|
So I am in act 1 inferno and started farming Halls of Agony, Warden and Butcher. Everywhere I read that one needs MF gear to do that effectively. Supposedly you have a much higher chance on getting rare (yellow) or legendary item drops when switching to it before the kill. I got some cheap MF gear (+147 %) to try it out.
For 4 runs (complete quest 9) I wrote down the number of yellows I got for each boss or elite kill while wearing or not wearing MF gear.
This is the average number of yellows I got per kill (in brackets number of kills):
Butcher with MF: 2.5 (2)
Butcher without MF: 3 (2)
Warden with MF: 2 (2)
Warden without MF: 2.5 (2)
Elite pack with MF: 0.875 (16, 10 of those with 5 stacks Neph valor)
Elite pack without MF: 0.65 (17, 7 of those with 5 stacks Neph valor)
Of course the number of kills here is way too small to say anything conclusive. But this is at least highly suspicious to me. I would have expected to get at least twice the number of yellows with +147% MF on gear. Yet I could not observe any real difference. The small difference with the elite packs can be easily explained with the on average higher stack of Nephalem valor during the kills with MF gear.
Does anyone have bigger data for this? Has anyone ever made an extensive statistical analysis about the effects of MF gear on item drops in D3? Because all I could find was a lot of hearsay about how it "should" work.
If switching to MF gear does indeed not work, I REALLY would like to know that. Because as a barb it is a real hassle to switch without dying, and good MF gear is quite expensive.
|
On May 30 2012 22:07 Redox wrote:
But this is at least highly suspicious to me. I would have expected to get at least twice the number of yellows with +147% MF on gear. Yet I could not observe any real difference.
MF affects the probability of finding blue or above items. You should be summing blues and above for this kind of analysis. The way it affects yellows/set/legendaries are a bit more complicated to calculate than simply multiplying the numbers together.
Also like you said, your sample size is too small. I say do 30 runs at least to see if there is any difference.
|
I think your sample size is way too small to be significant. Even with tons of magic find in D2 you could do dozens of runs without getting a good drop from a boss. I don't know for sure how D3 magic find works, but D2 magic find worked like this:
If a mob had a base 1% chance to drop a magic item with no magic find and you had 100% magic find, now the mob had a 2% chance to drop a magic item. If you had 20% magic find it would be 1.2% chance to drop a magic item. I don't know for sure if the same formula was used in D3, but I wouldn't be surprised if that was the case. Here is the D2 magic find info: classic.battle.net/diablo2exp/items/magic/magicfind.shtml
|
5 stack valor guarantees 2 yellow drops from bosses, aka warden and butcher (not Izual for some reason). Don't count that in. Testing with NV complicates things so much, you either take samples with only 5 stack NV or no NV at all.
And bigger sample size. I can take a sample size of 1 where a white trash mob drops a legendary item, but an elite doesn't, then QQ to the Blizz forums.
|
On May 30 2012 22:26 karazax wrote:I think your sample size is way too small to be significant. Even with tons of magic find in D2 you could do dozens of runs without getting a good drop from a boss. I don't know for sure how D3 magic find works, but D2 magic find worked like this: If a mob had a base 1% chance to drop a magic item with no magic find and you had 100% magic find, now the mob had a 2% chance to drop a magic item. If you had 20% magic find it would be 1.2% chance to drop a magic item. I don't know for sure if the same formula was used in D3, but I wouldn't be surprised if that was the case. Here is the D2 magic find info: classic.battle.net/diablo2exp/items/magic/magicfind.shtml
Can anyone confirm that the MF diminishing return formula on rare/set/uniques are still in place for D3?
|
On May 30 2012 22:20 Glacierz wrote:Show nested quote +On May 30 2012 22:07 Redox wrote:
But this is at least highly suspicious to me. I would have expected to get at least twice the number of yellows with +147% MF on gear. Yet I could not observe any real difference.
MF affects the probability of finding blue or above items. You should be summing blues and above for this kind of analysis. The way it affects yellows/set/legendaries are a bit more complicated to calculate than simply multiplying the numbers together. Also like you said, your sample size is too small. I say do 30 runs at least to see if there is any difference.
Number of blues:
Elite packs with MF: 1.29 (17 kills)
Elite packs without MF: 2.06 (16 kills).
So the overall number of dropped items was even much higher when NOT wearing MF gear.
But you are right about the sample size. I was just hoping someone had something better. I just cant believe that noone has ever recorded his drops in D3 and examined this question.
Well or if noone has any real data, maybe you can share your experiences? I mean every farmer must have done kills with and without wearing his MF gear.
|
On May 30 2012 22:29 Heh_ wrote:
5 stack valor guarantees 2 yellow drops from bosses, aka warden and butcher (not Izual for some reason). Don't count that in. Testing with NV complicates things so much, you either take samples with only 5 stack NV or no NV at all.
