Pages: [1] :: one page |
|
Author |
Thread Statistics | Show CCP posts - 0 post(s) |
Sale gueule
|
Posted - 2008.03.01 17:53:00 -
[1]
My skills for inventing are 5,4,4 The inventions were all made on gallente cruiser bpc
I made 21 tries with 0,8 chance modifier decryptor and only one worked
i repeat : Only one worked, that means something like 200m loss And, as it was written before, the cruisers success porba is (was?) 0,3
Speaking in therms of probability laws, with a probility of success of 0.24, just one success is likely to be impossible (as having 0 success for a 7 tries shots is like 5% proba, that means, for the 3 seven shots tries a probability of 1 success less than 0.02% ..................)
Lemme repeat please : ===> 0.02 % <=== (don t forget the %, that means 0,0002 on 1)
Do something changed ? Are my success rate informations wrong ??
Is the probability of success dependent of units on the market or somethign like that ?
Thanks for answering
|
Erotic Irony
0bsession
|
Posted - 2008.03.01 18:14:00 -
[2]
sample is not statistically significant
working as intended ___ Eve Players are not very smart. Support Killmail Overhaul
|
Anaalys Fluuterby
Caldari
|
Posted - 2008.03.01 18:16:00 -
[3]
Seems I got your successes...
Last week, 2 attempts at a Rapier, 2 successes
Originally by: Audri Fisher On the other, the emo tears being cryed in this thread tell me that just because you shoot somebody for a living, does not mean you aren't a carebear
|
Sale gueule
|
Posted - 2008.03.01 18:23:00 -
[4]
?? how can sample is statistically insignificant ?? You wan t a 99,95 trusting limit or what ? (sorry don t know the english name)
3 times 7¿tries is really significant while 0 success rate is 5% for one shot
if i made one 7 shot try, yeah, it wouldn t be statistically significant
second reply, the you things invented are not the sames as mine, so it really du nt help me =(
thks
|
Christari Zuborov
Amarr Ore Mongers R0ADKILL
|
Posted - 2008.03.01 19:11:00 -
[5]
Wrong forum to post this in. Go to Stations, Starbases, and Outposts forum - those guys know their stuff and there are stickies with great information in them.
|
Tamia Clant
New Dawn Corp New Eden Research
|
Posted - 2008.03.01 19:18:00 -
[6]
You may find this thread relevant to your problem.
Looking for queue-free research slots? Click here!
|
Zeba
Minmatar Pator Tech School
|
Posted - 2008.03.01 19:25:00 -
[7]
Originally by: Tamia Clant You may find this thread relevant to your problem.
Yup. Was just about to link that one. Invention takes deep pockets to make consistant isk. Yes quite a few get lucky and score big right off the bat but to make a reliable income from it you need to be able to absorb those irritating streaks of failure.
Originally by: MM0RPG EVE Review "if you aren't my friend you are my income."
|
WarlockX
Amarr Free Trade Corp
|
Posted - 2008.03.01 19:28:00 -
[8]
if the chance is greater then 0 then it has to happen to someone, somewhere, sometime, it happened to you. 0.02% means it would happen once in 5,000 there's alot more then 5,000 ppl logged in each day, that means this should statistically happen 4-5 times a day if everyone was inventing. That's not a small chance at all its actually quite large if you think about it that way. ----------------------------------------------- "I often quote myself. It adds spice to my conversation." |
gfldex
Kabelkopp
|
Posted - 2008.03.01 21:15:00 -
[9]
Originally by: Sale gueule ?? how can sample is statistically insignificant ??
You have to take a few hundred samples to get a trend. Some thousands to get okish numbers and a few million to get good numbers. That should have been teach to you in school.
--
There are countless games in the world. There are at least as many ppl that dont like one or more rules of said games. That never stopped smart game designers from creating good games.
|
Sale gueule
|
Posted - 2008.03.02 17:34:00 -
[10]
thanks for the replies exept gfldex
gfldex, very usefull comment, i admit
About mathematics : there s no confidences limits, either samples or trend in those calculs It s just a binomial distribution law, the 0.3 success rate is given by CCP and definelty true, at least in this post (the numbers i gave are made with this 0.3), there is no approximation; either samples. The probability to have "k" success on "n" tries with a GIVEN probability "p" is fixed and defined with no approximation.
That s why i asked if that success rate of 0.3 was true
Those things are teached at nusery school, you should know that, won t you ?
So next time both of you, think a little before saying useless things; stupid and wrong things who helps no one, you just appear to others as : people who want to seem great but knows nothing
|
|
Nizomi
Caldari Technopolis Ventures
|
Posted - 2008.03.02 17:45:00 -
[11]
Originally by: gfldex
Originally by: Sale gueule ?? how can sample is statistically insignificant ??
