63 réponses à ce sujet

[UK Wildcat]

I'd suggest not worrying whether it is completely RNG or not (look at the post describing how it could be completely RNG) and focusing on the most efficient ways you can actually use your credits to purchase the items you want/don't want in order.

Example:

I want a blackwidow, and currently don't have one. Specter packs are my most promising chance (perhaps premium veteran packs as well) to unlock one.

Assuming I am given a random range of results or a random number specifying which upgrade I am capable of, I can limit the number of possibilities by completely upgrading the *otherwise undesirable items that would appear* such as mods, crappy shotguns etc.

One good way to eliminate these items is to focus solely on purchasing premium veteran packs, veteran packs, recruit packs until I have all of the mods (done), a ton of weapons at Rank X (done). This ensures that in my specter packs / premium veteran packs, I won't be getting things I don't want, and will lower the pool from which I can receive items.

[LostJ]

MartialArtsSurfer wrote...

In consumer computer programs, the RNG really isn't random because it's based off of a seed value, usually the computer clock. 

The game uses a Psuedo-Random Number Generator (RNG) based off of a mathematical formula like any other computer program, with a seed value usually taken from the computer's clock (not sure if the server's clock or your computer clock) so if you want to avoid repeats, buy stuff at a different time.


It seems people get more repeats if they buy lots of packs at the same time instead of 1 & then later at a different time.

Stuff like the same characters, shotguns, widows at the same time 4 in a row ... they mostly were bought at the same time, especially since people play at the same times...

so if you don't like what you get in a pack buying it at 6 PM, buy your next pack at a different time like 11 PM or 9 AM

On the other hand, if you like what you get, buy them at the same time

It seems true for most of the people I know... true for you also?

I did get 2 Scorpions and 2 Claymore on a row when buying Specter packs at the same time. 

Makes sense to me.

[Phatose]

Threads like this only serve to show 2 things: 1, that humans are terrible at establishing cause and effect where chance is involved, and 2. Typical games know zip about programming.

[Ravenear]

MassEffectOnMe wrote...

Ravenear wrote...

My idea to get better items is:
1. buy Recruit Pack until you will get all basic weapons on X level - on Veteran Pack's & Spectre Pack's you will not get crapy unlocks of basic weapons; it's fast step.
2, buy Veteran Pack (Premium Veteran Pack) until you will get all uncommon weapons on X level & weapon modifications on V level - right now im on this step i have unlocked all uncommon weapons on X level and trying to unlock all uncommon weapon modifications on V level
This is my way

This idea is so crazy...it just might work! I'll try this out.

I think its working becouse most of the time i have in Spectre Pack's 1 uncommon weapon modification and 2 rare weapons. So i think its good drop

[vivanto]

Items are rolled server-side. /thread

[Fieryeel]

Ravenear wrote...

My idea to get better items is:
1. buy Recruit Pack until you will get all basic weapons on X level - on Veteran Pack's & Spectre Pack's you will not get crapy unlocks of basic weapons; it's fast step.
2, buy Veteran Pack (Premium Veteran Pack) until you will get all uncommon weapons on X level & weapon modifications on V level - right now im on this step i have unlocked all uncommon weapons on X level and trying to unlock all uncommon weapon modifications on V level
This is my way

Bumping this thread for more input to see if your idea works.

[Spellbound83]

Ravenear's post...sorry, board won't let me quote for some reason.

I started doing this because I spent 60k and got garbage in my packs. Use the google doc to keep track of what items you already have, and don't buy the next tier until you have everything from the previous one. The google doc is made by narida_L I think...search his name to find it, I can't link it now.

Do this and never again will you get a predator VI in your 60k spectre pack. However, it does nothing to solve the problem of getting character training cards after you've already maxed out appearances.

[Maylson]

I bought 3 spectre packs at once (had 180k saved up) and I got 2x Arc Pistols on the last 2 packages... Maybe what you're saying makes sense, I don't know, but good find

[MisterVQ]

Fieryeel wrote...

