BL S Kits vs Master Kits (Results)(Ectos)

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Geotherma.2395

Geotherma.2395

A quick note to forum mods so I don’t get infracted lol This is NOT the same thread as in Game Discussion, this is a complete comparison of BL Salvage Kits to Master kits.

So to begin-

I previously did a study of BL Salvage Kits on 200 rares. That info can be found at https://forum-en.guildwars2.com/forum/game/gw2/BL-Salvage-Kit-Results-200-Rares-Ectos/first#post1611719.

However, this thread I will compare them as this seems like a more appropriate forum for the thread.

My results for Black Lion Salvage Kits for 200 rare armor pieces are:
157 runes (5g 78s 15c)
257 ecto (60g 78s 70c)
159 thick leather
29 hardened leather
4 orichalcum
94 mithril
25 elder wood logs
7 gossamer scraps
112 silk scraps
Video: http://youtu.be/U-Mxe7ojb2w

My results for Black lion salvage kits #2
173 runes
239 ecto
108 thick leather straps
7 hardened leather section
3 orichalchum
123 mithril ore
17 gossamer scraps
203 silk scraps
Video:

My results for Master Salvage Kits for 200 rare armor pieces are:
143 runes
186 ecto
157 thick leather sections
20 hardened leather sections
3 orichalcum
37 mithril ore
0 elder wood
14 gossamer scraps
158 silk scraps
Video: http://youtu.be/DEAqfWCsJJI

My results for Master Salvage Kits for 200 rare armor pieces #2 are:
(200 Level 80 Rampager’s Banded Greaves of Balthazar
All contained Major Rune of Balthazar 0/4)
Runes 171
171 ecto
17 orichalchum
253 mithril
Video:
Note*I assume the type of armor could be a variable in the materials other than ectos.
(note that neither test calculated correctly the runes as all pieces did not have upgrades.Does NOT apply to master kit test #2)

So the reason behind all this was because I noticed a large amount of people complained that Arenanet made BL kits as a scam and they were no better than masters. So I wanted to post unbiased results from as many rares as I could do (money wise..) even at my own expense as you can see I lost a great deal of profit. But i hope this thread shows people how the kits work and they can factor in for themselves what is more worthwhile using.

Some important tips:
Buying kits directly for ectos is not as profitable over time.
It is best to use free kits you are given as rewards from BL chests or from rewards.
I recommend not using BL salvage kits on anything lower than level 77 rares. If you do, that’s up to you.
Don’t forget Master Kits cost 15s+ for 25. BL kits are sometimes given free, or ina decent amount as rewards for BL chest.

Thank you all for your time and hope this was helpful.

http://youtu.be/y5HHPuAoy-Y BL HD version
http://youtu.be/2emy9UYLzxg Master kit HD version

Total rares now tested-400 with BL kit, 400 with Master kit.

Intel i7 3.9ghz processor 16GB Ram 2TB HDD
Nvidia GTX 650 Win 7 64bit FFXI 4+yrs/Aion 4+ years Complete Noob~ Veteran OIF/OEF
http://everyonesgrudge.enjin.com/home MY GW2 Music http://tinyurl.com/cm4o6tu

(edited by Geotherma.2395)

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Davidjack.1507

Davidjack.1507

http://wiki.guildwars2.com/wiki/Mystic_Salvage_Kit 250 uses for 26s plus mystic forge stones. same chances as a master salvage kit.

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Zumy.6318

Zumy.6318

In my eyes 200 samples is jut not enough.
I have a way bigger difference between the two kits.

Thief | Zumy [Buka]
Legendary counter: Twilight, Bolt, Incinerator, Incinerator Nr. 2, Meteorlogicus, The Dreamer

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Geotherma.2395

Geotherma.2395

In my eyes 200 samples is jut not enough.
I have a way bigger difference between the two kits.

200 is the most I could do, but with the Wiki facts that support my evidence, the devs input on those threads, and my own experience I can say with very little doubt that what is on the wiki is very accurate.

But if anyone has video + written evidence to the contrary
I see no problem in posting that info here.

The general consensus from players is that BL works better for rarer mats and ectos, but is supported by the wiki/devs like I said and now this small sample test. I welcome any supporting info that devs would like to provide as well.

Intel i7 3.9ghz processor 16GB Ram 2TB HDD
Nvidia GTX 650 Win 7 64bit FFXI 4+yrs/Aion 4+ years Complete Noob~ Veteran OIF/OEF
http://everyonesgrudge.enjin.com/home MY GW2 Music http://tinyurl.com/cm4o6tu

(edited by Geotherma.2395)

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: munkiman.3068

munkiman.3068

This is exactly what i’d expect, on average i get 25-35% more ectos using BLSK. We had this guild conversation about the same thing, pertaining to you getting more ectos with the BLSK. Many people were under the impression if you were just going for ectos any kit would do. Obviously a wrong assumption. You really don’t need a bigger sample size to see a clear difference.

[TAO] Founder/Owner and Administrator for the NSP Server Website

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Ubi.4136

Ubi.4136

I had been wondering about how ‘bad’ the ecto rate would be from salvaging from master kits. I have 8 toons and 4 have world completion so lots of zone clears, and have been lucky to always have black lion salvage kits available…but I finally ran out.

Salvaging the items from 2 full storage panels, 1 with the last black lion kit and the other with the master kit netted different results than I thought. As I was only interested in ecto, it’s the only thing I kept track of.

From the black lion salvages I got 18 ecto. From the master salvages I got 26 ecto.
All of the items were rares from the dragon events/maw/etc chests.

We could definitely use a larger sample of the population and a better way to track it.

Ubi / Ghostly Krewe / Ghastly Krewe / Ghostly Night
Kronk Smash / Ghostly Gadget / Guardian Krewe / Serene Earth
Wolves of Valor (Wolf) – Tarnished Coast

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: DarkSpirit.7046

DarkSpirit.7046

In my eyes 200 samples is jut not enough.
I have a way bigger difference between the two kits.

