156 réponses à ce sujet

[sonofabumdooda]

I use at least 12 hamsters running over laptops so their tiny feet tap those keys... only when they each write one word of 'buy PSP's now' do I spend money. That's why all my U/R's smell of nuts.

[Storm-Reaver]

The way to get Ultra-Rare, Chromatix' method: Spend all your respec cards, find\\generate interesting build, farm 1 million, go to store, loot 2-3 UR's and no cards.

[PinchLion]

Before Retaliation, I'd get URs abundantly if I'd used that particular weapon recently. Can't say anything for sure, but that doesn't sound random to me.

[sonofabumdooda]

or use hamsters
monkeys are better but more expensive
but both will definitely get you more U/R's

[SgtAmazing]

Nerds.

[SweetWilly013]

Lexa_D wrote...

SweetWilly013 wrote...

This thread's still alive?

Yes they're pseudo random numbers. No they're not perfect. But a computer processor moves way too fast for you to be able to exploit it and the numbers might as well be completely random to any human, especially at the snails pace one is able to buy packs. (processors are in the GHz range, you can buy packs at fastest maybe around the 0.1 Hz,,,,,that's a 10^10 magnitude difference, good luck).

This "evidence" all sounds very placebo.

Or maybe I'm special like that, catching al the "good" numbers with my magic fingers.
Things you said have nothing to do with the case I'm advocating, please read into it.

One person is far too small a sample size to be a justification in this case. 

I've gotten 6-7 ultra-rares in a row...all bought separately each purchase separated by 1-2 gold games. 

The sample size you'd need to establish a valid correlation is much larger than you can ever hope to generate by yourself.

Modifié par SweetWilly013, 30 novembre 2012 - 09:22 .

[Tronar]

Mozts wrote...

That was debunked back in the demo.

A software that generates random number will come up with different results based on the same thing.

Well, I kept track of my purchases and wrote down which packs I bought and what I got out of them.

Since changing my buying pattern from "one PSP right after a match" to "saving 3 million and then following the advice above)" my UR drop rate increased from about 6% for the first months to sometimes over 20%.

Or in other words: 4 to 9 URs for every 3 million credits PSP shopping spree. My average dropping rate including the bad dropping rates from the time I did it the wrong way is currently at 15,27%.

My data is stored in an Excel file and covering 17 shopping sprees starting in August. I paid 49.104.000 credits for 496 PSPs during that time period. I got 84 UR drops from those packs. So I guess it is enough to be statistically significant and not just personal delusion resulting from my mind trying to see patterns where there aren't any.

To say it again: before that I could buy 40 PSPs in succession (but distributed over single purchases after finishing a successful game)  without getting a single UR card. Since I changed my pattern, I maxed out my URs a lot faster.

I am missing only 4 more UR drops and then I am maxed out. I could have been there earlier hadn't I wasted my money until August on my desire to buy packs immediately.

You can accept that or leave it. There are many people here supporting that pattern theory from personal experience, most of them with a nearly maxed out manifest as well.

So you can either believe those smart asses having only a theoretical debunk grip on that idea or those actually gotten to the point where you want to be.
This is not some esoteric science we are talking about. Maybe Bioware inadvertently uses a glitched RNG not producing a PERFECTLY randomized shopping experience that many smartasses just assume to be in place. Wouldn't be the first bugged thing, Bioware released, or would it!?!?!

Modifié par Tronar, 01 décembre 2012 - 11:41 .

[Tronar]

deleted - double post

Modifié par Tronar, 01 décembre 2012 - 11:40 .

[BlackbirdSR-71C]

I'll try to provide some data for this as well, starting to save up credits now.

Before using the "UR-streak" method:

I buy about 2 PSPs every day, and get about one UR per week, so:

1/14 = 7,14%

Modifié par BlackbirdSR-71C, 01 décembre 2012 - 12:11 .

[darkpassenger2342]

sonofabumdooda wrote...

use hamsters
monkeys are better but more expensive
but both will definitely get you more U/R's

[soulstriker09]

I agree with OP. This has been the tip/belief/superstition I've been handing out to people who look for advice on the store.

I've netted about 5 URs in ~1.5m shopping sprees before.


But this for me no longer holds true since retaliation released.... >__<

[Lexa_D]

New kit, more data.

During the week I bought 4 PSPs, no URs, stopped after 1st purchase. Just now bought 5 more PSPs to fully unlock Quarian Soldier, no URs. 9 in a row. More to come.

Modifié par Lexa_D, 04 décembre 2012 - 10:43 .

[Lexa_D]

cowwy wrote...

Lexa_D wrote...

You're using antireligious lingo while behaving like a "true believer".

By default, the store is perfectly random, cuase that's what you're told by your betters from BW (religion). I've provided data analysis which shows clustering significantly more probable than pure random, and my hypothesis is testable, it can be proven wrong (science). I have some data support, you have none, so unless you dig up some proof of the opposite, your words are empty.

