test content
What is the Arc Client?
Install Arc

Enchant fusion & tier 3 craft odds...my record log.

shamurai7shamurai7 Member, Neverwinter Beta Users Posts: 99
edited July 2013 in General Discussion (PC)
Hello all!
I'm usually not one to do this type of experiment. However after an experience of 7 consecutive failures on an enchant fusion I thought I'd start making a record of successes.
I also started logging successes of my cleric tier 3 profession crafting but have a MUCH smaller pool to get a true average.

First off the odds according to their 60% success rate of fusing tier 4 ->tier 5 enchants and failing 7 times is 0.163% or 16 in 10,000.
(.4 to the 7th power.)
Now for comparison the odds for success 7 times in a row at 60% success rate is 2.8%, or 280 in 10,000. This is unlikely but plausible.
(.6 to the 7th power.)
This to me seemed VERY strange unless i'm a mathematical anomaly.

So my results in fusing 194 tier 4 enchants and runes yielded 72 successes. This calculates to a 37% success rate. A LONG shot from the advertised 60%.
Coincidence? I don't believe in coincidences.... I believe in math.
I could continue this logging but I think I have a large enough pool to determine the truth.

And as promised my much smaller log of tier 3 profession crafting.
This is taken from cleric crafting specifically shirts and pants advertised @ 60% chance of tier 3.
Results of 12 attempts show 4 successes and 8 failures. This so far is an even 30% success rate against their advertised 60%.
Consider though the much smaller pool can yield less accurate results.

I felt obligated to share this with you all as a fellow gamer. It isn't proof of anything, but it is very strong evidence of a rigged lottery.
Good day.

Update 1: Here is a small update with recent additions over the last several days.
31 more tier 4->tier 5 fusions yielded 13 successes. Raising the average to a 38% success rate.
MORE IMPORTANTLY!
Tier 3 profession crafting........
6 more attempts yielded 1 success. This is an astounding 5 of 18. Lowering the average success to 28% vs the advertised 60%.
I've decided not to abandon this and keep updating it periodically until the trend changes and the devs fix the odds they advertise. They are false.

Update 2:
I stopped logging enchants... The trend seems to have changed and I think they snuck in a fix without telling us. However I have continued logging tier 3 professions and it is still just as bad.
12 more attempts yielded 3 successes. I am at a total of 8 out of 30, or 27% success rate vs the advertised 60%.

I recommend anyone investing in mithral profesion supplies to sell them off and forget about it. With the rising prices of dragon eggs you will statistically make roughly 20 thousand AD over the course of about 1 solid week of crafting for your massive 3+ million AD investment in purple assets.
The failed pieces net roughly 30 thousand loss per piece and successes net approximately 110k profit with an average sale. Subtract your average of 3 failures and your making 20k.

Sell your mithral junk before the majority catches on to the lie because no one will want to craft these shirts and pants once they realize the truth.
Post edited by shamurai7 on
«1

Comments

  • tang56tang56 Member, Neverwinter Beta Users, Neverwinter Guardian Users, Neverwinter Knight of the Feywild Users Posts: 0 Arc User
    edited June 2013
    I've never had a failed enchant aside from a single rank 2 fusing. Given the number of players in the game and how insignificant our individual contributions are it's perfectly possible. After all, someone has to be the unfortunate person who gets all the bad luck.
    RIP Neverwinter 26/06/2014
  • damianessdamianess Member, Neverwinter Beta Users, Neverwinter Guardian Users Posts: 283 Bounty Hunter
    edited June 2013
    It is 60% success rate for that one fusion only, not overall success rate.

    Edit: cleaned up my post a bit
  • morsitansmorsitans Member, Neverwinter Beta Users Posts: 1,284 Bounty Hunter
    edited June 2013
    damianess wrote: »
    It is 60% success rate for that one fusion only, not overall success rate.