200 is the most I could do, but with the Wiki facts that support my evidence, the devs input on those threads, and my own experience I can say with very little doubt that what is on the wiki is very accurate.

But if anyone has video + written evidence to the contrary
I see no problem in posting that info here.

The general consensus from players is that BL works better for rarer mats and ectos, but is supported by the wiki/devs like I said and now this small sample test. I welcome any supporting info that devs would like to provide as well.

Sorry but 200 salvages is not enough to prove that BLSK produces more ectos than Master. It is just luck.

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Buttercup.5871

Buttercup.5871

I salvaged more than the OP with both sets (mystic/master vs Black Lion), and have gotten similar results (regarding ectos). They fixed the Black Lion kit so that the ecto return rate is higher than the mystic/master kit. Makes sense too, because the Black Lion kit is very expensive when you actually buy it with gems.

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Syeria.4812

Syeria.4812

In my eyes 200 samples is jut not enough.
I have a way bigger difference between the two kits.

200 is the most I could do, but with the Wiki facts that support my evidence, the devs input on those threads, and my own experience I can say with very little doubt that what is on the wiki is very accurate.

But if anyone has video + written evidence to the contrary
I see no problem in posting that info here.

The general consensus from players is that BL works better for rarer mats and ectos, but is supported by the wiki/devs like I said and now this small sample test. I welcome any supporting info that devs would like to provide as well.

Sorry but 200 salvages is not enough to prove that BLSK produces more ectos than Master. It is just luck.

You (and others) keep coming up with this completely baseless claim and posting it all over.

Here’s the issue (this is directed at all of you statistically deficient people), sample sizes could always be bigger, but something having an insufficient sample size for a conclusion never proves the opposite of the hypothesis is true. A statistical test can have two results, reject the null hypothesis or fail to reject the null hypothesis. You cannot prove the null hypothesis. If this test failed to reject the null, then all we can say is “we lack sufficient data to make a conclusion.” But you people consistently take that as proof of the null.

Is 200 data points enough to prove Black Lion kits produce higher ecto per salvage results? Maybe, maybe not. I haven’t done the math to calculate the proper sample size to determine significance at 5% for these. However, I’d be shocked if the requisite sample size for determining solely that Black Lion kits have a better salvage rate (rather than determining what that rate is, simply qualifying that Black Lion are “better” than Master/Mystic) was more than a few hundred. When you add these data to any other set that has been tracked, it’s almost certain that the results are statistically significant.

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Syeria.4812

Syeria.4812

In regards to the actual test, thanks for doing it and sharing the results. I have a few questions about the data.

What level range of rares did you use? For example, were they all level 76-80 or some other range like that? Did you track the level during the testing? Were the samples for both Black Lion and Master kits all approximately the same distribution of item level?

In regard to the data, I’d like to compare it to the rates I’ve gotten from my salvage tracking:

Black Lion
Geotherma – 1.285 ecto/salvage
Syeria – 1.253 ecto/salvage

@ 200 salvages
Geotherma – 257 ectos
Syeria – 251 ectos (rounded to nearest full)

Master/Mystic
Geotherma – 0.93 ectos/salvage
Syeria – 0.88 ectos/salvage

@ 200 salvages
Geotherma – 186 ectos
Syeria – 176 ectos

Your data fits extremely well with mine, with a difference of about 2.5% for Black Lion and about 5.5% for Master/Mystic.

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Smooth Penguin.5294

Smooth Penguin.5294

It’s nice that you shared your research, but RNG is RNG. Someone could get much different results due to luck. For example, I’ve been getting more “3 ecto” drops from lv 80 rares with my Mystic Salvage then I was getting with my BLSK.

“You are the exception.”
- Mike O’Brien

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Behellagh.1468

Behellagh.1468

Well for a small % chance rolls, 200 samples may or may not be enough to infer drop rate with any degree of accuracy. That’s simple Prop and Stats 101.

400 is more helpful. And we may be dealing with two rolls for ectos. First if we get any and second how many within a range. Also the chances within the range may not be the same (ie 70% – 1 30% – 2). The actual mechanic is difficult to surmise only using the results of approx 1.25 for BL and 0.9 for Master/Mystic.

But this does appear to support the claim that BL kits aren’t worth it for just ectos as 0.35 ectos is worth at current sell price around 9s 23c while the difference is use between the two kits is 23s and change (100 Gems@1g98s).

We are heroes. This is what we do!

RIP City of Heroes

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Ursan.7846

Ursan.7846

In my eyes 200 samples is jut not enough.
I have a way bigger difference between the two kits.

You wanna calculate the confidence level of a 200 data points, then come back and tell me if the sample size is too small?

I am too lazy to do the calculations myself. But no matter how big your sample size is, you’re not going to prove anything 100%. It’s up to you to be satisfied with 95% confidence or 99% confidence, however. I wonder with a sample size of 200, how confident can you be with your results.

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Syeria.4812

Syeria.4812

Well for a small % chance rolls, 200 samples may or may not be enough to infer drop rate with any degree of accuracy. That’s simple Prop and Stats 101.

400 is more helpful. And we may be dealing with two rolls for ectos. First if we get any and second how many within a range. Also the chances within the range may not be the same (ie 70% – 1 30% – 2). The actual mechanic is difficult to surmise only using the results of approx 1.25 for BL and 0.9 for Master/Mystic.

But this does appear to support the claim that BL kits aren’t worth it for just ectos as 0.35 ectos is worth at current sell price around 9s 23c while the difference is use between the two kits is 23s and change (100 Gems@1g98s).