Ravenear wrote...

My idea to get better items is:
1. buy Recruit Pack until you will get all basic weapons on X level - on Veteran Pack's & Spectre Pack's you will not get crapy unlocks of basic weapons; it's fast step.
2, buy Veteran Pack (Premium Veteran Pack) until you will get all uncommon weapons on X level & weapon modifications on V level - right now im on this step i have unlocked all uncommon weapons on X level and trying to unlock all uncommon weapon modifications on V level
This is my way

Bumping this thread for more input to see if your idea works.

It does indeed work however once you max uncommons, you are more likely to unlock gold chracter cards than rare/N7 weapons.

Modifié par MisterVQ, 29 mars 2012 - 06:28 .

[JohnDoe]

Ayeba wrote...

You should be careful with your assumptions there. If I get two identical character or weapon upgrades in a row, that does not necessarily mean it wasn't random. Random doesn't mean getting the same item two (or even five) times in a row won't happen. In fact, it will, and the human mind will remember that much more clearly than all the streaks where you didn't get the same item multiple times.

Also consider that there seem to be one race for each class on Uncommon unlocks. So if you do get two characters in a row, it might be something like 1/6th chance of getting the same character twice.

Bottom line, I don't think people should pay too much attention to it unless it happens excessively often (not just once in a while) unless people start collecting statistics.

pretty sure its not asumptions.. I once bought a few spectre packs in a row.. and I GOT ALL SHOTGUNS.. again and again..
then started to get upgrades for them.. and I play roughtly in the same hours.

then the next day I bought another one.. got a falcon..
so if you compare.. it seems that there must be something like what he asumes..

Modifié par JohnDoe, 29 mars 2012 - 06:35 .

[DeckardWasAReplicant]

All i want is an asari adept. So should I save money for spectre packs. Or do I have a better chance with veteran packs because im able to buy more?

I have bought a lot of spectre packs woth no asari adept. So I was thinking of going with veteran packs. Good idea? Or just stick with spectre?

[Father Alvito]

Going with lower level packs won't help you get a Gold character card. If you wanted the Asari Vanguard, then spamming Veteran packs will eventually start giving you silver character cards most of the time.

You seem to get more rares per credit invested with Spectre packs than Veteran packs, so I'd keep fishing there. You also have that vanishingly small chance to get an N7 weapon with a Spectre pack, which is also attractive.

[SeaJayX]

Got 2 N7 weapons 20 minutes apart yesterday.

[Father Alvito]

SeaJayX wrote...

Got 2 N7 weapons 20 minutes apart yesterday.

Don't we all hate you.  I haven't opened one since the first week, and I've opened at least 100 Spectre packs since then.

[AndanteInBlue]

You guys realize random number generators are likely seeded once per server restart. "Seeding" generates a position in a "list" of numbers that is longer than all the number ever written down in the history of the Earth combined. Consecutive calls for random numbers on the server return the _next_ number in the list, so even if they are "closely spaced" together , they're still the next number in the list, not the same number again.

Also, computers operate on the scale of nanoseconds, which is several orders of magnitude faster than you can possibly click the mouse, much less purchase from the laggy store screen. Your "purchasing several items in a row" or "buying at the same time", might as well be eons apart as far as the server is concerned.

[vivanto]

AndanteInBlue wrote...

You guys realize random number generators are likely seeded once per server restart.

Don't bother, some of us already tried to explain, users don't even realize the drops are generated server-side. Apparently people prefer to believe in magic rather than science, I'm starting to see why Bioware went with the space-magic ending.

[Father Alvito]

vivanto wrote...

 Apparently people prefer to believe in magic rather than science, I'm starting to see why Bioware went with the space-magic ending.

We know that this is true because this thread is on the second page and people are still blathering rather than presenting data that would falsify the OP's hypothesis.  You can endlessly speculate about what algorithms the programmers used and how closely they would approximate truly random draws.  Or you can work together to compile a large enough sample that would permit inferences about whether or not the RNG behaves differently when Spectre packs are bought sequentially or separately.