    Um...those two are mathematically identical. You're saying "50% chance of heads applies to just one coin flip, not coin flips overall"

    If you have a 60% chance of making each fusion, then while you can happily get loads of failures or successes in a row, the global average should always tend to 60%.
  • silveralucardsilveralucard Member, Neverwinter Beta Users, Neverwinter Guardian Users Posts: 410 Bounty Hunter
    edited June 2013
    well i think that probably we need to check if in fact it is happening however to do that we would need to xcalculate how much of a portion of the total cases we need to track in order to have a group that we can use for the study, as they mention that is done so many times that 300 cases are not a good number for the study we would need at least 1000 or maybe more XD and in fact you ae correct the global aeverage of failure and success should tend to 60% you ca be unlucky a lot as well but not that unlucky and you mention and probably with an error of 10% in your calculaiton i think the true success rate in game is 50% and not 60% same applies for crafting is 10% below advertised but i doubht the Dev team will check on that since base on that is that they prefer that you buy the shard to be able to do them 100% the time :)
    Everything works out in the end . If it hasn't worked out yet, it isn't the end...
  • lerdocixlerdocix Member Posts: 897 Arc User
    edited June 2013
    shamurai7 wrote: »
    Hello all!
    I'm usually not one to do this type of experiment. However after an experience of 7 consecutive failures on an enchant fusion I thought I'd start making a record of successes.
    I also started logging successes of my cleric tier 3 profession crafting but have a MUCH smaller pool to get a true average.

    First off the odds according to their 60% success rate of fusing 4 tier 4 enchants and failing 7 times is 0.163% or 16 in ten thousand
    (.4 to the 7th power.)
    Now for comparison if the odds for success 7 times in a row at 60% success rate is 2.8%. or 280 in ten thousand. This is unlikely but plausible.
    (.6 to the 7th power.)
    This to me seemed VERY strange unless i'm a mathematical anomaly.

    So my results in fusing 194 tier 4 enchants and runes yielded 72 successes. This calculates to a 37% success rate. A LONG shot from the advertised 60%.
    Coincidence? I don't believe in coincidences.... I believe in math.
    I could continue this logging but I think I have a large enough pool to determine the truth.

    And as promised my much smaller log of tier 3 profession crafting.
    This is taken from cleric crafting specifically shirts and pants advertised @ 60% chance of tier 3.
    Results of 12 attempts show 4 successes and 8 failures. This so far is an even 30% success rate against their advertised 60%.
    Consider though the much smaller pool can yield less accurate results.

    I felt obligated to share this with you all as a fellow gamer. It isn't proof of anything, but it is very strong evidence of a rigged lottery.
    Good day.

    Umm, you never studied anything more advanced then adding and extracting on maths, did you?
    To get a statistically accurate view, you need a much larger pool, not silly and horrible extrapolation from a try that can give you about any result.

    Take a coin, throw it 5 times, you'll get say 4 heads and 1 tails, does that mean there is 80% probability of getting heads in 10.000 throws? According to your horribly flawed logic, yes. Meanwhile if you ACTUALLY did 10.000 throws you would be very close to 50% for heads or tails, which is accurate.

    Same case here. If you had 7 successful combines would you say its 100% rate? Well, you probably would, but you would still be horribly mistaken.

    Sorry, but your "tests" hold absolutely no value statistically UNLESS you really do 100-200 or preferably more fusions.
    And no, you aren't mathematical anomaly, you just have no clue about mathematic and statistic.
  • morsitansmorsitans Member, Neverwinter Beta Users Posts: 1,284 Bounty Hunter
    edited June 2013
    Dude, he did it 194 times. That's more than enough to trend to 60% reliably.

    ...or at least, "in the region of 60%". With 194 repeats, the chance of getting 37% or lower is very, very small.
  • shamurai7shamurai7 Member, Neverwinter Beta Users Posts: 99
    edited June 2013
    morsitans wrote: »
    Dude, he did it 194 times. That's more than enough to trend to 60% reliably.

    He's 12 man... don't argue with him. He said 'maths'.

    I'm aware my pool for enchants was plenty large enough. There should be no more than a 10% discrepancy. This tells me at most the true success rate is no higher than 47%....and at worst 27%.

    I think the true success is probably 40 and not 60.
  • lerdocixlerdocix Member Posts: 897 Arc User
    edited June 2013
    morsitans wrote: »
    Dude, he did it 194 times. That's more than enough to trend to 60% reliably.

