[TERA PC & Console] En Masse is closing, but TERA lives on! We will continue to support TERA PC (NA) and TERA Console until service is transferred. Stay tuned for more information.
[TERA Console] The Grotto of Lost Souls update (v85) is now live! Read the patch notes here: https://bit.ly/TERACon_v85
[TERA PC] The 64-bit update (v97) is now live. Check out all the changes delivered on August 11 here: https://bit.ly/tera64_patchnotes
[TERA PC & CONSOLE] Summerfest Part 2: The Beach Bash is on from August 11 until September 1! Participate in event activities to earn tokens redeemable for costumes, consumables, mounts, and more! Details: https://bit.ly/tera_sf20
Comments
Actually there could be some merit to this statement. However, you forgot that there is an extra 3% chance to succeed every time you fail an enchantment.
Let me attempt to shed some light on this statement. Disregarding this extra % chance, each enchantment attempt is independent of each other, i.e. each enchantment attempt do not affect the success rate of another such attempt. Hence, you can use the multiplicative rule here to calculate the % chance of only succeeding on the 4th try: 0.40*0.37*0.34*0.69 = 0.0347208 = 3.47% chance:
0.40 - chance of failing on 1st attempt
0.37 - chance of failing on 2nd attempt
0.34 - chance of failing on 3rd attempt
0.69 - chance of succeeding on the 4th attempt
By conducting a hypothesis testing at say a significance level of 0.05, and checking for the % chance of only succeeding at the 4th attempt, the % probability is 3.47%, which is lower than the 5% significance level. Hence, there could be a possibility that the success rate of this enchantment attempt is not 60% as stated. (There is also a possibility that you are just very unlucky.)
You should also include the total amount of item exp needed to maximize the enchantment success rate for each tier of equipment.
Thank you.
What in the...
your 3.47% is not a p-value, it's the actual likelihood of succeeding on the 4th and only 4th try, so not only is that percentage useless because what you really want to know is the likelihood of an event occurring within 4 tries (not just on the fourth), but it's not even something that you can deem significant or not significant based on your desired alpha cutoff of 0.05. Again, it's not a p-value.
???????
1. I agree that I have made a terrible mistake calculating the p-value, while I disagree on how you want to calculate the p-value. p-value should have been the probability of an event and its more extreme cases of occurring, which in this case is the summation of the probabilties of succeeding on the 4th attempt and beyond, up to 15 attempts (since the enchantment chance is 100% at the 15th attempt), and not either the probability of succeeding only on the 4th attempt, nor the probability of succeeding at the 1st to 4th attempt. Hence, I have recalculated the p-value, which gives 0.05032 = 5.032%, which is nearly equal to the alpha-value. Thus, it is debatable but it seems that there is insufficient evidence to conclude that the enchantment chance is not 60%.
2. I disagree with your point of me making up a random alpha-value. If you had conducted a simple Google search or read some textbooks on statistics, you will see that the most common alpha value used is 0.05, because it is a good balance between Type 1 error (when one incorrectly rejects the null hypothesis, in this case, when one concludes that the enchantment chance is not 60%, while in actual fact, it is) and Type 2 error (when one fails to reject the null hypothesis, in this case, when one concludes that the enchantment chance is 60%, while in actual fact, it is not). Hence, I did not conjour up the alpha-value by myself, but yes, other alpha-values like 0.1 and 0.01 have been used also in special circumstances.
Well Ninja isn't that strong imho. I'd like to think I'm a decent Ninja as my DPS is usually 1st or 2nd (never 3rd) in IMS. The problem is that Ninja needs Slaying in order to outdps a warrior who doesn't have slaying. So far the highest Ninja DPS (In all of TERA) I've seen is only 3M/s with Slaying. Gunner and Brawler are much stronger than Ninja and require less buttons pressed when soloing BAMs. Even though I'm not very skilled with Gunner or Brawler I can manage a nice 30-40 secs with +0 Twistshard gear and Bellum accessories on them.
Stormcry on a priest to solo mid-tier BAMs is a bit overkill and is on a whole new level compared to Twistshard or Frostmetal. I think OP isn't doing that bad considering their gear.
I never said you made up your alpha. Yes 0.05 is used ubiquitously across many fields and areas of research and statistics, that's fine. What you are misunderstanding is when a p-value is to be used to make conclusions about a hypothesis-based test.
I flip a coin, the probability of heads is 0.50, which is more than 0.05, so it is significant. Do you see the problem here? Significance is a property of a comparison, and does not describe a standalone event. Using significance to describe this is meaningless: The probability of heads is 0.50, hence it is significant, but so is tails, because it's also 0.50. There is no comparison here. The same can be said about tossing a six-sided die. Each number has a 0.167 chance of showing up, does that mean they're significant? What if it was a 50 sided die? Does that mean they are not significant? You don't use significance to describe these events, it's meaningless.
Instead if you were comparing two different dice, and maybe one of them was unfair, that's when you can use significance. Over a sufficiently large n, you simulate enough rolls to obtain a distribution of tosses and now are faced with the question if one die is SIGNIFICANTLY different from the other, that's where you use the p-value. This is where you calculate a test statistic (could be z or t or whatever distribution or assumptions you make). The test statistic CORRESPONDS to a p-value based on the distribution, (e.g. z = +/- 1.96 corresponds to a p of 0.05 for a two-tailed test), and then based on whether or not that p-value is more or less than your alpha, you can then accept/reject the null.
Regardless of if your calculation was correct or incorrect for his probability of success on enchanting a weapon in x number of tries, using significance or hypothesis testing is meaningless. You've already calculated precisely his chance of succeeding. That 3.47% or 5.03% is a discrete measurement of his likelihood of success, you don't need to further qualify that number as significant or not, the answer is already there. The only time you need significance to describe something is if you were going to compare something, like if one weapon had a 30% chance of succeeding while another weapon had a 50% chance. After simulating enchanting both weapons x number of times, is there a significant difference in the number of tries needed to succeed? Only in this situation can you formulate a testable hypothesis.
I don't think my Mystic can outdps my ninja but it does come close to the kill time.
ROFL