Whats the point ......

[Ninje]

This is happening far too much , That being gems failing to make.

Didnt bother me before because everything is just roll of the dice luck but fk me not even the most inlucky player in the world can keep getting fails on 97% 98% etc etc its just not right as its hard enough to get the ingrediants and getting them to apply to the items let alone failing at this % .

Even worse when trying to make gems with marbles. Please have look something has to be wrong.

 

[DeadGirl]

Yep had this a lot too. Don't go over 90% is my advice, unless it's 100% :P

 

[Samuel]

The code for this I think was part of the files... Or Td coded... Not entirely sure, I am happy to have a look to see of there are any obvious bugs though.

Sam

 

[Ninje]

The code for this I think was part of the files... Or Td coded... Not entirely sure, I am happy to have a look to see of there are any obvious bugs though.

Sam

ty fella

 

[Miriganu]

another thing ive noticed. every time i try to make a gem while on my horse = failed. Only tried twice, both over 90%..havnt tried making while on horse since lol

 

[Skyline]

I've not had one fail on 90+% lol, only when i risk it with no Ada it's failed.

But have heard a few people have them fail a few times at 90+%

 

[mStation]

is @bodyluck involved in gem crafting?

if i am correct i have heard that a few peeps have 5 bodyluck others 1 or even 2. if this is involved might explain why others are more "unlucky"

 

[Carribean]

In my opinion it's fine tbh Your just unlucky, some are gonna fail

 

[Skyline]

In my opinion it's fine tbh Your just unlucky, some are gonna fail

Agree with this to be honest.

The risk of a fail is all part of the gem making rather than buying. Otherwise it would be pointless selling gems.

 

[Foopah]

Agree with this to be honest.

The risk of a fail is all part of the gem making rather than buying. Otherwise it would be pointless selling gems.

I think the point was that he has had quite a few fail at 97% and 98% - Yeah fair enough, if say 2-3 out of every 100 attempts failed that would be realistic, however if 4 outta 5 gems fail to make at 97% then there's something wrong and it is not producing the chances it is saying it is.

I'll show my workings with this 97 gems out of 100 attempts... If something has 100% chance of working it is guaranteed to work... Right? So if something has a success rate of 97% (100-97 = 3) means that only 3 gems should fail out of every 100. Obviously chances and probability do not work to a set of rules, oh wait, they do... Something with 97% success rate should at some point (be it after 100 or 10,000) balance out and show the real percentage of success (And this should be 97%).

I'd be very interested in the results of a GM testing this at 97% success and see how many gems he can create out of 100 possible gems.

 

[Reverance]

In the same boat as TC

Multiple gems created with a 95%+ Chance breaking.

1/2 Chance of breaking and I see it happen 3/4 times out of maybe 10 gems.

It does feel as though something is going wrong at the higher % Chance.

 

[Samuel]

I think the point was that he has had quite a few fail at 97% and 98% - Yeah fair enough, if say 2-3 out of every 100 attempts failed that would be realistic, however if 4 outta 5 gems fail to make at 97% then there's something wrong and it is not producing the chances it is saying it is.

I'll show my workings with this 97 gems out of 100 attempts... If something has 100% chance of working it is guaranteed to work... Right? So if something has a success rate of 97% (100-97 = 3) means that only 3 gems should fail out of every 100. Obviously chances and probability do not work to a set of rules, oh wait, they do... Something with 97% success rate should at some point (be it after 100 or 10,000) balance out and show the real percentage of success (And this should be 97%).

I'd be very interested in the results of a GM testing this at 97% success and see how many gems he can create out of 100 possible gems.

I would imagine if I tested 100 gems I would see a much larger number than 3 failures...

Reason being the % ratio would work on a individual gem basis... e.g. on each gem there would be a piece of code that ran which looked something like:

With SuccessRate being 97

if Random(100) =< SuccessRate then
CreateGem
else
Destroy;

Random(100) could = 99 10 times in a row if you were unlucky enough...

Again this is an asumption on how I think the code should work, I have not actually looked at it yet.

Sam

 

[Liandrin2]

I would imagine if I tested 100 gems I would see a much larger number than 3 failures...

Reason being the % ratio would work on a individual gem basis... e.g. on each gem there would be a piece of code that ran which looked something like:

With SuccessRate being 97

if Random(100) =< SuccessRate then
CreateGem
else
Destroy;

Random(100) could = 99 10 times in a row if you were unlucky enough...

Again this is an asumption on how I think the code should work, I have not actually looked at it yet.