    Yea, I have missed that somehow, I suppose I can take back few words I've said earlier, his results should be better then, but still I'd keep counting for the sake of math, the bigger pool of tries, the more accurate the outcome. I've done way to many statistic calculations to think that pools that go below 1k repeats are reliable in any way.
  • morsitansmorsitans Member, Neverwinter Beta Users Posts: 1,284 Bounty Hunter
    edited June 2013
    Would be a fair conclusion to reach. Do you dare try it for level 5->6? :D
  • clortbagsclortbags Member, Neverwinter Beta Users, Neverwinter Guardian Users Posts: 0 Arc User
    edited June 2013
    Not another one of these threads >.<
  • silivrenasilivrena Member, Neverwinter Beta Users, Neverwinter Guardian Users Posts: 20 Arc User
    edited June 2013
    tang56 wrote: »
    I've never had a failed enchant aside from a single rank 2 fusing. Given the number of players in the game and how insignificant our individual contributions are it's perfectly possible. After all, someone has to be the unfortunate person who gets all the bad luck.

    He is talking about rank 3 to rank 4...Pretty sure rank 2 has a 90% success rate so not really surprised you don't fail on them...I don't either except on a rare occasion.
  • tang56tang56 Member, Neverwinter Beta Users, Neverwinter Guardian Users, Neverwinter Knight of the Feywild Users Posts: 0 Arc User
    edited June 2013
    silivrena wrote: »
    He is talking about rank 3 to rank 4...Pretty sure rank 2 has a 90% success rate so not really surprised you don't fail on them...I don't either except on a rare occasion.

    I said rank2 was the only enchant I've had fail, not that rank 2 is the only enchant I've ever done. I have had the luck of not having a rank 4 or 5 fail yet. :rolleyes:
    RIP Neverwinter 26/06/2014
  • silveralucardsilveralucard Member, Neverwinter Beta Users, Neverwinter Guardian Users Posts: 410 Bounty Hunter
    edited June 2013
    yes 194 should be trending to 60% however as lerdocix mention and also i didi that is also to few of a group to test this because the number of fusions is like 10k daily probably, so we would need at least the 10% of the cases of a group to be able to correctly trend that is using statistic models, as i mention to be fair the Error in the calculation with194 should be around 17% to 24% with that error gap he is still between the correct success rate, however it will be good to test with a Statistical hypothesis case :D let see who can do that and admin should be able to provide like the amount of fusion in the total muestral values lets said of to to tady for the 60% success rate :D and with that we can test on you true values within the complete universe so that will help us to probe if this is correct or incorrect

    PD: i'm not native speaker so sorry about my english
    Everything works out in the end . If it hasn't worked out yet, it isn't the end...
  • mconosrepmconosrep Member Posts: 0 Arc User
    edited June 2013
    lerdocix wrote: »
    UNLESS you really do 100-200
    lerdocix wrote: »
    o think that pools that go below 1k repeats are reliable in any way.

    So first you say that he needs to have done 100 at least, and then when it is pointed out he did 200, you change the goal posts......
  • shamurai7shamurai7 Member, Neverwinter Beta Users Posts: 99
    edited June 2013
    I have done MANY of 5-> 6 but haven't tracked them because nothing seemed odd about it.
    My mind keeps a mental note of rough odds and 5 ->6 feels about right where it should be.

    I wasn't inspired to track 4 ->5 until the 7 failures in a row.

    I'll continue to track tier 3 crafting though because it doesn't 'feel' right so far. It would probably take me alone another year to get a reputable result from crafting though.
  • thebrimanthebriman Member Posts: 218 Arc User
    edited June 2013
    It's funny, I had the exact opposite experience. All of my first 5, and 7 of my first 8, Rank 4 fuses were successful and I thought to myself "Wow, is it really just a 60% rate? Sure seems like higher." I just chalked it up to small sample size. I haven't tracked my fuses since, but I usually seem to have a few more succeed than fail, so 60% sounds about right overall, based on my experience.
  • morsitansmorsitans Member, Neverwinter Beta Users Posts: 1,284 Bounty Hunter
    edited June 2013
    I've just run an excel macro to check deviation from expected percentage with 174 attempts, and the chances of it getting outside of 5% either way are small. The chance of it being out by 20% is vanishingly small (though, of course, not impossible :))
  • tang56tang56 Member, Neverwinter Beta Users, Neverwinter Guardian Users, Neverwinter Knight of the Feywild Users Posts: 0 Arc User
    edited June 2013
    shamurai7 wrote: »
    I have done MANY of 5-> 6 but haven't tracked them because nothing seemed odd about it.
    My mind keeps a mental note of rough odds and 5 ->6 feels about right where it should be.