Your call.

Modifié par Father Alvito, 29 mars 2012 - 07:29 .

[vivanto]

Father Alvito wrote...
We know that this is true because [..]

You mean you know this because you didn't graduate in computer engineering?  

Results you want?, I tought the thread with the guy spending lotsa $$ on specter packs already did that, but I'll give in. Last week I accumulated ~1 mil creds with the intent of buying 10 recruit, 10 premium and 10 sepcter packs. Since from the premium packs it appeared that I've still got some unlocking to do, I sticked with buying just them. Regardless, they are stuff generated randomly in rapid succession by Bioware's magic box of wonders.

You might try to do the same, so you'll understand how annoying it is to just buy packs nonstop for 20 minutes and write them all down, hence why I just decided to **** it after the 13th and bought the rest without logging.

Last two items in premiums:
Raptor, Tempest
Asari Vanguard, Quarian Infiltrator
Viper, Locust
Tempest, Eviscerator
Asari Vanguard, Eviscerator
Mattock, Drell Adept
Salarian Engineer, Scimitar
Scimitar, Locust
Turian Sentinel, Phalanx
Viper, Tempest
Drell Adept, Incisor
Eviscerator, Asari Vanguard
Scimitar, Quarian Infiltrator

Total equipments:
medigel: 9
ops pack: 9
rockets: 12
ammo packs: 9

Seems pretty random. Equipment count is especially representative as you can easily draw quantitative conclusions based on them.

But sure, you can believe in space jesus if you prefer, we have freedom of belief father.

[LikeReD]

I unlocked all modes and weapons except N7s and four human character modification. The method that I tried was starting with recruit pack then veteran and spectre. And for the result, it wasn't necessary. As soon as you can clear gold challenge, then I suggest you go strait for the spectre pack. Adept, Sentinel and Engineer don't need fancy weapons so start with those classes.

If I start MP from the beginning, I would suggest go from spectre pack to veteran and recruit.

Modifié par LikeReD, 29 mars 2012 - 10:04 .

[MartialArtsSurfer]

vivanto wrote...

Father Alvito wrote...
We know that this is true because [..]

You mean you know this because you didn't graduate in computer engineering?  

Results you want?, I tought the thread with the guy spending lotsa $$ on specter packs already did that, but I'll give in. Last week I accumulated ~1 mil creds with the intent of buying 10 recruit, 10 premium and 10 sepcter packs. Since from the premium packs it appeared that I've still got some unlocking to do, I sticked with buying just them. Regardless, they are stuff generated randomly in rapid succession by Bioware's magic box of wonders.

You might try to do the same, so you'll understand how annoying it is to just buy packs nonstop for 20 minutes and write them all down, hence why I just decided to **** it after the 13th and bought the rest without logging.

Last two items in premiums:
Raptor, Tempest
Asari Vanguard, Quarian Infiltrator
Viper, Locust
Tempest, Eviscerator
Asari Vanguard, Eviscerator
Mattock, Drell Adept
Salarian Engineer, Scimitar
Scimitar, Locust
Turian Sentinel, Phalanx
Viper, Tempest
Drell Adept, Incisor
Eviscerator, Asari Vanguard
Scimitar, Quarian Infiltrator

Total equipments:
medigel: 9
ops pack: 9
rockets: 12
ammo packs: 9

Seems pretty random. Equipment count is especially representative as you can easily draw quantitative conclusions based on them.

But sure, you can believe in space jesus if you prefer, we have freedom of belief father.

I go by EMPIRICAL evidence, not "it's supposed to work this way according to what I read in a book"

As a someone who's scientifically trained (UCBerkeley) who's a software developer/Oracle admin for 10+ years, I know enough when someone's programmed faulty RNG & how even the best RNG algorhytmns can be buggy  & aren't really random but pseudo random

--spend enough years working in the real world & you realize that there's never any program that's not buggy & book examples that work perfectly bug-free in some ideal dream world exists only in the classroom ..

