Jump to content

Reverse Engineering is not 20%


Darth_Sweets

Recommended Posts

  • Replies 160
  • Created
  • Last Reply

Top Posters In This Topic

What's the point? People will say the sample size is not big enough no matter how many I do.

 

You are right that even a thousand attempts is considered a small sample, but it is far more significant than the typical 20-50 most forum-goers use to "prove" their point. So if you do a thousand attempts and actually post your data others will listen. But if Khevar and I (and others) are right, you will find that the RE RNG is in fact working just fine.

Link to comment
Share on other sites

Do you REALLY think that if the system were actually broken that BioWare would not fix it? If so, you better get your tinfoil hat out :eek:

Ok, Ive tried to stay out of this but I cant anymore. You're making this statement about EA??? (Bioware isnt even a company any more). The company with one of the worse records for customer service in the world. Not to mention their being voted the worse company in America 2 years in a row. You really believe they would be honest and tell us it is broken or just deny it like all the other bugs they wont acknowledge. Why fix this one thing out of all the broken code in this game.

Link to comment
Share on other sites

Ok, Ive tried to stay out of this but I cant anymore. You're making this statement about EA??? (Bioware isnt even a company any more). The company with one of the worse records for customer service in the world. Not to mention their being voted the worse company in America 2 years in a row. You really believe they would be honest and tell us it is broken or just deny it like all the other bugs they wont acknowledge. Why fix this one thing out of all the broken code in this game.

 

Nice selective quoting. Did you actually bother to read my whole post? In short, random =/= even and no one player's sample is going to be accurate, but Bioware...oh wait I'm sorry EA has the data we don't.

Edited by psandak
Link to comment
Share on other sites

Found this topic kind of interesting, so I wrote a (very simple) little console app to run some RNGs over x-number of samples within a given range to see average hits and determine how spot-on RNGs would be.

It's pretty unlucky to hit anywhere below 15%, even with only 300 samples. But any good number of samples (1K to 10K for example) with a default RNG gets pretty spot on the 20%. Only small sample numbers and a tight range of random values to sample from can cause those extremes below 15% or above 25%.

 

NOTE: I have occasionally spotted absurd gaps though. Once had a gap of 69 fails before I got a hit, but it's rare. Still, gaps of 20-40 are quite common.

 

One thing that did bother me during all these discussions here though (and I'm not sure if this is applicable), but didn't anybody consider the fact that crafting with companions could influence the outcome, as some companions are more likely to succeed with certain skills than others? Resulting quality of an item is something that can greatly influence the numbers and is something one simply can't test with just RNGing, without knowing what the real code does to factor in those companion stats. Or doesn't the quality of a crafted item not influence the chances during RE? (maybe I'm just misunderstanding those things)

If it does, wouldn't it be a matter of other factors influencing the formula, rather than of RNG?

 

If the quality of an item is irrelevant, the test at the very least shows how larger sample numbers and ranges indicate solid RNG average hit results.

 

PS: Please excuse if I get terminologies wrong. Not really a math guy and most of the time don't really care about the inner workings of stats that much,... usually.

 

Quick video capture of a few runs: watch/download

Edited by nGAGE
Link to comment
Share on other sites

Companions have no impact on reverse engineering other than creating a extra unit through a crit.

 

When trying to learn schematics, getting two of something (item modifications, and biochem consumables) creates a scenario where the chance of getting a schematic is now 36% rather than 20% (20%^2=36%), because you now have two attempts rather than one. Critically succeeding at crafting a piece of gear - that results in an augment slot being added to that gear - has no impact on RE. In fact there is a thread suggesting that augmented items should have a higher RE chance. I cannot say I disagree with the suggestion but the implementation is probably not nearly as simple as the OP of that thread thinks/suggests.

Link to comment
Share on other sites

  • 2 weeks later...
Found this topic kind of interesting, so I wrote a (very simple) little console app to run some RNGs over x-number of samples within a given range to see average hits and determine how spot-on RNGs would be.

It's pretty unlucky to hit anywhere below 15%, even with only 300 samples. But any good number of samples (1K to 10K for example) with a default RNG gets pretty spot on the 20%. Only small sample numbers and a tight range of random values to sample from can cause those extremes below 15% or above 25%.

 

NOTE: I have occasionally spotted absurd gaps though. Once had a gap of 69 fails before I got a hit, but it's rare. Still, gaps of 20-40 are quite common.