This sounds nice and impressive to skeptics, but fortunately for stats, it actually doesn’t matter if there’s 1 “roll,” 2 “rolls,” or 300 “rolls.” The math works out exactly the same. If a black lion kit has a 70% chance to produce ectos (note, I’m not just making up numbers, although I am aggressively rounding) but produces them in this distribution 1 @ 50%, 2 @ 25%, 3 @ 25%, that’s exactly statistically the same as producing an average of 1.225 ectos per salvage. The only effect of the different distribution is a higher standard deviation, which may force a very slightly larger sample size requirement.

As to your concern about a sample size of 200, if he had 400 for the sample size, you’d be demanding 800. If he had 800 you’d be demanding 1600. If he had 1600 you’d be demanding 3200. If you’re actually concerned about the sample size, do the math yourself and show us that the sample is statistically insufficient. If you can’t do that math (estimate the values you don’t know, but tell us your estimates and why) then I’m afraid you’re not qualified to criticize the sample size.

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Behellagh.1468

Behellagh.1468

This sounds nice and impressive to skeptics, but fortunately for stats, it actually doesn’t matter if there’s 1 “roll,” 2 “rolls,” or 300 “rolls.” The math works out exactly the same. If a black lion kit has a 70% chance to produce ectos (note, I’m not just making up numbers, although I am aggressively rounding) but produces them in this distribution 1 @ 50%, 2 @ 25%, 3 @ 25%, that’s exactly statistically the same as producing an average of 1.225 ectos per salvage. The only effect of the different distribution is a higher standard deviation, which may force a very slightly larger sample size requirement.

As to your concern about a sample size of 200, if he had 400 for the sample size, you’d be demanding 800. If he had 800 you’d be demanding 1600. If he had 1600 you’d be demanding 3200. If you’re actually concerned about the sample size, do the math yourself and show us that the sample is statistically insufficient. If you can’t do that math (estimate the values you don’t know, but tell us your estimates and why) then I’m afraid you’re not qualified to criticize the sample size.

If you are trying to determine the probability of an unknown system more samples are always better since, as you point out, it conforms better to the theoretical standard distribution (assuming our understanding of the mechanism behind it is correct). More samples simply improve the odds that the mean we see is accurate.

Lets go back to the old mainstay of probability, the coin flip (or a die roll). If I flip a coin 10 times and get 7 heads it would be incorrect for me to assume that there’s a 70% chance for heads. If I flip a coin 100 times the result should be closer to 50% and 1000 flips should be closer still. Except we know that there should be a 50% chance with a coin or a 1/6th chance with a die. Here we are blind to what we should be seeing so we don’t know how accurate our results are since we don’t have a base line to test it against.

There is probably a way to do it, I’m sure a math or science major could pop in and school me on my 30 year old knowledge of one prop and stat course I took in college.

We are heroes. This is what we do!

RIP City of Heroes

(edited by Behellagh.1468)

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Syeria.4812

Syeria.4812

This sounds nice and impressive to skeptics, but fortunately for stats, it actually doesn’t matter if there’s 1 “roll,” 2 “rolls,” or 300 “rolls.” The math works out exactly the same. If a black lion kit has a 70% chance to produce ectos (note, I’m not just making up numbers, although I am aggressively rounding) but produces them in this distribution 1 @ 50%, 2 @ 25%, 3 @ 25%, that’s exactly statistically the same as producing an average of 1.225 ectos per salvage. The only effect of the different distribution is a higher standard deviation, which may force a very slightly larger sample size requirement.

As to your concern about a sample size of 200, if he had 400 for the sample size, you’d be demanding 800. If he had 800 you’d be demanding 1600. If he had 1600 you’d be demanding 3200. If you’re actually concerned about the sample size, do the math yourself and show us that the sample is statistically insufficient. If you can’t do that math (estimate the values you don’t know, but tell us your estimates and why) then I’m afraid you’re not qualified to criticize the sample size.

If you are trying to determine the probability of an unknown system more samples are always better since, as you point out, it conforms better to the theoretical standard distribution (assuming our understanding of the mechanism behind it is correct). More samples simply improve the odds that the mean we see is accurate.

Lets go back to the old mainstay of probability, the coin flip (or a die roll). If I flip a coin 10 times and get 7 heads it would be incorrect for me to assume that there’s a 70% chance for heads. If I flip a coin 100 times the result should be closer to 50% and 1000 flips should be closer still. Except we know that there should be a 50% chance with a coin or a 1/6th chance with a die. Here we are blind to what we should be seeing so we don’t know how accurate our results are since we don’t have a base line to test it against.

There is probably a way to do it, I’m sure a math or science major could pop in and school me on my 30 year old knowledge of one prop and stat course I took in college.

That’s a false equivalency though. We’re not saying we’ve looked at 10 salvages and decided a rate based on it. We’re trying to determine two things: are Black Lion kits better than Masters/Mystic, and what are the rates. For the former, the proper analogy is that we’ve done 200 six-sided die rolls and 200 coin flips. We’ve determined from this that coin flips are more likely to produce a single given outcome than a dice roll.

On top of that, you’re trying to frame things very poorly. You’ve apparently decided that a 99.9999999% accuracy with a confidence interval of +/- 0.00000000001 is the only acceptable result. No one would ever actually want or expect that from this analysis. If we accept a reasonable confidence interval of say +/- 0.1 at 95% confidence, the required sample size is going to be relatively tiny. Required sample sizes are always way smaller than people expect.

If you’re going to claim the sample is too small, you’d better actually do the math to determine that fact. The OP probably should provide that math himself, but really he hasn’t claimed the sample is sufficient either. If you want to disprove his data, the onus is on you. Just shouting “you need more data” isn’t enough unless you can actually mathematically show that more data is needed.

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Ensign.2189

Ensign.2189

Do you happen to have the raw numbers on 0’s, 1’s, 2’s, and 3’s salvaged per rare? It isn’t legible from the youtube videos, and that information is essential if you want to put confidence intervals on these things.