"ship first, debug later & patch later"

Let's take your  packs you just bought in a row that you just posted:

Out of the hundreds of different weapons, char cards, weapon mods, what are the odds & do you really think the RNG algorythmn is s really random & working perfectly?

I gathered your data & sorted & organized them for analysis:

You got REPEATS OF:

Eviscerator
Eviscerator
Eviscerator  (total 3 repeats)

Tempest  
Tempest    
Tempest    (total 3 repeats)
  
Scimitar,
Scimitar
Scimitar, (total 3 repeats)

Asari Vanguard,
Asari Vanguard
Asari Vanguard, (total 3 repeats)

4 items repeated 3 times

Quarian Infiltrator
Quarian Infiltrator (total 2 repeats)

Drell Adept
Drell Adept (total 2 repeats)

Locust
Locust (total 2 repeats)

3 items repeated 2 times

No repeats:

Raptor,
Viper,
Mattock,
Salarian Engineer,
Turian Sentinel,
Phalanx
Viper,
Incisor

The vast majority of what you got were repeated 3 times & then 2 times... out of the hundreds of possible weapons, characters, & weapon mods

The evidence suggests the RNG algorhytmn THAT'S SERVER SIDE isn't working as intended

(gee, computer program algorythmns not working as intended & buggy--  wow, that should never happen since we all know programs are never buggy & work just as well as they're supposed to like you learned in "Intro to Programming/Engineering" class like the book says)

Mass Effect 3 has no bugs so therefore the Mass Effect 3 RNGalgorythmn running server-side must have no bugs either , LOL

You said it looked random to you? Really, since you just listed it without any organization but a jumbled mess & no analysis?

It seems your own data contradicts your own preconceived conclusions & you just went by ideology & faith in what you learned in your first intro to computer engineering class

I DON'T CARE EITHER WAY WHETHER buying packs all at once or separately works for getting packs (I have  NO PERSONAL STAKE in whether it's buggy or not
--I want to find out if Mass Effect's RNG is buggy
(which case it should be patched & fixed)
or
if not & test it, & if it is buggy & not fixed,
then test & find the best method that will best get good unlocks instead of repeats (or get repeats if you get something good)

But we can't do that if people are too busy trying to defend Bioware's RNG honor & how the RNG program Bioware is using must be perfectly bug-free because we all  know that game companies would NEVER have buggy code nor SHIP A PRODUCT or GAME that wasn't BUG-FREE, LOL

Modifié par MartialArtsSurfer, 29 mars 2012 - 10:38 .

[RogenTAU]

I found that the repeating of content in the packs can be sometimes useful. I leveled some silver weapons to X while i still have majority os silver guns around level V. I got Pahestron X, in one day from level II. Also leveled phalanx, vindicator and mattock.

So my advice is to save credits, until you have a larger sum, then buy few packs, (1-3). If you are happy with the content, continue right away with more packs. The chance that you get desired weapon is much highier. If you don;t recieve second copy of the desired weapon in reasonable amount of packs, stop then, do some missins and again try the test seed of packs, and repeat the entire proces.

By this you can get desired weapon quite fast }provided you have enough funds.

And it work with common uncommon items, and it seems to work even with rare items.

[MartialArtsSurfer]

Phatose wrote...

Threads like this only serve to show 2 things: 1, that humans are terrible at establishing cause and effect where chance is involved, and 2. Typical games know zip about programming.

It's not superstition, it's the likelihood  that the RNG code is BUGGY (server-side too!) and that's because 10+ years as a business software developer/Oracle admin lets me know that most code is BUGGY & most game companies ship buggy code:

"ship first! debug & patch later!" ..including the server-code

Modifié par MartialArtsSurfer, 29 mars 2012 - 10:48 .

[MartialArtsSurfer]

RogenTAU wrote...