 

One thing that did bother me during all these discussions here though (and I'm not sure if this is applicable), but didn't anybody consider the fact that crafting with companions could influence the outcome, as some companions are more likely to succeed with certain skills than others? Resulting quality of an item is something that can greatly influence the numbers and is something one simply can't test with just RNGing, without knowing what the real code does to factor in those companion stats. Or doesn't the quality of a crafted item not influence the chances during RE? (maybe I'm just misunderstanding those things)

If it does, wouldn't it be a matter of other factors influencing the formula, rather than of RNG?

 

If the quality of an item is irrelevant, the test at the very least shows how larger sample numbers and ranges indicate solid RNG average hit results.

 

PS: Please excuse if I get terminologies wrong. Not really a math guy and most of the time don't really care about the inner workings of stats that much,... usually.

 

Quick video capture of a few runs: watch/download

Cool video. What you're not counting are the win streaks. For example look at the win stream at 2:44 at 9796. You hit 5 in a row and 8 out of 10. Curious, anyone here ever pop 5 in a row REs? How about 8 out of 10?

Link to comment
Share on other sites

Cool video. What you're not counting are the win streaks. For example look at the win stream at 2:44 at 9796. You hit 5 in a row and 8 out of 10. Curious, anyone here ever pop 5 in a row REs? How about 8 out of 10?

 

I've had many 3 of 5 and 2 in a row streaks.

 

But there is another human factor to consider in "win" streaks: getting a schematic but one that is not desired e.g. all the Presence purples, or Redoubt Cunning blues, or worse a player is looking for a VERY specific purple and not getting it until getting the other three or four. Many - if not most - players in those situations will emotionally consider those successes as failures.

Link to comment
Share on other sites

I've had many 3 of 5 and 2 in a row streaks.

 

But there is another human factor to consider in "win" streaks: getting a schematic but one that is not desired e.g. all the Presence purples, or Redoubt Cunning blues, or worse a player is looking for a VERY specific purple and not getting it until getting the other three or four. Many - if not most - players in those situations will emotionally consider those successes as failures.

 

Though I don't consider learning any schematic as a Fail. I Agree w/ the Point Psandak is making. SolQitta my Healer is also my Armormech. I craft all of the gear that she wears so I want very specific Purples. Because I am looking for Expert/Vehemence or Endowment/Supremacy for her to wear it can seem sometimes that RE'ing is being less productive than is actually happening. In the end all you need is patience (& have a love of WZs or a 55 toon to run dailes & give you a never ending supply of credits for which to keep getting supplies to make the items to RE till you get what you are looking for). :rak_09:

Link to comment
Share on other sites

My point is if people are missing 60 in a row, someone must also be getting 10 in a row. Never happens.

Let me see if I follow your logic. You're saying:

 

1. The chance to fail 60 times in a row is the same chance to succeed 10 times in a row.

2. People have reported failing 60 times in a row.

3. Nobody has reported succeeding 10 times in a row.

4.Therefore, the RE chance is broken.

 

Is that right?

Edited by Khevar
Link to comment
Share on other sites

My point is if people are missing 60 in a row, someone must also be getting 10 in a row. Never happens.

 

These two things are not even remotely equal.

 

Let's say you are starting from scratch, with no prior history. I'm also assuming we are RE'ing blues to purples, with a 10% chance of success.

 

You start to RE. Your chance of failing 60 times in a row, before you begin, is about 0.18%. Not very likely, but with all the players in the game we can safely assume that this happens multiple times a day. To the person is it happening to, it feels incredibly frustrating. I don't know that I have ever failed 60 times, but I know for sure I failed 50, and it was painful.

 

The counterpart to failing 60 times in a row is succeeding 3 times in a row, which is almost half as likely at 0.10%. I don't think I've ever had this happen, but it is again likely that every day, a few lucky souls out there hit this.

 

Even if you are doing RE's with a 20% success rate, it is ten times more likely that you will fail 60 times in a row than succeed 10 times in a row. The odds of these are very low (1 in a million for 60 failures, one in 10 million for 10 successes), but the failure is on the order of something that might be happening once a month or every few months.

Link to comment
Share on other sites

My point is if people are missing 60 in a row, someone must also be getting 10 in a row. Never happens.

 

I got five in a row two days ago, does that count :)

 

Another thread was started on this subject with the OP having a 1 for 26 bad luck streak (0 for 12, 1 success, then 0 for 13). As I was planning to level a crafting skill anyway, I decided to track my RE rate. You can view my data: https://drive.google.com/file/d/0By58u0ps4f7gcHdtVy1OZlJDeU0/edit?usp=sharing

 

The first day, I got to 410 skill in armstech and had an RE rate of ~24% over 239 attempts, with a long streak of 14 failures on one item. Day two I got to the armstech skill to 450 but my RE rate was a lot worse: ~14% over 64 attempts with a 2 in 42 streak over four items and a single item streak of 1 in 25 (with the final attempt being a success). On day three I am still working on one schematic from day two (active streak of 0 in 16), but am slightly better overall than day two: ~16% over 38 attempts.

