The tool tips say the chance to successfully reverse engineer a green item as being 20%. This is clearly rubbish. I've produced MANY MANY more greens than needed for a 1 in 5 hit rate.

I'm not sure if it's the SWTOR random number generator that's broken, or if the tool tip is wrong. Either way it should be fixed..

(and yes I know, 20% chance doesn't guarentee a 1 in 5 hit, but it should average out to that over time - and clearly, in this game it doesn't).

The tool tips say the chance to successfully reverse engineer a green item as being 20%. This is clearly rubbish. I've produced MANY MANY more greens than needed for a 1 in 5 hit rate.

 

I'm not sure if it's the SWTOR random number generator that's broken, or if the tool tip is wrong. Either way it should be fixed..

 

(and yes I know, 20% chance doesn't guarentee a 1 in 5 hit, but it should average out to that over time - and clearly, in this game it doesn't).

No, it shouldn't. 1/5 and 20% are two different things, and the number of times you'd actually have to make the roll in order to see if it's 20% or not is huge(at least a thousand), beyond that, computers aren't truly random.

The tool tips say the chance to successfully reverse engineer a green item as being 20%. This is clearly rubbish. I've produced MANY MANY more greens than needed for a 1 in 5 hit rate.

 

I'm not sure if it's the SWTOR random number generator that's broken, or if the tool tip is wrong. Either way it should be fixed..

 

(and yes I know, 20% chance doesn't guarentee a 1 in 5 hit, but it should average out to that over time - and clearly, in this game it doesn't).

Each item has a 20% chance to teach you something new but also an 80% chance to do nothing at all. Every time you reverse engineer, you roll the dice again with the same odds.

The tool tips say the chance to successfully reverse engineer a green item as being 20%. This is clearly rubbish. I've produced MANY MANY more greens than needed for a 1 in 5 hit rate.

 

I'm not sure if it's the SWTOR random number generator that's broken, or if the tool tip is wrong. Either way it should be fixed..

 

(and yes I know, 20% chance doesn't guarentee a 1 in 5 hit, but it should average out to that over time - and clearly, in this game it doesn't).

Just means you weren't very lucky. I remember a night when I've managed to get successful RE about 8 times out of 10 for Cybertech.

Good times. Good times....

Each item has a 20% chance to teach you something new but also an 80% chance to do nothing at all. Every time you reverse engineer, you roll the dice again with the same odds.

yes.. i know basic probability theory.. still says that in a large sample, 1 in 5 should pop.

yes.. i know basic probability theory.. still says that in a large sample, 1 in 5 should pop.

No, that isn't how it works. You clearly aren't understanding "basic probability theory".

yes.. i know basic probability theory.. still says that in a large sample, 1 in 5 should pop.

You are talking about Law of Large Numbers. The caveat is that the sample size must be big enough for it to be applicable. It's also not a certainty either.

So in effect, I stole all your crits.

it's not rubbish, you clearly do not know what you're doing

yes, cause it's easy to stuff up reverse engineering.. takes great skill.

The tool tips say the chance to successfully reverse engineer a green item as being 20%. This is clearly rubbish. I've produced MANY MANY more greens than needed for a 1 in 5 hit rate.

 

I'm not sure if it's the SWTOR random number generator that's broken, or if the tool tip is wrong. Either way it should be fixed..

 

(and yes I know, 20% chance doesn't guarentee a 1 in 5 hit, but it should average out to that over time - and clearly, in this game it doesn't).

Averages out for me. Was doing a lot of crafting on my new Gunslinger, 1 was the smallest number and 9 was the largest and it was definitely averaging at about 5 reverses.

yes.. i know basic probability theory.. still says that in a large sample, 1 in 5 should pop.

Yes. What you're failing to understand is that, in terms of probability, your entire crewskill production from launch until this moment isn't a 'large sample'.

Let me say nicely what the other insensitve people are saying.

It's random, and yeah, it feels like it's broken, but really it isn't. The other night I RE'd the same thing 34 times before it popped, then I RE'd something else and it popped immediately.

It sucks sometimes, but it's not broken.

I feel your pain.

Let put it the other way, you don't get 20% chance of each try to success.

You are getting 80% chance of each try to fail.

So instead of 20% x 5 = 100% success in RE. You are facing 80% x 5 = 400% FAIL!!! in RE.

Does it make you feel you were actually very very luck to sucess in only 1 out of 10 attempts now???

I'd say you have about a 10% chance of not getting a new schematic after 10 tries on a single item.

Proof is left as an exercise for the reader.

Yes. What you're failing to understand is that, in terms of probability, your entire crewskill production from launch until this moment isn't a 'large sample'.