I found that the repeating of content in the packs can be sometimes useful. I leveled some silver weapons to X while i still have majority os silver guns around level V. I got Pahestron X, in one day from level II. Also leveled phalanx, vindicator and mattock.

So my advice is to save credits, until you have a larger sum, then buy few packs, (1-3). If you are happy with the content, continue right away with more packs. The chance that you get desired weapon is much highier. If you don;t recieve second copy of the desired weapon in reasonable amount of packs, stop then, do some missins and again try the test seed of packs, and repeat the entire proces.

By this you can get desired weapon quite fast }provided you have enough funds.

And it work with common uncommon items, and it seems to work even with rare items.

Exactly.. this strategy is how many people got their Widow X or Carnifex X.

[vivanto]

MartialArtsSurfer wrote...

I go by EMPIRICAL evidence, not "it's supposed to work this way according to what I read in a book"

That only works with a large enough sample size, obviously mine, or any sample that a single user posts can't be compared to the number of permutations given based on every items that exists in the packs. You're disregarding a lot of other factors while exagerating the ones that support your point.

MartialArtsSurfer wrote...

As a someone who's scientifically trained (UCBerkeley) who's a software developer/Oracle admin for 10+ years, I know enough when someone's programmed faulty RNG & how even the best RNG algorhytmns can be buggy  & aren't really random but pseudo random

Please list the exact course, not everyone lives in the USA and simply looking up a university can (and does) give hundreds of courses. But yeah, I am working on the field, Nokia Siemens Networks right now, prevously Cisco and even some local Oracle distributor. If you are a developer, then you should know that every random generator is pseudo random, not just the buggy ones. It is by design, that doesn't matter since the level of determinism is completely different than the level of application.

MartialArtsSurfer wrote...

Out of the hundreds of different weapons, char cards, weapon mods, what are the odds & do you really think the RNG algorythmn is s really random & working perfectly?

There are 34 possible weapons, not hundreds. Filter them by category of the pack (don't know the figure, I'll be generous and give uncommon weapons 50%), that leaves 17. You forgot that I may already have a couple of them already maxed. That leaves 10-15% probabilty of getting any single weapon, which is a LOT and still less than the actual numbers would be, especially since between the first and the last pack this chance just goes upwards.

Some user posted that weapons distributed that you already have at lvl X are replaced with a random character card of the same rarity. Don't know if it's true or what his source was, but might take into consideration when drawing conclusions based on character cards (hence I didn't post my list of 10 consecutive recruit packs).

MartialArtsSurfer wrote...

The evidence suggests the RNG algorhytmn THAT'S SERVER SIDE isn't working as intended

And what makes you think how it is intended? The items are obviously weighted, that's not a secret nor a design flaw. What people don't understand that it cannot be statistically weighted, especially since you can simply disable user feedback. I might do some more experiments on this subject on the weekend tho, I plan on tracing some packets to determine my ping anyways.

MartialArtsSurfer wrote...


(gee, computer program algorythmns not working as intended & buggy--  wow, that should never happen since we all know programs are never buggy & work just as well as they're supposed to like you learned in "Intro to Programming/Engineering" class like the book says)[/b]

Mass Effect 3 has no bugs so therefore the Mass Effect 3 RNGalgorythmn running server-side must have no bugs either , LOL

 
Not sure if sarcastic or simply troll. We all know there are bugs and design flaws, instantiating a random generator however can hardly be classified as a bug, or did you work on the 10+ year old and proven libraries the developers used?

[vivanto]

MartialArtsSurfer wrote...

Phatose wrote...

Threads like this only serve to show 2 things: 1, that humans are terrible at establishing cause and effect where chance is involved, and 2. Typical games know zip about programming.

It's not superstition, it's the likelihood  that the RNG code is BUGGY (server-side too!) and that's because 10+ years as a business software developer/Oracle admin lets me know that most code is BUGGY & most game companies ship buggy code:

"ship first! debug & patch later!" ..including the server-code

Yeah, and we all know that Oracle was founded by university dropouts who didn't even know what a regular expression is.