So if 5 stacks of valor give 2 guaranteed drops, would 75% of MF from gear give the same? Or is there a diffrence between MF % from valor buff and from gear? Because this does just sound like it.
|
On May 30 2012 22:43 Redox wrote:Show nested quote +On May 30 2012 22:29 Heh_ wrote:
5 stack valor guarantees 2 yellow drops from bosses, aka warden and butcher (not Izual for some reason). Don't count that in. Testing with NV complicates things so much, you either take samples with only 5 stack NV or no NV at all.
So if 5 stacks of valor give 2 guaranteed drops, would 75% of MF from gear give the same? Or is there a diffrence between MF % from valor buff and from gear? Because this does just sound like it.
No, valor gives 2 rares by default, 75% MF does not guarantee anything, only guarantee bigger chance.
|
On May 30 2012 22:44 bgx wrote:Show nested quote +On May 30 2012 22:43 Redox wrote:On May 30 2012 22:29 Heh_ wrote:
5 stack valor guarantees 2 yellow drops from bosses, aka warden and butcher (not Izual for some reason). Don't count that in. Testing with NV complicates things so much, you either take samples with only 5 stack NV or no NV at all.
So if 5 stacks of valor give 2 guaranteed drops, would 75% of MF from gear give the same? Or is there a diffrence between MF % from valor buff and from gear? Because this does just sound like it. No, valor gives 2 rares by default, 75% MF does not guarantee anything, only guarantee bigger chance.
Source?
|
On May 30 2012 22:47 Redox wrote:Show nested quote +On May 30 2012 22:44 bgx wrote:On May 30 2012 22:43 Redox wrote:On May 30 2012 22:29 Heh_ wrote:
5 stack valor guarantees 2 yellow drops from bosses, aka warden and butcher (not Izual for some reason). Don't count that in. Testing with NV complicates things so much, you either take samples with only 5 stack NV or no NV at all.
So if 5 stacks of valor give 2 guaranteed drops, would 75% of MF from gear give the same? Or is there a diffrence between MF % from valor buff and from gear? Because this does just sound like it. No, valor gives 2 rares by default, 75% MF does not guarantee anything, only guarantee bigger chance. Source?
Well, don't know if it counts, but I've beaten Kulle in Act 2 several times with more than 80% MF (no stacks) and I've never gotten a yellow from him.
|
On May 30 2012 22:49 Tobberoth wrote:Show nested quote +On May 30 2012 22:47 Redox wrote:On May 30 2012 22:44 bgx wrote:On May 30 2012 22:43 Redox wrote:On May 30 2012 22:29 Heh_ wrote:
5 stack valor guarantees 2 yellow drops from bosses, aka warden and butcher (not Izual for some reason). Don't count that in. Testing with NV complicates things so much, you either take samples with only 5 stack NV or no NV at all.
So if 5 stacks of valor give 2 guaranteed drops, would 75% of MF from gear give the same? Or is there a diffrence between MF % from valor buff and from gear? Because this does just sound like it. No, valor gives 2 rares by default, 75% MF does not guarantee anything, only guarantee bigger chance. Source? Well, don't know if it counts, but I've beaten Kulle in Act 2 several times with more than 80% MF (no stacks) and I've never gotten a yellow from him.
That would kinda reinforce my suspicion that the MF from items does not work. 
That is indeed the kind of info I was looking for.
|
On May 30 2012 22:51 Redox wrote:Show nested quote +On May 30 2012 22:49 Tobberoth wrote:On May 30 2012 22:47 Redox wrote:On May 30 2012 22:44 bgx wrote:On May 30 2012 22:43 Redox wrote:On May 30 2012 22:29 Heh_ wrote:
5 stack valor guarantees 2 yellow drops from bosses, aka warden and butcher (not Izual for some reason). Don't count that in. Testing with NV complicates things so much, you either take samples with only 5 stack NV or no NV at all.
So if 5 stacks of valor give 2 guaranteed drops, would 75% of MF from gear give the same? Or is there a diffrence between MF % from valor buff and from gear? Because this does just sound like it. No, valor gives 2 rares by default, 75% MF does not guarantee anything, only guarantee bigger chance. Source? Well, don't know if it counts, but I've beaten Kulle in Act 2 several times with more than 80% MF (no stacks) and I've never gotten a yellow from him. That would kinda reinforce my suspicion that the MF from items does not work. That is indeed the kind of info I was looking for.
Honestly, that's not the kind of information you should be looking for. You should be looking for something like "I did 50 butcher runs with 5 stacks of NV and got X blues, Y yellows and Z legendaries, then did 50 butcher runs with +150% MF and 5 stacks of NV and got X blues, Y yellows and Z legendaries". Something like "I did that boss a few time with MF and got nothing" doesn't actually mean that much.
|
One stack of valor guarantees one extra magical item, and 4 stacks gives one guaranteed rare, 5 stacks two guaranteed rares. Tome of Secret count as a magical item. With 5 stacks you should get five extra drops from "boss monsters" like Warden or The Butcher.