You have to take a few hundred samples to get a trend. Some thousands to get okish numbers and a few million to get good numbers. That should have been teach to you in school.
You are wrong, you dont need millions of tries to get a good number.. I guess you dont know much of statistics either?
|
Tortun Nahme
Minmatar Heimatar Services Conglomerate
|
Posted - 2008.03.02 17:49:00 -
[12]
they also teach the difference between the words "certainty" and "probability" in nursery school
there is a "probability" of 30% success, not a "certainty"
Originally by: Cecil Montague They should change that warning on entering low sec to:
"Go read Crime and Punishment for a few days then come back."
|
Sale gueule
|
Posted - 2008.03.02 17:50:00 -
[13]
:(
in this cases, there are no good numbers, either samples, as i just explained
everythings is on the given number of 0.3 for ship inventing by ccp
I m just asking if this number is true or not.
I d just like to know if i have been really unlucky or if i have wrong informations
|
Tortun Nahme
Minmatar Heimatar Services Conglomerate
|
Posted - 2008.03.02 17:55:00 -
[14]
it's true, some one just griefed your luck, probably podded it on its way up to high sec as in passed through .4 space... 1 jump from safety
Originally by: Cecil Montague They should change that warning on entering low sec to:
"Go read Crime and Punishment for a few days then come back."
|
Sale gueule
|
Posted - 2008.03.02 17:56:00 -
[15]
Originally by: Tortun Nahme they also teach the difference between the words "certainty" and "probability" in nursery school
there is a "probability" of 30% success, not a "certainty"
You smoked too much mate ?
|
Sale gueule
|
Posted - 2008.03.02 17:58:00 -
[16]
Originally by: Tortun Nahme it's true, some one just griefed your luck, probably podded it on its way up to high sec as in passed through .4 space... 1 jump from safety
hehehehe
i hope you really done that ! he just deserved it
|
Sale gueule
|
Posted - 2008.03.02 18:04:00 -
[17]
No way to have an answer like :
Yes your 0.3 success rate information is true, you have been unlucky or No it's wrong, it's actually : ..........................
Bah, thanks anyway
|
Tortun Nahme
Minmatar Heimatar Services Conglomerate
|
Posted - 2008.03.02 18:06:00 -
[18]
man both my replies were witty and answered your question, you are no fun
yes the chances are .3 you were unlucky
but my other ways of saying it were better
Originally by: Cecil Montague They should change that warning on entering low sec to:
"Go read Crime and Punishment for a few days then come back."
|
Sale gueule
|
Posted - 2008.03.02 18:08:00 -
[19]
!!
sorry i did not get it
this forum lacks so much of fun that i missed the meaning of your reply
thanks for answering man
|
Tortun Nahme
Minmatar Heimatar Services Conglomerate
|
Posted - 2008.03.02 18:12:00 -
[20]
lol no prob
Originally by: Cecil Montague They should change that warning on entering low sec to:
"Go read Crime and Punishment for a few days then come back."
|
|
Sphynx Stormlord
Gallente Anqara Tech
|
Posted - 2008.03.02 18:14:00 -
[21]
The thing is, there are streaks. Some people get streaks of successes, some people get streaks of failures.
Everyone who gets what they think is a streak of failures, comes and posts on the forums that invention is broken.
It isnt, 100's of people are inventing away just fine. If they were not, hac's would be costing 200 mil each, for the good ones, and covert ops cloaks would be back up to 50 mil each.
|
ShardowRhino
Caldari The Older Gamers R0ADKILL
|
Posted - 2008.03.03 01:37:00 -
[22]
you cant say that 21 attempts truly defines the % chance of success. You can't say that your rate based on 21 or so attempts is somehow the norm. What you can say is that out of your attempts that you got X% success rate BUT thats ALL you can really say.
Saying that you now know what the real chances of success is by such a small sampling is insane. An example,though opposite, is someone buying 21 lottery tickets for the big jackpot. The person has never played the lottery before but in his handful of tickets he has the sole winning ticket. He can't say that everyone's chances of winning are 1 out of 21. Just because he happened to have the winning ticket in his hand doesn't honestly prove what the chances are, especially since the actual statistics will say that his chance is 1 in millions. Its the opposite of bad luck but it shows how your small sampling is in no way an indicator of the true chances of success.
