. I don't have any way (w/o Kupower) to test TH1 atm, so that would be most helpful, but like I said, any data is welcome.

Oh and edit: moonphase doesn't affect drop rate, I've tested that on Stirges in Ranguemont Pass. So don't let that stop ya.

[Septimus]

If I could play I could test TH1. My second character got a TH1 Tarutaru Sash, he could BLU the fuck out of the lizards.

Silly hard drive.

[vagus]

Wanna lvl bst and test out dipper while you're at it. Please and thank you.

[Motenten]

I don't believe that that method of sampling will give you fully reliable data.

Reference post I made analyzing drop rates of items with multiple slots available: http://forum.square-enix.com/ffxi/th...l=1#post210053

It -appears- as if the second and following items on multi-drop items are dependent variables, not independent variables. As such, you need totals on a per-count basis.

In addition, you indicate that the lizards' loot pool has a max of 6 eggs, 5 skins and 1 tail, giving a total of 12 items. Hasn't it already been shown that the max loot pool is 10 items? As such it seems like there would be limits on what you could get from a lizard, separate from its full potential.

[Mojo]

To be honest, even the last round of testing was not conclusive. There were situations where the assertions you've made simply did not hold true. You ought to focus on finding out why it falls short first before moving on, imo.

[Yugl]

If I could play I could test TH1. My second character got a TH1 Tarutaru Sash, he could BLU the fuck out of the lizards.

Silly hard drive.

Tbh, I agree with this. BLU can much more readily kill mobs from the TH1-3 range.

[Enedin]

We're not in the the theories part yet, so don't bring up your view on how/why TH works, unless it's backed by statistical observation. For now, we're just killing stuff and gathering data.

^

Edit: I've been trying to find a maximum for drops, which, if it is indeed 10, would indeed skew my results. But still, data is data. I'd just like to see something that shows that the max is 10.

[Byrthnoth]

What is your current maximum for drops? Iirc, it is 10 per monster. Some of your Eggs/Skins (probably many of them) share slots. You have to figure:
1 Slot: Crystal (3 slots? 6 slots? How many crystals can drop off one mob with multiple people with Signet?)
1 Slot: Beastman/Kindred Seal
5 Slots: Lizard Skin
6 Slots: Lizard Egg

If we assume a maximum of 10, then most of the Skin/Egg slots must be shared. The Crystal and BS/KS slots may be weird, like they only occur to replace an open slot or something.

[Mafai]

What is your current maximum for drops? Iirc, it is 10 per monster. Some of your Eggs/Skins (probably many of them) share slots. You have to figure:
1 Slot: Crystal (3 slots? 6 slots? How many crystals can drop off one mob with multiple people with Signet?)
1 Slot: Beastman/Kindred Seal
5 Slots: Lizard Skin
6 Slots: Lizard Egg

If we assume a maximum of 10, then most of the Skin/Egg slots must be shared. The Crystal and BS/KS slots may be weird, like they only occur to replace an open slot or something.

10 per monster isn't enough, unless your doing it on a monster basis? I've seen drops distribute before the whole treasure pool is loaded a couple times. Maybe this is a special case for NMs? The only time I've noticed it happen is from Fafnir but I'm sure its happened other times.

Edit: As Byrthnoth stated, crystals were more then likely involved in my over 10 items dropping. Doesn't conquest determine crystal drops, not the mob?

[noodles355]

Regarding 10 drop slots and Crystals/Seals replacing one of those slots, wasn't there instancdes of Pandemonium Warden dropping all 10 drops as well as crystals/seals?

[Enedin]

I've never seen 10 items drop from the Labyrinth Lizards, max I saw was 9, min I saw was 1 (always tail). But the possibility of a shared slot is there. This is the first time I tackle these Lizards (been thinking about it long time) and all I see is that the skin drop rate doesn't climb as much as the egg drop rate when killing with a higher TH tier. So eggs might be hogging the slots in higher TH?

I'm no expert on multi-drops, I've only tested singles so far. Currently killing a few hundred Pit Hares in pally mines. Those are /ra one-shot, so TH can't increase. If you'd like to help out, that is also a possibility.

I remember people convinced that crystals took up a spot in the drop pool so they would ask us to not get Signet for Sky.

[Mojo]

11+ drops are certainly possible although like everybody else I'm sure that it only occurs when you get crystals (and possibly seals?) Honestly though I haven't spend a lot of time killing mobs that have more than 10 items in their loo pool.

Either way, I recommend you make your data collection a bit more comprehensive. A bulk drop rate leaves out a lot of details about how pools of items could specifically be affected by TH.

https://docs.google.com/spreadsheet/...hl=en_US#gid=1

Here's my old data. Every drop combination that occurred was recorded separately. It's laborious but given what is already known about drop pools and TH there isn't a way around it unless you're aiming to overlook something potentially important.

I can potentially help come winter break although at the moment I'm hella bogged down.

[Enedin]

I've been killing Pit Hares today. I know my sample sizes aren't large enough, but it's a start. I had already tested TH0/1/2, today I did some 3/4/5:

TH0: 79/400 kills = 20%
TH1: 162/440 kills = 37%
TH2: 97/200 kills = 49%
TH3: 97/200 kills = 49%
TH4: 53/100 kills = 53%
TH5: 57/100 kills = 57%

These are very easy to test since I can one-shot them with a ranged attack. I'll go to Ghelsba tomorrow and find some new monsters. For now, I'll stay away from multi-drop mobs, 'cause there's clearly something different about the way they drop their items.

[Ophannus]

Is this real?

http://img15.imageshack.us/img15/720...1119214313.png

http://img528.imageshack.us/img528/9...1119214318.png

http://img196.imageshack.us/img196/4...1119214324.png

http://img64.imageshack.us/img64/480...1119214548.png

[Septimus]

Is this real?

http://img15.imageshack.us/img15/720...1119214313.png

http://img528.imageshack.us/img528/9...1119214318.png

http://img196.imageshack.us/img196/4...1119214324.png

http://img64.imageshack.us/img64/480...1119214548.png

Having TH8 while facing NE on full moon darksday while wearing a subligar decreases drop rates.

He got his urban legand wrong. The rumor was that TH10 was worse than TH9. Mocchi apparently tested it and found that TH10 was better than TH9. The general problem is that since the effect of TH is so low, it requires a large sample size to help factor out randomness. And by large, I mean thousands to tens of thousands of data points. It is all a giant pain in the ass because people see a small amount anecdotal evidence and confuse it with fact because they don't understand concepts like "Regression to the mean". (i.e. They see two full treasure pools with THF7, but a lousy treasure pool with TH8 and assume it is a fact, ignoring all of the lousy treasure pools with TH7 that happened after that.)

[cdgreg]

It's really amazing what people will believe because item X dropped a few times with a certain TH number.

[Ophannus]

Getting 0/1 with TH8 and getting 1/1 on Loki's also helped to reinforce them further >_>

Whatever, enjoying my new THF body.

[Septimus]

Getting 0/1 with TH8 and getting 1/1 on Loki's also helped to reinforce them further >_>

Whatever, enjoying my new THF body.

Because those numbers imply statistical significance! Somewhere there is a mathematician doubled over in pain. As long as you got it, that is what matters, just don't listen to them about game mechanics. lol

[Darkmagi]

Face palmed so hard at those pics...wtf seriously?

[Ophannus]

The second they saw TH8 they literally stopped curing and turned their backs to let Turul kill them so they can HP and come back to have it reset. This was while it was at 20% with Blue Proc on.