Magic find Test

[Dragon Brother 619]

But do you include normal monsters as well?

[locolagarto 15]

I just did 3 Guardian runs with Gran Canon

 

1st run:

1 Rare in hidden location on Start isle(lucky)

15 Rare

4 Magic

 

2nd Run

no hidden location but first chest in machine gave only medium Gold(unlucky)

1 unique

8 Rare

9 Magic

 

3rd run

hidden spot gave a white hoop(unlucky)

1 set

2 unique

7 Rare

5 magic

 

Hmmp not the results I thought I would get. I still think there are lucky and unlucky servers, and characters, but the 1 hidden spot on start island is probably not the best indicator. As my third run looks to be the lucky one.

[Shattered Rift 2]

I don't mean to rain the parade, but...

 

If you get 1 drop how do you conclude the server is 'lucky'? Wouldn't that be like concluding that hitting a homerun was due to lucky socks if the day you wore those socks you hit a home run? If you win once at roulette does that mean you should bet all your money on the next round because you are 'lucky'?

 

I mean clearly there is a seed, but I believe it is pseudorandom. Which means that just because you find 6 good drops at hiding spot doesn't mean you will get more drops from boss. Or just because grunwald dragon drops something good doesn't mean swamp dragon will..

 

...

 

Another example is 6 6sided dice rolled. If you get 4 sixes does that mean your roll had a good seed? And next you get no sixes so thats a bad seed?

Precisely.

 

Pseudorandomizers, as I recall, are designed in such a way as to be "perfectly" random. Suppose a pseudorandomizer is only going to be used to simulate a six-sided die. The numbers 1, 2, 3, 4, 5, and 6 all exist in the pseudorandomizer in a specific order (let's say the order is 4, 3, 1, 6, 2, 5). (In a full-scale pseudorandomizer, the magnitude is much, much higher than a mere six numbers.)

 

You roll the die once, and you start over it if you don't roll a 6. But when you roll a 6, your next three results are going to be 2 (poor), 5 (very good), and 4 (good). Because you started off with 6, two of your next three rolls were guaranteed to be good.

 

Granted, Sacred 2's pseudorandomizer is infinitely more complex than this (and far more balanced since you're rolling at least D1000 or D10000 if not more (which would then be divided back down to size)). Without knowing the exact pseudorandomizer and the various formulas, we'll never be able to determine anything about the seed. (If we assume there is no special "luck" factor, starting off with a rare/unique/it's-been-too-long-since-I-played-Sacred-to-remember-the-right-terms would actually make you very slightly less likely to get another good item.)

 

... this went in a different direction than I expected when I started typing the post. In short: if there's a luck factor, it would probably modify either the rolls themselves or your drop bins. And, let's face it, we're still not entirely sure whether magic find% adds or multiplies the sizes of the drop bins (though I'd argue adding to them is more likely).

 

Its like this Oh I said abracadabra and then I rolled two heads in a row with flipping a coin.... No you would need more data to prove that was true than 2 coin tosses. It might take 10 tosses 9 heads and 1 tails. And then there would still be a chance that it was a coincidence. Maybe 100 tosses 80 heads 20 tails and you could show to a high confidence level that the magic word 'abracadabra' makes it heads more often.

 

 

Another thing noticed in psychology called confirmation bias has shown that the brain is wired (my words and I am not a psychologist) to notice information that confirms our beliefs and not notice information that contradicts. This is why you have to be careful when studying things such as randomness in video games.

To use another example: I'm sure everyone has played Pokémon at some point in their lives. Most of us will remember, back in the day, that each of us had a different button combination we pressed when throwing a PokéBall, and we were sure it made it more likely that we would capture the Pokémon. However, there was no button input during this time: it was all speculation and superstition. There are also various studies with rodents (and probably people as well) to back up Claudius' second paragraph in the above post.

 

Perhaps we should define a set of parameters and have everyone test them?

 

1. Check the hidden spots on Start Island, note down what was found.

2. Run a selection of predefined areas in a close proximity to a portal. Boar, guardians, poison lord, blood forest, etc.

3. Note the items that drop.

4. Rinse, repeat.

 

Do this 10 times and note any correlation between the initial spots and the gear found. The more people that do it, the more data we will be able to collect. This still won't prove anything conclusively, but it will give us a much higher level of confidence as claudius said.

 

If nothing else, you should collect some loot/xp for your troubles

Even knowing what we do about the bins, it's not that simple. We know that the randomizer is compressed to a D1000. You would need to grab a fairly large number of drops to make sure that each run took a fair sample. (I'm not sure how many, but 100 drops would be a bare minimum. You'd probably want a solid 500, 1,000, or more.)

 

 

There are some things in this game that can be figured out. This is one of the ones that will probably never be conclusively (or even substantially) proven.