Sam

You should have the same chance of seeing more than 3 failures as you do of seeing less than 3 failures....

If it's not averaging out at 3 failures then the 97% is wrong.

The problem is you would have to make thousands to get a decent enough sample size.

 

[Skyline]

I think the point was that he has had quite a few fail at 97% and 98% - Yeah fair enough, if say 2-3 out of every 100 attempts failed that would be realistic, however if 4 outta 5 gems fail to make at 97% then there's something wrong and it is not producing the chances it is saying it is.

I'll show my workings with this 97 gems out of 100 attempts... If something has 100% chance of working it is guaranteed to work... Right? So if something has a success rate of 97% (100-97 = 3) means that only 3 gems should fail out of every 100. Obviously chances and probability do not work to a set of rules, oh wait, they do... Something with 97% success rate should at some point (be it after 100 or 10,000) balance out and show the real percentage of success (And this should be 97%).

I'd be very interested in the results of a GM testing this at 97% success and see how many gems he can create out of 100 possible gems.

Not really, because as Sam has already said, it's not a 97% on 100 gems it's a 97% per gem.

Sam could make 100 successful gems... He could then only make 60 successful out of the next 100.

It's all down to chance really... I'll happily do some testing myself and see if there is actually something not quite right, but I personally have not had any issues.

 

[mStation]

lol let's make some math shall we

what as been said above is unfortunately all correct..

linear probablity would say that 97% chance means get a clean 97 results over 100.. this would be linear which means will have the same effect even if 100 is 1 or 10000

the problem here is that random numbers from CPU's aren't really random.. they unfortunately follow aswell the gaussian curve



in probability there is a generic rule which say's that if you for instance have a 97% of chance.. this would rappresent the median value of the gaussian curve.. this means that you should have most of the probabilty to get a 97% but at the same time as you can see you can get also a much higher value or a much lower value according to the "luck" of the extracted number.


now while this is true even for these server files.. we should now find out if there is really a bug or not.. and i think the only real way is by reading the code and looking at what get involved (number side) when crafting. because trying to trial and error such a probabilistic thing is quite difficult

 

[Liandrin2]

they unfortunately follow aswell the gaussian curve

Are you sure? o.o

At least as far as programming languages are concerned I thought they all used linear...

 

[TheMasterGee]

Are you sure? o.o

At least as far as programming languages are concerned I thought they all used linear...

I'm kinda mystified, not in a disrespectful way, but although I quit machine language programming 25/8 before the 486 came about, I always derived random numbers in my coding as a factor of the program execute address (non absolute), the bios time/clock and the range of number you wanted randomising.

That's not cpu based.

Maybe compilers do it differently. Not heard of gaussian or linear tbh, which probably means it's taught in college.

 

[Kai]

Besides mstation's post (I'm not familiar with cpu random number generation), the math is just simple probabilities.

Chance of 1 fail is 3/100. 97% is 97/100 there is absolutely no difference.

Chance of 2 fails in a row is (3/100)^2=0.0009 or 0.09%

Chance of 3 fails in a row is (3/100)^=0.000027 or 0.0027%

Edit: Sam the chance of an individual gem according to the code you wrote is still 97% to work, if someone rolled that 10 times in a row its worse odds than winning the lottery. (ie. 0.000000000000059049%)

 

[Samuel]

Besides mstation's post (I'm not familiar with cpu random number generation), the math is just simple probabilities.

Chance of 1 fail is 3/100. 97% is 97/100 there is absolutely no difference.

Chance of 2 fails in a row is (3/100)^2=0.0009 or 0.09%

Chance of 3 fails in a row is (3/100)^=0.000027 or 0.0027%

Edit: Sam the chance of an individual gem according to the code you wrote is still 97% to work, if someone rolled that 10 times in a row its worse odds than winning the lottery. (ie. 0.000000000000059049%)

Sure... But what I was trying to point out was although it might be 97% that does not mean that your are only going to see three breaks out of a 100...

You could have 25 break out of a hundred and that does not mean that the calculation is broken, it means that you have been unlucky and hit that 3%.

Sam

 

[Kai]

Ah all I meant to say is that if its working properly 25/100 is such astronomically bad odds that it will never happen unless something is broken.

Your right it won't always be 97/100 but since its a % it means that given infinite tries it absolutely will always be 3% chance to break.

Some people will have bad luck most will never see a fail at that rate. I haven't had any fail yet at 90%+ thankfully!


Edit: Just curious, is there a log generated by the craft npc?