 

But in my overall RE rate is still above expected at 21.41%: 73 successes in 341 attempts. Mostly greens to blue, but some blue to purple.

Link to comment
Share on other sites

OP needs to learn how basic probability works.

 

2/10 chance =/= guaranteed 2 successes in 10 trials.

 

That is all.

While there are people in this thread that have basic confusions on the subject of probability, the OP (Darth_Sweets) is certainly not one of them.

 

He used a type of statistical analysis called a confidence interval, and performed sophisticated math analyzing his raw data. The only disagreement I had with his analysis was the sample size, not his methods or understanding of probability.

 

Either you didn't actually read the OP, or you didn't understand what he was saying.

Link to comment
Share on other sites

While there are people in this thread that have basic confusions on the subject of probability, the OP (Darth_Sweets) is certainly not one of them.

 

He used a type of statistical analysis called a confidence interval, and performed sophisticated math analyzing his raw data. The only disagreement I had with his analysis was the sample size, not his methods or understanding of probability.

 

Either you didn't actually read the OP, or you didn't understand what he was saying.

 

While I agree with your general point (perhaps the poster you replied to was referring to another contributor in this thread) - the bolded is a pretty huge problem. OP stated it's "almost impossible that the 20 percent is the true rate of getting a new plan" which is patently ridiculous based on the extremely limited data. I don't care how sophisticated his methods are (I agree he does seem to have a good grasp of the tools involved), but sample size is such a critical factor here that I can't really blame anyone for assuming OP has some gaps in his understanding of probability.

 

No offense @OP - go ahead and run another few thousand trials, and re-run your analysis. Then we'll talk. :)

Link to comment
Share on other sites

I did, and i got that re-- confidence interval (although I wonder how strictly relevant that is given his much-too-small sample size? Yeah, my Stats-training was a loong time ago now, so maybe I'm forgetting something.)

 

And yeah, "three-sigma rule," as someone else mentioned, but again, how valid is even that given the extremely small sample size in this case?

 

More --much more-- data needed.

 

E:

 

To continue from the poster above me: OP should feel free to run several thousand more trials, and get all their friends to do so, and get all their friends to do so, then post those results.

Edited by midianlord
Link to comment
Share on other sites

I did, and i got that re-- confidence interval (although I wonder how strictly relevant that is given his much-too-small sample size? Yeah, my Stats-training was a loong time ago now, so maybe I'm forgetting something.)

The basic problem in his argument was this statement:

Actually the confidence interval test is valid with almost any sample size since, the tightness of the bounds will change with the number of samples used.
Edited by Khevar
Link to comment
Share on other sites

Hey - I ended up getting 0 for about 20 just now. Must mean that RE is broken.

 

Oh wait... it could also mean I was multi-tasking (posting on Facebook, reading the forums, and playing SWTOR on a 2nd screen), not paying attention whatsoever, and didn't notice the "no research available." D'oh! Luckily it wasn't anything expensive.

Link to comment
Share on other sites

Far as I can tell it is just a 20% of something happening. Not a 20% chance of happening, there for after 5 times it should = 100% sort of thing. Each time it is just a straight up 20% chance.

 

 

That and RNG hates some people more then others :D

Link to comment
Share on other sites

Far as I can tell it is just a 20% of something happening. Not a 20% chance of happening, there for after 5 times it should = 100% sort of thing. Each time it is just a straight up 20% chance.

 

 

That and RNG hates some people more then others :D

 

^^ This

 

RNG is also temperamental like my 4 yr old son. :rak_09: Case in point, My Armormech had Re'd 24 (i kept the 25th one cuz it was augmented & I needed to upgrade my gear) Overkill Heavy Exoskeleton Belts w/o getting a single purple schematic. However, she RE'd 5 Overkill Heavy Exoskeleton Gloves & got 2 Purples from it.

Link to comment
Share on other sites

RE's 24 items, got 11 crits. You are right. That is no where near 20%

 

Update : 39 items 18 crits

Now up to 68 REs and 21 crits. OMG ONLY 3 CRITS OUT OF THE LAST 29 THE SYSTEM IS BROKEN!!!!

-But still the overall rate since I started tracking is 30.88% for me. I have a hunch I know what is causing the issue with those with lower numbers but it will take several more days of research and RE attempts with different types of items.

Edited by HelinCarnate
Link to comment
Share on other sites


×
×
  • Create New...