 

But have fun with it, guys, . (Sorry if I came across as a downer: I'm just trying to emphasize how insanely difficult it would be to even compile the evidence to suspect that there's any validity to lucky characters or lucky servers.)

[gogoblender 3,070]

But have fun with it, guys, .

 

That's it. We want to have some fun with this!

 

So to return the topic back to Furian's original direction... has anyone seen findings like Furian's that they wish to comment upon?

 

We all understand how statistics and odds have been theorized, and now we're all head up with how larger samples are needed etc etc... and how a gut feeling is baaaaaaaad etc etc ... and now this thread is full of them ...while we have lost the initial spirit and precise question of what Furian is asking here.

 

We're instead looking to see if anyone has seen what Furian is seeing, and has, perhaps, some insight into what is happening with some of their own research or great insight into his take on this with Furian's original topic direction in mind.

 

We are theorizing and starting off this topic with the assumption that... that there is something else at work here.

 

We would like to know if anyone has conjecture on what it could be only if they are willing to start off firstly with the premise that there is something else at work here.

 

That is the "camp" of this thread ^^

 

Ideas please

 

Post away!

 

 

gogo

 

p.s. Of course, there's nothing stopping anyone from opening up a new topic of "gut" feeling versus large statistical sampling. If the topic is interesting, I encourage people to start up a new topic on that as well so that it does not, again, take Furian's topic off topic. Cheers!

[essjayehm 58]

I will "stoke the fire" I guess.

 

I have seen evidence of this, but only "qualitative" - that is to say there is no "real" numbers to back up the theory, but only observations.

 

A while back, when I was building Trystesta, I was in dire need of runes, because VD dryads don't pick up a lot. So, I would create a Shadow Warror, do a "run", which was a clockwise loop down to Gold Coast, then up the path along the mountains killing kobolds (mostly) looking for runes. About half of the SW's I ran picked up 8 or 10 runes in a run (about 15 mins). I did notice that the other half of the time, I would get either 2 or 3 runes, or 20 runes; with the "low" number happening more often than the "high" number.

 

As a rough guess, I would say that 33% "unlucky", picking up 2 or 3 runes; 50% "standard" getting 8 or 10, and 17% "lucky". Using ONLY the starting equipment (I didn't equip whatever armour dropped out of starting chest, and only used the beginning sword/shield), I felt this was a nicely standardized test, and that there was "something else" going on with the drop system other than "kill speed".

 

I did this run for a few days... so I'll guess that I ran it about 75 or 100 times, and at that time, I was pretty "sold" on the idea of a lucky toon, the guy who got the 20 runes in a run. I didn't fathom the idea of a lucky server, but that does seem like a plausible theory as well. Even a lucky toon maybe having a better chance for a lucky server seems to be just as viable as either explaination. But all these things cannot ever really be proven, no matter how many tests, or the sample size, or whatever other "statistical" proofs that can be done... because of the random nature of the game.

 

Remember the "golden" rule of science - observing something affects the behaviour of the system being observed!

[FrostElfGuard 9]

I started a new character 6 times before I got a gold drop on the start island.

I'm playing on PC (HC multiplayer) with very limited resources... so not buckets of find special item.

To date the character has only found one set item (a quest reward), but the char is still low level.

 

One thing I would like to add as a wrinkle to the discussion... is that I 'found' {in S1} when players entered a server I believe everybody's magic find seed was reset. And... reset to a good number. So: I would suggest when you enter a server---get to a boss RIGHT AWAY --{of course with courteous hellos and can I take the Griffin {or whatever boss is your flavour of the month}}.

Edited August 11, 2010 by FrostElfGuard

[Antitrust 32]

Are you aware that different spots have different drops? Some of them always drop yellow items, some always drop junk, and others drop a mix of them. I hope you're not making 1000 runs on a spot that can't drop unique items...

 

I'm not sure if the above methodologies can provide any evidence for anything

[Dragon Brother 619]

Remember the "golden" rule of science - observing something affects the behaviour of the system being observed!

 

Lol...but only when you're making physical measurements on almost an atom scale is it actually not negligible...still cool though

 

Are you aware that different spots have different drops? Some of them always drop yellow items, some always drop junk, and others drop a mix of them. I hope you're not making 1000 runs on a spot that can't drop unique items...

 

I'm not sure if the above methodologies can provide any evidence for anything

 

That would be just our luck...we wanna test something and then dont even test the right spot.

[Furian67 15]

That would be just our luck...we wanna test something and then dont even test the right spot.

That would throw a wrench in the works.

[FrostElfGuard 9]

Let me say this about spots:

 

I got an unique to drop from a box in Khorum. The box outside the merchant near the portal.

 

That box can drop uniques. A known good spot.