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Geotherma.2395

Geotherma.2395

In my eyes 200 samples is jut not enough.
I have a way bigger difference between the two kits.

200 is the most I could do, but with the Wiki facts that support my evidence, the devs input on those threads, and my own experience I can say with very little doubt that what is on the wiki is very accurate.

But if anyone has video + written evidence to the contrary
I see no problem in posting that info here.

The general consensus from players is that BL works better for rarer mats and ectos, but is supported by the wiki/devs like I said and now this small sample test. I welcome any supporting info that devs would like to provide as well.

Sorry but 200 salvages is not enough to prove that BLSK produces more ectos than Master. It is just luck.

How about the fact that the wiki states it does, and devs have posted that it does? Does that cover it?

Intel i7 3.9ghz processor 16GB Ram 2TB HDD
Nvidia GTX 650 Win 7 64bit FFXI 4+yrs/Aion 4+ years Complete Noob~ Veteran OIF/OEF
http://everyonesgrudge.enjin.com/home MY GW2 Music http://tinyurl.com/cm4o6tu

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Geotherma.2395

Geotherma.2395

In regards to the actual test, thanks for doing it and sharing the results. I have a few questions about the data.

What level range of rares did you use? For example, were they all level 76-80 or some other range like that? Did you track the level during the testing? Were the samples for both Black Lion and Master kits all approximately the same distribution of item level?

In regard to the data, I’d like to compare it to the rates I’ve gotten from my salvage tracking:

Black Lion
Geotherma – 1.285 ecto/salvage
Syeria – 1.253 ecto/salvage

@ 200 salvages
Geotherma – 257 ectos
Syeria – 251 ectos (rounded to nearest full)

Master/Mystic
Geotherma – 0.93 ectos/salvage
Syeria – 0.88 ectos/salvage

@ 200 salvages
Geotherma – 186 ectos
Syeria – 176 ectos

Your data fits extremely well with mine, with a difference of about 2.5% for Black Lion and about 5.5% for Master/Mystic.

I wanted to keep some things constant, unfortunately I didn’t think about upgrades until after.. But the gear itself was all Armor, was various types, and was ONLY level 80. Aside from the lacking a average for upgrades which would have been cool to find out as well, that’s the best i could do without spanning this over like 3 months of buying very specific gears heh

Intel i7 3.9ghz processor 16GB Ram 2TB HDD
Nvidia GTX 650 Win 7 64bit FFXI 4+yrs/Aion 4+ years Complete Noob~ Veteran OIF/OEF
http://everyonesgrudge.enjin.com/home MY GW2 Music http://tinyurl.com/cm4o6tu

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Geotherma.2395

Geotherma.2395

Well for a small % chance rolls, 200 samples may or may not be enough to infer drop rate with any degree of accuracy. That’s simple Prop and Stats 101.

400 is more helpful. And we may be dealing with two rolls for ectos. First if we get any and second how many within a range. Also the chances within the range may not be the same (ie 70% – 1 30% – 2). The actual mechanic is difficult to surmise only using the results of approx 1.25 for BL and 0.9 for Master/Mystic.

But this does appear to support the claim that BL kits aren’t worth it for just ectos as 0.35 ectos is worth at current sell price around 9s 23c while the difference is use between the two kits is 23s and change (100 Gems@1g98s).

Aside from the gem cost if you “do” buy them i just want to note that the kits themselves can be free and allow for 100% return of an upgrade. Just in case newer people are reading and weren’t aware of this difference. Sadly my results weren’t tracking that, but I learned after trial #1 to toss the runes in mystic forge which netted me a large amount of my spent money back ^^

Intel i7 3.9ghz processor 16GB Ram 2TB HDD
Nvidia GTX 650 Win 7 64bit FFXI 4+yrs/Aion 4+ years Complete Noob~ Veteran OIF/OEF
http://everyonesgrudge.enjin.com/home MY GW2 Music http://tinyurl.com/cm4o6tu

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Geotherma.2395

Geotherma.2395

Do you happen to have the raw numbers on 0’s, 1’s, 2’s, and 3’s salvaged per rare? It isn’t legible from the youtube videos, and that information is essential if you want to put confidence intervals on these things.

I think I have an HD version of the BL kits, I can upload that sometime soon. I would look at it directly but I’m lacking much time to do so today.

Intel i7 3.9ghz processor 16GB Ram 2TB HDD
Nvidia GTX 650 Win 7 64bit FFXI 4+yrs/Aion 4+ years Complete Noob~ Veteran OIF/OEF
http://everyonesgrudge.enjin.com/home MY GW2 Music http://tinyurl.com/cm4o6tu

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Behellagh.1468

Behellagh.1468

This sounds nice and impressive to skeptics, but fortunately for stats, it actually doesn’t matter if there’s 1 “roll,” 2 “rolls,” or 300 “rolls.” The math works out exactly the same. If a black lion kit has a 70% chance to produce ectos (note, I’m not just making up numbers, although I am aggressively rounding) but produces them in this distribution 1 @ 50%, 2 @ 25%, 3 @ 25%, that’s exactly statistically the same as producing an average of 1.225 ectos per salvage. The only effect of the different distribution is a higher standard deviation, which may force a very slightly larger sample size requirement.

As to your concern about a sample size of 200, if he had 400 for the sample size, you’d be demanding 800. If he had 800 you’d be demanding 1600. If he had 1600 you’d be demanding 3200. If you’re actually concerned about the sample size, do the math yourself and show us that the sample is statistically insufficient. If you can’t do that math (estimate the values you don’t know, but tell us your estimates and why) then I’m afraid you’re not qualified to criticize the sample size.