Actually in statistical terms it would be, assuming he REs regularly, representative samples can seem quite small.

There are no true random numbers when it comes to computers. Computers use an algorithm and a seed number (often the CPU clock) to generate a seemingly random number. Technically, they are pseudo-random numbers.

If anyone didn't know, the RE was bonked when the game first came out. I suspect that it is still off a bit. Though, it is MUCH better than it was. I sure am curious as to what algorithm and seed they are using.

There are no true random numbers when it comes to computers. Computers use an algorithm and a seed number (often the CPU clock) to generate a seemingly random number. Technically, they are pseudo-random numbers.

Sure, but if you want to break it down that far, we have no method of generating actual random numbers at all. The closest we've got is a place that uses multi-step randomizing based off precision atmospheric measurements, but even that would be deterministic with enough knowledge. Algorithm + unpredictable seed gives you 'functionally random', which is plenty close enough to truly random for the purposes of measurement.

Let put it the other way, you don't get 20% chance of each try to success.

 

You are getting 80% chance of each try to fail.

 

So instead of 20% x 5 = 100% success in RE. You are facing 80% x 5 = 400% FAIL!!! in RE.

 

Does it make you feel you were actually very very luck to sucess in only 1 out of 10 attempts now???

That's still not true. In order for 5 tries of 20% to beome 100%, you have to eliminate tried choice.

For example, let's say there are 5 numbers; 1, 2, 3, 4, 5. You have 20% probability of picking 5. If I pick 2 then I'm within the 80% failure bracket. If I put 2 back to the mix, I still have 20% probability of picking 5. If I remove 2, then I have 25% chance of picking 5.

In this case, we are putting the 2 back into the mix. So you will ALWAYS have 20% chance of picking 5. In theory if I pick enough times, I SHOULD have 20% success rate. However, this isn't really guaranteed. Refer to Law of large numbers.

Let put it the other way, you don't get 20% chance of each try to success.

 

You are getting 80% chance of each try to fail.

 

So instead of 20% x 5 = 100% success in RE. You are facing 80% x 5 = 400% FAIL!!! in RE.

 

Does it make you feel you were actually very very luck to sucess in only 1 out of 10 attempts now???

no.. succeeding 1 in 10 attempts in 10% chance, not 20...

by the way i realise it's not guaranteed.. im just saying it should approximate a 1 in 5 win.. and it's not even close to that.. maybe my losses are making up for everyone else 1 shotting them

Let put it the other way, you don't get 20% chance of each try to success.

 

You are getting 80% chance of each try to fail.

 

So instead of 20% x 5 = 100% success in RE. You are facing 80% x 5 = 400% FAIL!!! in RE.

 

Does it make you feel you were actually very very luck to sucess in only 1 out of 10 attempts now???

I prefer the "glass is half full" way of thinking......if I have 5 items to RE, then I have 0.8 to the 5th power (0.32768, or let's say 33%) chance to NOT get a schematic.

I keep telling that to myself as that 0.8 to the first power keeps kicking my *** over and over.....

The 20% RE does NOT work like this:

RE 1 item = 20% chance; therefore RE 5 items = 100% chance! I r so smart

Rather, it works like THIS:

Will RE X items:

Item 1 has 20% chance of success, 80% chance to fail

Item 2 has 20% chance of success, 80% chance to fail

Item 3 has 20% chance of success, 80% chance to fail

...

Rinse and repeat until you hit the 20% chance.

In short: The chances of success are rated individually, not cumulative.

yes.. i know basic probability theory.. still says that in a large sample, 1 in 5 should pop.

You are falling into the gambler's fallacy. Yes, the average is expected to come out to 1 in 5 as the sample gets very large (as in, the limit approaching infinity), but each individual attempt is always 20%. When looking at probability for independent events, what already happened doesn't matter. There is no corrective force that says that because you failed 20 times, the 21st time has to succeed because you're not getting a 1 in 5 average for your actual attempts. Your next RE is still a 20% chance to succeed.

To put it another way, if you said you should expect around 200 successful REs out of your next 1000 tries, that'd be reasonable. If you've done 999 REs and failed and you expect your next attempt to be anything but a 20% chance at success, you'd be wrong. (And probably cursed.)

No, that isn't how it works. You clearly aren't understanding "basic probability theory".

No he isn't. More specifically, the 'theory part. IE: Just because the theory states that, with a large sample size you'll find the probability to be correct, it doesn't mean it always is.

I do wonder what the % it's come out with overall is for the population of single server, though, as that would certainly be classed as a huge sample size.