I've had streaks where all of my invention jobs went through. Then theres been times where they all fail. Theres also times where its 50% or 75% success out of the batch i just finished. I could say its 0% or 100% or anything between , it doesn't really matter unless I track a multitude of attempts.
|
Celege Je'thries
Gallente Antares Fleet Yards
|
Posted - 2008.03.03 02:13:00 -
[23]
With 1500 attempts you might be within 1% of the actual success rate. With 7-30 you can be off by 50% quite easily.
|
ghosttr
Amarr ARK-CORP
|
Posted - 2008.03.03 02:19:00 -
[24]
Its well known that the lv5 invention skills will through your success rate off. CCP continues to claim that its working, but wont provide any details.
For now if you have lv5 invention skills you are screwed My Blog |
wammeh II
|
Posted - 2008.03.03 03:38:00 -
[25]
I think 0.02 is very high considering how many that does invention - quite a few I'm sure, but not that many. Will it happen? Yes. Here's a different question... Should it happen?
I got 6 t2 bpo agent offers back in the lottery days. Though I did run 4 agents. Not great prints either but ok. 3 offers from same agent. 2 from another. And 1 more from a third. Accepted 3.
I'm not saying this is a great sampling either, but I think the BPO lottery for one was, well, bugged. Several very lucky people got several amazing prints. The probabilities were insane.
So, yes I feel streaks has a tendency to occur in eve and I know about the psychology behind seing patterns. I also know a thing or two about pseudorandom number generation and it's not easy to create one that doesn't ditch out crappy numbers. imo... the RND is either broken or inadequate.
Should it happen? A streak can completely destroy a battle. Should it be left up to probability? I thought it was a game of skill?
|
Dimagus
|
Posted - 2008.03.03 05:11:00 -
[26]
The biggest mistake in perception that people make is thinking in terms of "invention" probability, or "mission" probability, or "loot" probability. There is no troll in the server that utilizes magic or psychic phenomenon to predict the who, what, where, when, why and how of your game.
Your 7 tries or 21 or whatever are not consecutive, they are not one array block from the Random Number Generator. Those tries are completely unrelated to each other and there are literally hundreds of other random instances for the server and other players calculated in between. Someone or something else ended up with the good rolls, and you got the crappy ones. Deal with it.
|
Sale gueule
|
Posted - 2008.03.03 10:36:00 -
[27]
Originally by: ShardowRhino
you cant say that 21 attempts truly defines the % chance of success. You can't say that your rate based on 21 or so attempts is somehow the norm. What you can say is that out of your attempts that you got X% success rate BUT thats ALL you can really say.
Saying that you now know what the real chances of success is by such a small sampling is insane. .
omg
seems your really don t understand : 1/ statistics and proba laws or 2/ what i said
I ll repeat a second time :
the 0.3 success rate IS GIVEN BY CCP
i don t have to do stats about real chances of success as it is GIVEN, there is no sampling to take into account in what i said
you say : "Saying that you now know what the real chances of success is by such a small sampling is insane."
Sure i know this success rate, CCP GAVE IT
So, if you do 1 billion cruisers inventions you'll have the 0.3 success rate But i don t give a f*** about that, IT S GIVEEEEEEEEEEEEN
|
Sale gueule
|
Posted - 2008.03.03 10:59:00 -
[28]
So, to end this stupid discussion, as my answer was just : is the 0.3 success rate for cruisers invention is a good information, or is it wrong ?
I Give you the binomial distibution law. It works like that : If every shot is independent of any other ones (like when you roll a dice, the number you ll have do not matter of any other rolls you made). This is our case, if you do 2 inventions, the fact that the second invention is successfull or not is independent about the first invention, it s still 0.3 no matter what
This laws gives you the probability to have "k" success in "n" tries KNOWING THE PROBABILITY "p" GIVEN BY CCP (ffs) (obviously, iy wont work if you don t have the REAL value of "p") the prbability "p" is 0.3, the number of tries "n" is 21 k P(X=k) <=== this is the probability that numbers of success equals "k" 0 0.000559 1 0.005027 2 0.021544 3 0.058476 4 0.112776 5 0.16433 6 0.187806 . . . . . . 21 0.00000000000105
Sums of all the P(X=k) for k going from 0 to "n", is 1 (obviously) If you the the sums of all the P(X=k) for k going from 2 to 21, you ll have the probability to have 2 success OR MORE for 21 tries If you do 1-this probability, you have the proba to get 1 or less seccess (so 1 or 0) and it is 0.005585 (whose simply the sum of P(X=0)+(P(X=1)
You understood now ?
If i hear once more : you can t say it s a good sampling, i kill you all
|
Ardent Rellik
Gallente MURAKAMI INDUSTRIES
|
Posted - 2008.03.03 11:17:00 -
[29]
Nope, they haven't, so far I have had same success rate per/100 jobs as I always did.
|
|
|
|
Pages: [1] :: one page |
First page | Previous page | Next page | Last page |