If you are trying to determine the probability of an unknown system more samples are always better since, as you point out, it conforms better to the theoretical standard distribution (assuming our understanding of the mechanism behind it is correct). More samples simply improve the odds that the mean we see is accurate.

Lets go back to the old mainstay of probability, the coin flip (or a die roll). If I flip a coin 10 times and get 7 heads it would be incorrect for me to assume that there’s a 70% chance for heads. If I flip a coin 100 times the result should be closer to 50% and 1000 flips should be closer still. Except we know that there should be a 50% chance with a coin or a 1/6th chance with a die. Here we are blind to what we should be seeing so we don’t know how accurate our results are since we don’t have a base line to test it against.

There is probably a way to do it, I’m sure a math or science major could pop in and school me on my 30 year old knowledge of one prop and stat course I took in college.

That’s a false equivalency though. We’re not saying we’ve looked at 10 salvages and decided a rate based on it. We’re trying to determine two things: are Black Lion kits better than Masters/Mystic, and what are the rates. For the former, the proper analogy is that we’ve done 200 six-sided die rolls and 200 coin flips. We’ve determined from this that coin flips are more likely to produce a single given outcome than a dice roll.

On top of that, you’re trying to frame things very poorly. You’ve apparently decided that a 99.9999999% accuracy with a confidence interval of +/- 0.00000000001 is the only acceptable result. No one would ever actually want or expect that from this analysis. If we accept a reasonable confidence interval of say +/- 0.1 at 95% confidence, the required sample size is going to be relatively tiny. Required sample sizes are always way smaller than people expect.

If you’re going to claim the sample is too small, you’d better actually do the math to determine that fact. The OP probably should provide that math himself, but really he hasn’t claimed the sample is sufficient either. If you want to disprove his data, the onus is on you. Just shouting “you need more data” isn’t enough unless you can actually mathematically show that more data is needed.

Where do you think I’m asking for 99.999999999% confidence? All I’m was saying is more is better. We had 200 of each before, now we have 400 of each. Confidence is improved, end of story.

Understanding the mechanism behind it however could help to determine if it gets accidentally broken or intentionally changed. But when attempting to model an unknown random event, more samples simply provides more clues to the underlying mechanic.

I guess not everyone took a clock apart to see how it worked as a kid.

We are heroes. This is what we do!

RIP City of Heroes

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Ensign.2189

Ensign.2189

I think I have an HD version of the BL kits, I can upload that sometime soon. I would look at it directly but I’m lacking much time to do so today.

EDIT – Nevermind, the video has enough information in it to calculate the percentages of each amount – we don’t need the HD versions. Except the last two salvages cut off of the BLK video. GAH.

Just to be clear, the frequency of each drop amount determines the variance between drops, which has a really big effect on how fast your results regress to the mean, and consequently, your sampling error margins.

For example, at a 0.9 ectos/salvage average with 200 data points, we could add on a 95% confidence interval of +/- 0.0425 ectos per salvage if 90% of the salvages were for 1 ecto and 10% were for 0; that’s a very robust result.

At the same time, if they came from getting 3 ectos 30% of the time and 0 ectos 70% of the time (a 0.9 ecto/salvage average) with the same 200 data points, our 95% confidence interval would be +/- 0.195 ectos per salvage – more thakittentimes as large!

So we really do kind of need that information if we want to make any kind of firm statement about whether the amount of data we have is ‘enough’.

(edited by Ensign.2189)

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Obtena.7952

Obtena.7952

You (and others) keep coming up with this completely baseless claim and posting it all over.

It’s not baseless at all. If someone was so inclined and had sufficient background in statistics, they could tell you the level of confidence. Unfortunately, mine is too rusty for that. I can only provide the ‘200 is pretty reasonable’ sample size without actually telling you that level of confidence.

More samples won’t give you a radically different number of average ecto per salvage, it will simply give you a more accuracy and more powerful statistics The information and precision we are after, 200 samples is more than sufficient. For arguments sake, if the number is 0.9 ecto per salvage, then doing more samples will simply give you more decimal places. It won’t move too much from 0.9

You also have to remember that we only have only 4 possible outcomes from a salvage (0, 1, 2 or 3 ectos). That means you build the distribution much faster (approach mean value faster) than you would for an unbound range.

Sorry naysayer’s, 200 is just fine.

(edited by Obtena.7952)

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Ensign.2189

Ensign.2189

Standard error for these tests is:

SQRT( ( P0*avg^2 + P1*(avg-1)^2 + P2*(avg-2)^2 + P3*(avg-3)^2 ) / (samplesize-1) )

Where:

“avg” is the average number of ectos per salvage over the entire sample;
“P0” is the percentage of salvages that returned 0 ecto;
“P1” is the percentage that returned 1 ecto;
“P2” is the percentage that returned 2 ecto;
“P3” is the percentage that returned 3 ecto; and
“samplesize” is the total number of salvages made.

If you have no information on the percentages of each salvage result, use the worst case scenario, which is P3 equal to the total ectos salvaged divided by three times the number of salvages made, and P0 being the remainder.

EDIT – wasn’t rendering properly without the pre tags.

(edited by Ensign.2189)

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Geotherma.2395

Geotherma.2395

Added HD versions to the original post.

Intel i7 3.9ghz processor 16GB Ram 2TB HDD
Nvidia GTX 650 Win 7 64bit FFXI 4+yrs/Aion 4+ years Complete Noob~ Veteran OIF/OEF
http://everyonesgrudge.enjin.com/home MY GW2 Music http://tinyurl.com/cm4o6tu

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Geotherma.2395

Geotherma.2395

I think I have an HD version of the BL kits, I can upload that sometime soon. I would look at it directly but I’m lacking much time to do so today.

EDIT – Nevermind, the video has enough information in it to calculate the percentages of each amount – we don’t need the HD versions. Except the last two salvages cut off of the BLK video. GAH.