It's not my goal to "prove" anything to anyone. I've made a hypothesis, test it rigorously and post the results here. It might show that my initial assumption was wrong. If it helps some people, all the better. Can't see how it can do harm in a purely random setup.

You're right, I believe that BioWare is going to deliver us out of the hell we're currently in via the date the Mayans predicted.

Clearly your goal is to "prove" something because you keep dismissing our assertions that your data are wrong. I don't have to run any tests to know that your testing is arbitrary and wrong. Where you're wrong is that you are taking cases that prove your point, and ignoring the ones that show otherwise.

My point is that you're looking for a pattern where there is none. You stated that you provided data analysis, and now I'm doing the same. Your testing and analysis were full of bias.

What do you mean by saying "data is wrong"? Biased sample? Small sample? Selection bias? I don't understand and think neither do you. You've proven nothing.

Instead you attack my credibility, but why go that far - I have no way to prove I really get the outcomes I describe here, start at that point.

Modifié par Lexa_D, 04 décembre 2012 - 11:23 .

[Lexa_D]

Zso_Zso wrote...

OK, guys here is the REAl -- guaranteed method to get black weapons:

1. Take any gold / silver weapon into the "weapon bench" menu, i.e. changing mods
2. add the penetration mod to the weapon
3. TADA, you got a Black weapon !

Go enjoy, you can thank me later!

Also true  Close-up Raider, Reegar and Piranha are better than any UR weapon. We're just trying to do some hypothesis testing here

[Lexa_D]

CmnDwnWrkn wrote...

Lexa_D wrote...

iOnlySignIn wrote...

Reverse Gambler's Fallacy

Well, so far the coin is spinning the right (for me) way. Please provide some arguments for random.

You are the one claiming that packs cluster and URs can be obtained more frequently using your method, so the burden is on you to provide sufficient supporting evidence.  Yes, you have provided some evidence, but it isn't nearly enough to conclude anything reasonably with respect to your claim.  What you have done is the equivalent of flipping a coin 20 times, and when it comes up Heads 15 times, you conclude that a coin flip has a 75% chance of Heads.  Even if we're to take your "evidence" as 100% accurate, it just isn't enough.

Indeed, and I'm gathering more data testing the hypothesis. We therefore have null hypothesis of purely random distribution and alternative of it being not purely random (not necessarily the way I described). 

To put it short, I see anomaly and say "the distribution is not purely random", you say "no it is". We have no better way to test it than running hypothesis testing on empirical data unless we see the code, which I find unlikely. The data so far (much more than 20 flips, btw) does not support the random theory, it shows clustering which makes the approach I propose more probable to yield URs than just buying stuff whenever you have money.

[Lexa_D]

dbushik wrote...

Lexa_D wrote...
...
Doing that on my own, although with larger sums usually. Anyone is welcome to post their results here if they wish.

Okay, but wanted to follow some standard method so any data would be fully compatible (gathered with an identical method).  The initial discription of your method is loose.

Ok, to clarify what I do (edited excerpt from OP):

 hoard the cash, occasionally[once in a game day, which is 3-4 times a week for me] buy a Premium Spectre Pack. If you get nothing [no URs], move along [exit shop]. If you get Black, start spending. No further occurences in next 3 packs - might have been a fluke on the first one, move along [exit shop]. Otherwise - spending spree [spend all cash].

[12323432543]

I don't usually stash and buy after I have enough and at the moment I usually get 1 UR every day. So that would be about 4 to 5 packs. Sometimes I get more, sometimes I get less, but it evens out.

I've noticed however that there are times where it can be streaky. Bought a pack containing a wraith 9. Didnt have enough credits since I dont store up, went for another game and the next pack gave me a double black card.

[Lexa_D]

SweetWilly013 wrote...

Lexa_D wrote...

SweetWilly013 wrote...

This thread's still alive?

Yes they're pseudo random numbers. No they're not perfect. But a computer processor moves way too fast for you to be able to exploit it and the numbers might as well be completely random to any human, especially at the snails pace one is able to buy packs. (processors are in the GHz range, you can buy packs at fastest maybe around the 0.1 Hz,,,,,that's a 10^10 magnitude difference, good luck).

This "evidence" all sounds very placebo.

Or maybe I'm special like that, catching al the "good" numbers with my magic fingers.
Things you said have nothing to do with the case I'm advocating, please read into it.

One person is far too small a sample size to be a justification in this case. 

I've gotten 6-7 ultra-rares in a row...all bought separately each purchase separated by 1-2 gold games. 

The sample size you'd need to establish a valid correlation is much larger than you can ever hope to generate by yourself.

Well, actually, no, as I'm testing the general law, I don't need large sample. Something like 100 data points will be all too sufficient. I've announced the law before I started most of the testing, and I'm getting meanignful results, as Bayes is my witness. Feel the difference with a case of a particularly "lucky" person in a large sample of store-grinders who got (and noticed) the seemingly clustered sequence of purely random variable outcomes. If he starts testing the "rule" after his claim, the data will not follow the rule. Not the case here so far.