Magic Find increases your chance of finding a magical item, I don't think it increases your chances of finding an item, only that the white items you find have a higher chance of being blue, or better.
|
Also if you look at the link I posted for Diablo 2 magic find, Blizzard considered 400% average, and 600% plus was good. Even with 600 then you were only boosting a mob with a 1% chance of having a magical item drop to having a 6% chance, not guaranteeing 6x as many magic items from a boss.
|
Against bosses with 5 stack of valor, you will notice that the bonus magical items seemingly appear out of nowhere, at the very end of the item drop spam. If you don't believe us, then go test it out yourself. Or do you want us to link a scientific paper in a peer-reviewed journal as proof?
|
Elite pack with MF: 0.875 (16, 10 of those with 5 stacks Neph valor)
Elite pack without MF: 0.65 (17, 7 of those with 5 stacks Neph valor)
Can you provide the items dropped per run in an excel format:
Run (no mf) - Loots
1 - 3
2 - 0
3 - 3
etc.
With that data, we can run a normality test and bootstrap / t-test the data and determine whether or not it's statistically significant.
|
On May 30 2012 22:51 Redox wrote:Show nested quote +On May 30 2012 22:49 Tobberoth wrote:On May 30 2012 22:47 Redox wrote:On May 30 2012 22:44 bgx wrote:On May 30 2012 22:43 Redox wrote:On May 30 2012 22:29 Heh_ wrote:
5 stack valor guarantees 2 yellow drops from bosses, aka warden and butcher (not Izual for some reason). Don't count that in. Testing with NV complicates things so much, you either take samples with only 5 stack NV or no NV at all.
So if 5 stacks of valor give 2 guaranteed drops, would 75% of MF from gear give the same? Or is there a diffrence between MF % from valor buff and from gear? Because this does just sound like it. No, valor gives 2 rares by default, 75% MF does not guarantee anything, only guarantee bigger chance. Source? Well, don't know if it counts, but I've beaten Kulle in Act 2 several times with more than 80% MF (no stacks) and I've never gotten a yellow from him. That would kinda reinforce my suspicion that the MF from items does not work. That is indeed the kind of info I was looking for.
it's not that simple, i remember for example that i heard in a developers talk about a year ago they dont want people to switch gear like in d2. it could be that they calculate it based of the items you wear when you entered the fight.
i watched kungens stream when they where doing goblin runs and they did it mainly with MF gear but also sometimes without and then there was a big difference in drops.
|
MF from Valor and MF from gears are difference at least in my experience. When I first started Butcher Inferno, with no NV stack I switch to 100+ MF at the end but the boss only drop blue.
Once I was able to farm him with 5 stacks of NV, 2 yellows are guaranteed, you get 3 yellows many times and sometime(very rare) a fourth yellow appear.
NV Magic Find>>> Item magic find, at least on boss fight.
|
On May 30 2012 23:19 Caphe wrote:
MF from Valor and MF from gears are difference at least in my experience. When I first started Butcher Inferno, with no NV stack I switch to 100+ MF at the end but the boss only drop blue.
Once I was able to farm him with 5 stacks of NV, 2 yellows are guaranteed, you get 3 yellows many times and sometime(very rare) a fourth yellow appear.
NV Magic Find>>> Item magic find, at least on boss fight.
yes, NV is way way better than an equivalent MF.
I saw it explained this way, both increase the chance of magic items dropping but NV ups the minimum amount that can drop as well, so you are guaranteed better loot.
EDIT: Looking at your example again, it looks like you only used NV when farming elites, in which case it should be the exact same as +magic find so it may not have mattered.
|
On May 30 2012 23:05 karazax wrote:
Also if you look at the link I posted for Diablo 2 magic find, Blizzard considered 400% average, and 600% plus was good. Even with 600 then you were only boosting a mob with a 1% chance of having a magical item drop to having a 6% chance, not guaranteeing 6x as many magic items from a boss.
We are not talking about mobs with 1 % drop chance here though.
If bosses (or elites) have a 1% chance of dropping additional loot and with 100 % MF I increase it to 2% chance, then switching to MF becomes totally pointless and I would stop doing it.
On May 30 2012 23:13 RAGEMOAR The Pope wrote:
Can you provide the items dropped per run in an excel format:
Run (no mf) - Loots
1 - 3
2 - 0
3 - 3
etc.
With that data, we can run a normality test and bootstrap / t-test the data and determine whether or not it's statistically significant.
Too lazy. Tbh I was more hoping someone else had such data or a link to it.
But I can tell you right away that my data would fail a t-test. 
On May 30 2012 23:24 rastaban wrote:
EDIT: Looking at your example again, it looks like you only used NV when farming elites, in which case it should be the exact same as +magic find so it may not have mattered.
I always had 5 stacks NV when doing the bosses, I only did complete runs.
|
|
|
|
|
|
EPT