    I wasn't inspired to track 4 ->5 until the 7 failures in a row.

    I'll continue to track tier 3 crafting though because it doesn't 'feel' right so far. It would probably take me alone another year to get a reputable result from crafting though.
    Are you in a guild? You can always ask your guild to do it, or open a request for volunteers to help gather data in the forum. The events are all independent so it shouldn't matter who does the fusing.
    RIP Neverwinter 26/06/2014
  • lerdocixlerdocix Member Posts: 897 Arc User
    edited June 2013
    mconosrep wrote: »
    So first you say that he needs to have done 100 at least, and then when it is pointed out he did 200, you change the goal posts......

    No, not really, I have addressed his 7 tries with the 100-200 thing, then added he would need to do much more anyway to get accurate result, I know selective reading is very comfortable, but what comes after coma is just as important in a sentence, so for the future please, read it all. And silveralucard said why. 200 repeats would give more accurate result, but would still leave a wide gap for error and look, its enough to be closer to the goal, but small enough to still show basically anything and exactly that happened. Repeating tries for as much as he can do(preferably thousands for most accurate result) would give the real outcome, but I doubt he or anyone would have the patience. I know I wouldn't.
  • shamurai7shamurai7 Member, Neverwinter Beta Users Posts: 99
    edited June 2013
    thanks morsitan! :) very nice calculation!

    I knew I was of course skewed very slightly by choosing to start with my initial 7 failures. I'm still convinced it's a 40% success rate. I do plenty of game development and I know how easy it is to let a mistake like that into your product.
    However in this circumstance i'm offended because this costs players expensive green wards. Or even gives players a false sense of advantage.
    The graphical user interface for this has NOTHING to do with the math happening in the background.
    Without access to their numbers I can't prove anything.

    Perhaps i'll look into the game files tonight... I almost guarantee though that this data is stored server side to protect from exploits though.
  • morsitansmorsitans Member, Neverwinter Beta Users Posts: 1,284 Bounty Hunter
    edited June 2013
    If I were cynical, and sneaky, I would program the game to monitor sum(enchants(levelX)) and adjust the probability on the fly to stop enchant flooding.

    Let's hope they're not cynical and sneaky, eh? :P
  • ulyxosulyxos Member, Neverwinter Beta Users, Neverwinter Hero Users Posts: 49
    edited June 2013
    With 194 tries you are pretty much hitting assymptotic behaviour. Their random is flawed somehow. This is the same for skill kits. You often get succession of 5+ failure which are very unlikely given the advertised rates of success. The most suspect part is that failure of kits rarely come by singleton which should be the most common figure. My take is that like a lot of their effect (all the bugs where a increase is replaced by a decreased) they invert the ratio somehow... A 75% success becomes a 25% success as soon as you fail once.
  • shamurai7shamurai7 Member, Neverwinter Beta Users Posts: 99
    edited June 2013
    ulyxos wrote: »
    With 194 tries you are pretty much hitting assymptotic behaviour. Their random is flawed somehow. This is the same for skill kits. You often get succession of 5+ failure which are very unlikely given the advertised rates of success. The most suspect part is that failure of kits rarely come by singleton which should be the most common figure. My take is that like a lot of their effect (all the bugs where a increase is replaced by a decreased) they invert the ratio somehow... A 75% success becomes a 25% success as soon as you fail once.