Edited August 12, 2010 by FrostElfGuard

[rhs408 7]

Another thing noticed in psychology called confirmation bias has shown that the brain is wired (my words and I am not a psychologist) to notice information that confirms our beliefs and not notice information that contradicts. This is why you have to be careful when studying things such as randomness in video games.

 

Cheers.

 

This topic reminds me a lot of the faster kill = better drops discussion... I remember in my early playing days, in the two caves just north of Clearview, I found one set item in a chest in each cave. Extremely stoked, I immediately thought of those caves to be "lucky" caves. Since then I've probably been back to those caves 15-20 times on other characters (I often bee-lined to those caves early on thinking I'd nab a good item), haven't found a unique or set again, nor any other memorable item. It's all just luck/coincedence with what comes out of the chests, same with what drops from monsters. JMHO.

 

Are you aware that different spots have different drops? Some of them always drop yellow items, some always drop junk, and others drop a mix of them. I hope you're not making 1000 runs on a spot that can't drop unique items...

Can you give us any examples? I've yet to find any spot that always drops the same quality of item every time. The only thing I have noticed is that you will never find an armor piece (excluding jewelry) or weapon from a vase. Magical Hiding Places and Chests seem to give pretty much anything.

[Antitrust 32]

Let me say this about spots:

 

I got an unique to drop from a box in Khorum. The box outside the merchant near the portal.

 

That box can drop uniques. A known good spot.

If I remember correctly, that chest drops only a few jewelry pieces, but it does drop them often But, what does it have to do with hidden spots?

 

 

Can you give us any examples?

Here's one spot from top of my head (in the bandit camp near the river, behind a tent). There's also one near the bridge (north-west of it) in the jungle. There should also be one on the edge of the dock, easily accessible via the Na'Fian portal. All of them should always drop yellows.

Edited August 13, 2010 by Antitrust

[FrostElfGuard 9]

Let me say this about spots:

 

I got an unique to drop from a box in Khorum. The box outside the merchant near the portal.

 

That box can drop uniques. A known good spot.

If I remember correctly, that chest drops only a few jewelry pieces, but it does drop them often But, what does it have to do with hidden spots?

 

 

 

MHS and chests are different yes... I concede the point.

[rhs408 7]

Here's one spot from top of my head (in the bandit camp near the river, behind a tent). There's also one near the bridge (north-west of it) in the jungle. There should also be one on the edge of the dock, easily accessible via the Na'Fian portal. All of them should always drop yellows.

Hmm the links aren't working for me, getting a "502 Bad Gateway" error message. I think I can find the first and third of those though, I'll check them out.

[Antitrust 32]

Maplib seems to be having issues...

 

Nothing beats the old-fashioned screenshots of maps with arrows on them.

[Llama8 8]

Where are the magical hiding places/etc on the start island? I'm sure I found 2 ages ago but I can't remember where they are anymore...

[rhs408 7]

Can you give us any examples?

Here's one spot from top of my head (in the bandit camp near the river, behind a tent). There's also one near the bridge (north-west of it) in the jungle. There should also be one on the edge of the dock, easily accessible via the Na'Fian portal. All of them should always drop yellows.

2 for 2 so far at the dock in Na'Fian, got a nice run speed w/ spell intensity ring the last time. I'll assume the others always drop yellows/rares as well. I like it.

[Bluebehir 0]

I hate to throw a spanner in the works.....

but....

 

(And I'm not really sure how to word this very well...)

 

You're testing too many variables.

 

Let's pretend that you know you have a 'lucky toon' and you believe you have a 'lucky server'.

So you go and kill a specific boss - one that you know can drop blues, rares, sets and uniques, and you know they will never drop greys or whites.

 

You still don't know what the range of a drop can be.

 

Let's take a drop of:

1 unique,

1 set,

3 rares.

2 blues.

 

Is that a lucky drop or an unlucky drop?

On one hand, you got 5 potentially high calibre items out of 7.

But on the other hand, you got 7 'hits' and up maybe 10 'miss' when it came to dropping *something*

 

A second example.

 

You know you have a lucky toon. You don't know if this server is lucky or not.

You kill the same boss.

You get

3 sets

9 rares

4 blues.

 

That's a pretty good drop. You'd probably list it as lucky.

But did you get the maximum lucky DROPROLL on an otherwise UNLUCKY server?

Or did you get a reasonable roll on a lucky server with an unlucky toon?

Or a bad roll on a lucky server with a lucky toon?

 

The only way to test this, is to kill the same boss over and over and over again with the same toon on the same server, without resetting the map seed (or drop seed, or whatever other randomiser seeds Sacred uses.)

 

This is coming from someone who read similar discussions on the single player forum for Diablo 2. (And tried some for himself)

At least there, we could make constant the map seed, and the character, and the boss, and we could just test the map seed by running it 1000 times.

 

Anyway, food for thought.