Yea… Game froze at the most stupid time…

Intel i7 3.9ghz processor 16GB Ram 2TB HDD
Nvidia GTX 650 Win 7 64bit FFXI 4+yrs/Aion 4+ years Complete Noob~ Veteran OIF/OEF
http://everyonesgrudge.enjin.com/home MY GW2 Music http://tinyurl.com/cm4o6tu

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Leablo.2651

Leablo.2651

As pointed out in another of the threads OP made for this, the sample sets appear to be a random assortment of rares. The sample size is moot since it doesn’t appear to be a controlled sample anyway.

I appreciate the effort that was put into this but unfortunately the results don’t tell us anything. If you are pursuing this issue out of academic curiosity, you have to use scientific method, otherwise you are just wasting money. It is better to have a controlled sample of 100 items than to use a random set of the 200 cheapest items you can find.

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Geotherma.2395

Geotherma.2395

As pointed out in another of the threads OP made for this, the sample sets appear to be a random assortment of rares. The sample size is moot since it doesn’t appear to be a controlled sample anyway.

I appreciate the effort that was put into this but unfortunately the results don’t tell us anything. If you are pursuing this issue out of academic curiosity, you have to use scientific method, otherwise you are just wasting money. It is better to have a controlled sample of 100 items than to use a random set of the 200 cheapest items you can find.

If a dev can verify that all armor types of the same level will yield the same results then the results would be fine. There will NEVER be a way to see 100% accuracy to figure this out, but like I said before since devs DO have the tools to know and have stated as such its a good indicator that what they say is true. My evidence is merely unbiased results and I am in no way going to buy 200 specific armor types to help some TP player profit from this. All items were Armor types and all were level 80 Rare.

I understand the scientific method, however, I also understand the dollar/time method. Meaning I’m not going to spend 100’s of gold to appease a few people with a conspiracy theory as I have stated unbiased evidence and they have not.

Intel i7 3.9ghz processor 16GB Ram 2TB HDD
Nvidia GTX 650 Win 7 64bit FFXI 4+yrs/Aion 4+ years Complete Noob~ Veteran OIF/OEF
http://everyonesgrudge.enjin.com/home MY GW2 Music http://tinyurl.com/cm4o6tu

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Geotherma.2395

Geotherma.2395

I’ve added a second set of results of 200 more rares with BL salvage kits.

Intel i7 3.9ghz processor 16GB Ram 2TB HDD
Nvidia GTX 650 Win 7 64bit FFXI 4+yrs/Aion 4+ years Complete Noob~ Veteran OIF/OEF
http://everyonesgrudge.enjin.com/home MY GW2 Music http://tinyurl.com/cm4o6tu

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Geotherma.2395

Geotherma.2395

I’ve added 200 more rares with master salvage kits.

However, please note that these were:
200 Level 80 Rampager’s Banded Greaves of Balthazar
All contained Major Rune of Balthazar 0/4

Results are top of page.

Intel i7 3.9ghz processor 16GB Ram 2TB HDD
Nvidia GTX 650 Win 7 64bit FFXI 4+yrs/Aion 4+ years Complete Noob~ Veteran OIF/OEF
http://everyonesgrudge.enjin.com/home MY GW2 Music http://tinyurl.com/cm4o6tu

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Geotherma.2395

Geotherma.2395

For those curious on how the test went coin wise, well I ended up better than I started heh. Thanks to the Mystic Forge change I started doing after the 2nd test I made quite a few superior runes including the Superior Divinity Rune. Thanks to that my losses with master kits were covered.

Started with 48g
Ended entire test with 80g after all fees,travel, overpriced test #2 rares etc.

Intel i7 3.9ghz processor 16GB Ram 2TB HDD
Nvidia GTX 650 Win 7 64bit FFXI 4+yrs/Aion 4+ years Complete Noob~ Veteran OIF/OEF
http://everyonesgrudge.enjin.com/home MY GW2 Music http://tinyurl.com/cm4o6tu

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Geotherma.2395

Geotherma.2395

I think I have an HD version of the BL kits, I can upload that sometime soon. I would look at it directly but I’m lacking much time to do so today.

EDIT – Nevermind, the video has enough information in it to calculate the percentages of each amount – we don’t need the HD versions. Except the last two salvages cut off of the BLK video. GAH.

Just to be clear, the frequency of each drop amount determines the variance between drops, which has a really big effect on how fast your results regress to the mean, and consequently, your sampling error margins.

For example, at a 0.9 ectos/salvage average with 200 data points, we could add on a 95% confidence interval of +/- 0.0425 ectos per salvage if 90% of the salvages were for 1 ecto and 10% were for 0; that’s a very robust result.

At the same time, if they came from getting 3 ectos 30% of the time and 0 ectos 70% of the time (a 0.9 ecto/salvage average) with the same 200 data points, our 95% confidence interval would be +/- 0.195 ectos per salvage – more thakittentimes as large!

So we really do kind of need that information if we want to make any kind of firm statement about whether the amount of data we have is ‘enough’.

I have the numbers but not sure how to display them on here. In excel:
1.285 Average
200 Total Rares
257 Total Ecto
Results with COUNTIF 0’s,1’s,2’s,3’s.
45 Zeros
91 Ones
26 Twos
38 Threes

These are from test 1 Black lion kit. As this took an incredible amount of time to figure out lol.. It may take some time to post all results.

Intel i7 3.9ghz processor 16GB Ram 2TB HDD
Nvidia GTX 650 Win 7 64bit FFXI 4+yrs/Aion 4+ years Complete Noob~ Veteran OIF/OEF
http://everyonesgrudge.enjin.com/home MY GW2 Music http://tinyurl.com/cm4o6tu

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Obtena.7952

Obtena.7952

As pointed out in another of the threads OP made for this, the sample sets appear to be a random assortment of rares. The sample size is moot since it doesn’t appear to be a controlled sample anyway.