    A totally agree with skill kits....I got a very nasty vibe from them after my first 20 or so.....
    Without even collecting data on these I can promise anyone that the 75% is a lie. I wasn't so concerned about these because they don't effect the players in such a negative way as enchant fusing does.
  • lerdocixlerdocix Member Posts: 897 Arc User
    edited June 2013
    ulyxos wrote: »
    With 194 tries you are pretty much hitting assymptotic behaviour. Their random is flawed somehow. This is the same for skill kits. You often get succession of 5+ failure which are very unlikely given the advertised rates of success. The most suspect part is that failure of kits rarely come by singleton which should be the most common figure. My take is that like a lot of their effect (all the bugs where a increase is replaced by a decreased) they invert the ratio somehow... A 75% success becomes a 25% success as soon as you fail once.

    And how would you explain scoring 5 critical hits in row with 25% chance in your stats? As soon as you crit once your crit goes to 100% for next few hits?
    Thats how chances work. Each try is independent, you can have a string of crits or no crit at all for lengthy time or actually one crit for every 4 attacks. What is a good way to tell how it works? A huge sample size with small error gap.
  • shamurai7shamurai7 Member, Neverwinter Beta Users Posts: 99
    edited June 2013
    lerdocix wrote: »
    And how would you explain scoring 5 critical hits in row with 25% chance in your stats? As soon as you crit once your crit goes to 100% for next few hits?
    Thats how chances work. Each try is independent, you can have a string of crits or no crit at all for lengthy time or actually one crit for every 4 attacks. What is a good way to tell how it works? A huge sample size with small error gap.

    Your absolutely correct, but most people will not grasp that concept. Many will swear that previous results have an effect on subsequent results.

    This can only apply in finite sample data such as drawing clubs from a deck of cards... The more clubs you pull the less likely youll pull another because the sample size of clubs was lessened by 1.

    In the game though this doesn't apply. Your not drawing numbers from a hat with 100 pieces of paper marked 'win' or 'lose'.
    Each drawing in the game starts you with a fresh 100 chances. (At least I hope that's how it works unless they have some jank system they are using.)
  • jalfreyjalfrey Member Posts: 0 Arc User
    edited June 2013
    morsitans wrote: »
    Um...those two are mathematically identical. You're saying "50% chance of heads applies to just one coin flip, not coin flips overall"

    If you have a 60% chance of making each fusion, then while you can happily get loads of failures or successes in a row, the global average should always tend to 60%.

    But only with a significant amount of test data - 194 tests is nothing even like a significant amount of test data it is likely there are millions of fusions attempted each week and probably 10s of millions since the game launched. Maybe if you tested 500 000 fusions you would see better results, but 194 simply is not a significant sample.
  • shamurai7shamurai7 Member, Neverwinter Beta Users Posts: 99
    edited June 2013
    jalfrey wrote: »
    But only with a significant amount of test data - 194 tests is nothing even like a significant amount of test data it is likely there are millions of fusions attempted each week and probably 10s of millions since the game launched. Maybe if you tested 500 000 fusions you would see better results, but 194 simply is not a significant sample.

    Yes it is.
  • kindyrekindyre Member Posts: 101
    edited June 2013
    It's very prone to streaks. Very. Might be due to a flaw in their random number generation, I don't know.

    But it does average to the advertised amount. That I know for fact because I've done... god knows how many thousands of fusions for profit. And I'm not the type to gamble... so I calculated the projected costs and profits and, in the long run, I always came out with what I was counting on. Only very minor variation across large enough volumes.

    But I did have huge streaks. Exceedingly improbable streaks. In both directions.
  • clortbagsclortbags Member, Neverwinter Beta Users, Neverwinter Guardian Users Posts: 0 Arc User
    edited June 2013
    shamurai7 wrote: »
    Yes it is.

    With a 40% fail chance every fuse it's not even close to a large enough sample to draw any accurate conclusions...
  • jalfreyjalfrey Member Posts: 0 Arc User
    edited June 2013
    clortbags wrote: »
    With a 40% fail chance every fuse it's not even close to a large enough sample to draw any accurate conclusions...

    Some people have no understanding of inferential statistics or probability, it isn't worth wasting our time trying to explain to them if they are not willing to listen.
Sign In or Register to comment.