[Lexa_D]

12323432543 wrote...

I don't usually stash and buy after I have enough and at the moment I usually get 1 UR every day. So that would be about 4 to 5 packs. Sometimes I get more, sometimes I get less, but it evens out.

I've noticed however that there are times where it can be streaky. Bought a pack containing a wraith 9. Didnt have enough credits since I dont store up, went for another game and the next pack gave me a double black card.

Have some gut memory of clustering by the day too, hence the rule of buying once a game day, but no data recorded to prove, so won't make any claims.

[Lexa_D]

Tronar wrote...

Mozts wrote...

That was debunked back in the demo.

A software that generates random number will come up with different results based on the same thing.

Well, I kept track of my purchases and wrote down which packs I bought and what I got out of them.

Since changing my buying pattern from "one PSP right after a match" to "saving 3 million and then following the advice above)" my UR drop rate increased from about 6% for the first months to sometimes over 20%.

Or in other words: 4 to 9 URs for every 3 million credits PSP shopping spree. My average dropping rate including the bad dropping rates from the time I did it the wrong way is currently at 15,27%.

My data is stored in an Excel file and covering 17 shopping sprees starting in August. I paid 49.104.000 credits for 496 PSPs during that time period. I got 84 UR drops from those packs. So I guess it is enough to be statistically significant and not just personal delusion resulting from my mind trying to see patterns where there aren't any.

To say it again: before that I could buy 40 PSPs in succession (but distributed over single purchases after finishing a successful game)  without getting a single UR card. Since I changed my pattern, I maxed out my URs a lot faster.

I am missing only 4 more UR drops and then I am maxed out. I could have been there earlier hadn't I wasted my money until August on my desire to buy packs immediately.

You can accept that or leave it. There are many people here supporting that pattern theory from personal experience, most of them with a nearly maxed out manifest as well.

So you can either believe those smart asses having only a theoretical debunk grip on that idea or those actually gotten to the point where you want to be.
This is not some esoteric science we are talking about. Maybe Bioware inadvertently uses a glitched RNG not producing a PERFECTLY randomized shopping experience that many smartasses just assume to be in place. Wouldn't be the first bugged thing, Bioware released, or would it!?!?!

Thank you, Tronar, for the very valuable input.

Modifié par Lexa_D, 04 décembre 2012 - 11:22 .

[Lexa_D]

New data

5 daily purchases of a single PSP w/o URs in a row, then leavung shop.

Today got Talon in1st, bought 2 more - no URs. However, decided to go further (dry season weighing on me) to see what happens. Another Talon at 7th pack, then none until I called it a day. 15 packs bought total, not stopping after 3 packs as the strategy prescribes caused me to spend 1.2mil to get a single UR - a whooping 4.2% drop rate. Altogether the sequence of purchases since last "lucky streak" is 21 pack with 2 drops, or 4.7% vs clustered 31% and 16.7%. Not a dry spell, but a quite sub-par outcome nevertheless.

Update: 5 empty PSPs while unlocking Saboteur, then PPR folowed by 3 empty PSPs - droped according to trategy.

Modifié par Lexa_D, 13 décembre 2012 - 09:20 .

[YachiruChin]

If your theory has only 1% probability of being true, it's still better trying than not. xD

I have actually kept track of my last 240 PSPs and surprisingly I do have some UR streaks and a lot of pretty nasty null sequences. I usually save about 1~2 mi credits before spending it all too.

Too bad I haven't been playing a lot lately.. but I'll keep that in mind for future tests. xD

[Lexa_D]

update: 22 single PSPs and one set of 5 PSPs starting with Talon - only 10 cards to go, risk pays off more. Result (with edit of previous post): 2 Black Cards in 32 PSPs - another piece of data against random UR occurence.

The big disadvantage of this strategy - might take long between "lucky" strikes indeed.

[Lexa_D]

ok, so I returned from skiing in Alps on Sunday (or was ot Monday already) and hit the gold vein - 2 URs in first 3 PSPs. Spent all my 4mn I bagged up to this point.

First 15 PSPs - 7 URs, a 23.3% drop rate, statisticaly significant above average
Other 25 PSPs - 2 URs, 4% drop rate, a bit below average.
It could be that I had two clusters in 1 purchase, or just one with peculiar distribution, no way to tell from a single data point. Even in the latter case overall drop rate was 11.25%, significantly above 6.9% drop rate for Artificial-Ignorance (thanx for the link, Shampoohorn).

I will update the OP soon to reflect the outcome of this rather extended test. With only 3 Typhoon ranks left to unlock, I will no longer follow the strategy, so happy hunting to everybody!

Modifié par Lexa_D, 18 janvier 2013 - 01:00 .