I think there is a reasonable balance to be struck here. I think more people are interested in a practical result vs. the scientifically rigorous one. If the more random sample gives more variation in the final result than the controlled one, I think that’s actually more interesting because it’s really what happens ingame when you loot a rare item. Sample size isn’t ever moot … a bigger sample provides you a higher confidence level. With the number of outcomes available for this experiment, you can come to a reasonable conclusion pretty fast.

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Obtena.7952

Obtena.7952

some great work

If you can or know how, please post your standard deviation (Yes, I know it’s not a very good statistic in this case)

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Geotherma.2395

Geotherma.2395

some great work

If you can or know how, please post your standard deviation (Yes, I know it’s not a very good statistic in this case)

Okay, so I’m no math whiz but I gave it a shot lol..

BL Kit Test 1
Total Numbers 200
Mean (Average):1.285
Standard deviation:1.0193
Variance(Standard deviation):1.03897
Population Standard deviation:1.01675
Variance(Population Standard deviation):1.03377

Master Kit Test 1
Total Numbers: 200
Mean (Average): 0.93
Standard deviation: 0.81141
Variance(Standard deviation): 0.65839
Population Standard deviation: 0.80938
Variance(Population Standard deviation): 0.6551

BL kit Test 1
If it follows the normal distribution
The 68.3% measure confidence range, ? 0.26570129537695 – 2.3042987046231
The 90% measure confidence range, 1.645? -0.39174636910492 – 2.9617463691049
The 95% measure confidence range, 1.960? -0.71282546106118 – 3.2828254610612
The 99% measure confidence range, 2.576? -1.340713463109 – 3.910713463109
The 99.9% measure confidence range, 3.291? -2.0695120369145 – 4.6395120369145

Master Kit Test 1
If it follows the normal distribution
The 68.3% measure confidence range, ? 0.11858644342173 – 1.7414135565783
The 90% measure confidence range, 1.645? -0.40477530057125 – 2.2647753005713
The 95% measure confidence range, 1.960? -0.66037057089341 – 2.5203705708934
The 99% measure confidence range, 2.576? -1.1602013217456 – 3.0202013217456
The 99.9% measure confidence range, 3.291? -1.7403620146991 – 3.6003620146991

BL Kit Test 1
Population size:200

Mean (?): 1.285
Median: 1
Mode: 1
Lowest value: 0
Highest value: 3
Range: 3
Interquartile range: 1
First quartile: 1
Third quartile: 2
Variance (?2): 1.033775
Standard deviation (?): 1.0167472645648
Quartile deviation: 0.5
Mean absolute deviation (MAD): 0.8376

Master Kit Test 1
Population size:200

Mean (?): 0.93
Median: 1
Mode: 1
Lowest value: 0
Highest value: 3
Range: 3
Interquartile range: 1
First quartile: 0
Third quartile: 1
Variance (?2): 0.6551
Standard deviation (?): 0.80938248066041
Quartile deviation: 0.5
Mean absolute deviation (MAD): 0.5394

This is just for the first test of BLvsMaster salvage test so far. Based on 200rares vs 200rares.

Intel i7 3.9ghz processor 16GB Ram 2TB HDD
Nvidia GTX 650 Win 7 64bit FFXI 4+yrs/Aion 4+ years Complete Noob~ Veteran OIF/OEF
http://everyonesgrudge.enjin.com/home MY GW2 Music http://tinyurl.com/cm4o6tu

(edited by Geotherma.2395)

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Xynn.2748

Xynn.2748

Wow, I had no idea that BL was better for ectos. I thought it was the same as Master’s, but it guaranteed the upgrade was salvaged. Thanks!

(edited by Xynn.2748)

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: DarkSpirit.7046

DarkSpirit.7046

Is 200 data points enough to prove Black Lion kits produce higher ecto per salvage results? Maybe, maybe not. I haven’t done the math to calculate the proper sample size to determine significance at 5% for these. However, I’d be shocked if the requisite sample size for determining solely that Black Lion kits have a better salvage rate (rather than determining what that rate is, simply qualifying that Black Lion are “better” than Master/Mystic) was more than a few hundred. When you add these data to any other set that has been tracked, it’s almost certain that the results are statistically significant.

And we are saying that 200 may not be enough of a sample size to account for the “luck” factor. I won’t be surprised if someone repeats this little experiment and come up with the opposite conclusion after just 200 salvages.

(edited by DarkSpirit.7046)

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Syeria.4812

Syeria.4812

Is 200 data points enough to prove Black Lion kits produce higher ecto per salvage results? Maybe, maybe not. I haven’t done the math to calculate the proper sample size to determine significance at 5% for these. However, I’d be shocked if the requisite sample size for determining solely that Black Lion kits have a better salvage rate (rather than determining what that rate is, simply qualifying that Black Lion are “better” than Master/Mystic) was more than a few hundred. When you add these data to any other set that has been tracked, it’s almost certain that the results are statistically significant.

And we are saying that 200 may not be enough of a sample size to account for the “luck” factor. I won’t be surprised if someone repeats this little experiment and come up with the opposite conclusion after just 200 salvages.

Then you have no idea what statistics are.

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Ursan.7846

Ursan.7846

And we are saying that 200 may not be enough of a sample size to account for the “luck” factor. I won’t be surprised if someone repeats this little experiment and come up with the opposite conclusion after just 200 salvages.

Hahahahahahaha

Wait what? Is 201 be large enough to account for the “luck” factor? 202? 300? 400? What arbitrary number do you think is enough of a sample size to account for the “luck” factor, I’m curious?

Though I’d like to give you the benefit of the doubt, the way you worded it makes it seem like you fail to grasp the concept of statistics.

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Syeria.4812

Syeria.4812

Standard error for these tests is:

SQRT( ( P0*avg^2 + P1*(avg-1)^2 + P2*(avg-2)^2 + P3*(avg-3)^2 ) / (samplesize-1) )

Where:

“avg” is the average number of ectos per salvage over the entire sample;
“P0” is the percentage of salvages that returned 0 ecto;
“P1” is the percentage that returned 1 ecto;
“P2” is the percentage that returned 2 ecto;
“P3” is the percentage that returned 3 ecto; and
“samplesize” is the total number of salvages made.

If you have no information on the percentages of each salvage result, use the worst case scenario, which is P3 equal to the total ectos salvaged divided by three times the number of salvages made, and P0 being the remainder.

EDIT – wasn’t rendering properly without the pre tags.

Using this formula against my personal data (randomly acquired rares of at least level 76) my results (converted to confidence interval at 95%):

Black Lion
1.253 ectos/salvage +/- 0.163 (13%)

Master/Mystic
0.889 ectos/salvage +/- 0.081 (9%)

This assumes I converted the standard error to confidence interval correctly (it’s been at least 7 years since I’ve done any of this, so that’s really just a question of whether the equation I googled was accurate or not).

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Ursan.7846

Ursan.7846

Using this formula against my personal data (randomly acquired rares of at least level 76) my results (converted to confidence interval at 95%):

Black Lion
1.253 ectos/salvage +/- 0.163 (13%)

Master/Mystic
0.889 ectos/salvage +/- 0.081 (9%)

This assumes I converted the standard error to confidence interval correctly (it’s been at least 7 years since I’ve done any of this, so that’s really just a question of whether the equation I googled was accurate or not).

Huh, well thanks for taking time to do the calculations. I’d actually opine that 200 is a bit too small a sample size then, assuming your calculations are correct.

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: munkiman.3068

munkiman.3068

The kits aren’t RNG, they are based on actual hard numbers. The only RNG you could attribute to it is the amount of ectos and even then it’s within 5%.

Every post i’ve seen and the pretty large now samples sizes i’ve done are all within 3-5% variance. As a matter of fact, if you get’s the kits for free (BLSK) it’s far cheaper to buy a stack of rare armor @ 23s each and sell of the runes and mats.

[TAO] Founder/Owner and Administrator for the NSP Server Website

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Geotherma.2395

Geotherma.2395

For those who didn’t notice, I’ve done 200 more tests making the total tests 400 BL kits vs 400 Master kits. I’m still working on the test#2 individual results. The updated qty is at the original post at top of page.

Intel i7 3.9ghz processor 16GB Ram 2TB HDD
Nvidia GTX 650 Win 7 64bit FFXI 4+yrs/Aion 4+ years Complete Noob~ Veteran OIF/OEF
http://everyonesgrudge.enjin.com/home MY GW2 Music http://tinyurl.com/cm4o6tu

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Obtena.7952

Obtena.7952

^^ Is waiting for the obligatory “400 isn’t enough” post.

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Irenicus.9801

Irenicus.9801

Geo, my understanding is that you have two of each sample (2×2×200 for BL and Master kits). Could you provide more info about the other samples’ yield per salvage as you have for one sample already, whenever you have time of course

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Geotherma.2395

Geotherma.2395

Geo, my understanding is that you have two of each sample (2×2×200 for BL and Master kits). Could you provide more info about the other samples’ yield per salvage as you have for one sample already, whenever you have time of course

Yea that’s what I’m working on now. The carpal tunnel really kicking in haha. See I have to watch the video and in excel press what gets salvaged “0,1,2,3” and if I sneeze, blink, have an itch within the 15 minutes I have to start all over lols But I’m working on it ^^

Intel i7 3.9ghz processor 16GB Ram 2TB HDD
Nvidia GTX 650 Win 7 64bit FFXI 4+yrs/Aion 4+ years Complete Noob~ Veteran OIF/OEF
http://everyonesgrudge.enjin.com/home MY GW2 Music http://tinyurl.com/cm4o6tu

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Geotherma.2395

Geotherma.2395

^^ Is waiting for the obligatory “400 isn’t enough” post.

Ikr :P

15conspiracy theories

Intel i7 3.9ghz processor 16GB Ram 2TB HDD
Nvidia GTX 650 Win 7 64bit FFXI 4+yrs/Aion 4+ years Complete Noob~ Veteran OIF/OEF
http://everyonesgrudge.enjin.com/home MY GW2 Music http://tinyurl.com/cm4o6tu

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Ensign.2189

Ensign.2189

Using this formula against my personal data (randomly acquired rares of at least level 76) my results (converted to confidence interval at 95%):

Black Lion
1.253 ectos/salvage +/- 0.163 (13%)

Master/Mystic
0.889 ectos/salvage +/- 0.081 (9%)

This assumes I converted the standard error to confidence interval correctly (it’s been at least 7 years since I’ve done any of this, so that’s really just a question of whether the equation I googled was accurate or not).

That’s consistent with the original poster’s data (95% confidence interval is trivially close to +/- 2 standard errors). Thanks!

BL S Kits vs Master Kits (Results)(Ectos)

in Black Lion Trading Co

Posted by: Ensign.2189

Ensign.2189

So the reason I’m pressing for standard deviation data is to calculate standard errors. When you know the standard errors of multiple samples of the same distribution, you can combine that data using inverse-square error weighting to get a joint distribution with an even smaller standard error.

To wit, at 95% confidence:

Black Lion Salvage Kit:
Syeria: 1.253 +/- 0.163
Geotherma: 1.285 +/- 0.1442
Joint: 1.271 +/- 0.108

Master Salvage Kit:
Syeria: 0.889 +/- 0.081
Geotherma: 0.93 +/- 0.1148
Joint: 0.903 +/- 0.0662

Obviously more data always helps, but their data combined is starting to look pretty robust, especially for the Master Salvage Kit case.