Homam Corazza Triple Attack rate

[mallister]

Bumping this with good reason: Homam Corazza looks to be borked.

Quoting Lobivopis from the alla thief forums:

this is the start of your problem. that guy's sort of a known idiot.

[Valyana]

Based on the quoted data, we can conclude with 95% confidence that THF TA rate w/o Corazza is 9.08-11.64% and TA rate w/ Corazza is 8.74-11.44%. Assuming that we know that the rate w/o Corazza is 10%, this would seem to rule out a 2.5% boost, but is entirely consistent with a 1% boost. The sample size is too small to determine whether or not there's any effect.

Confidence Interval for Proportions Calculator

[Ambir]

HAI GUYS IGNORE THIS DATA, THE FIRST FIVE SHOWED 20%, THE NEXT FIVE SHOWED 0%. YOU CAN'T JUST COMBINE THEM LOL

Then you tell me, brainiac, if you do the same experiment twice and get results as different as 9.77 and 11.07%, and then add 1% proc rate to the gear, what result will you get, EXACTLY?

Based on the quoted data, we can conclude with 95% confidence that THF TA rate w/o Corazza is 9.08-11.64% and TA rate w/ Corazza is 8.74-11.44%. Assuming that we know that the rate w/o Corazza is 10%, this would seem to rule out a 2.5% boost, but is entirely consistent with a 1% boost. The sample size is too small to determine whether or not there's any effect.

Confidence Interval for Proportions Calculator

Actually no,
For 1000 data points and a 10% proc rate (is that what we're using for a Base THF?) we can say with 95% confidence that the result will be between 8.14 and 11.86%.

If we then add a 1% proc on top (if Homams 1%), we'll be 95% sure the result will be between 9.06 and 12.94%

2.5% would be 95% chance of between 9.81-15.19

Theres just too much variation added because of the THFs basic proc rate to accurately measure the Homam proc rate accurately, but if the choice is only 1% or 2.5, (and not 0 like Allah says ) then it's probably 1%.

[Deadgye]

lawlz. Here's the problem. You instantly see "alla" and disregard that it has any backing. Maybe if I was pulling something from =10(The Main Forums) you'd have a point. Thief Forums ≠ Main forums. Doubting information because it came from alla's thief forums is retarded. Guess where it was found out that shark bites 200TP fTP modifier is 2.5? etc.

With that being said:

How about they test it with something other than a job that has Triple Attack as a natural trait.

How about no? What was being tested was the triple attack rate incease on THIEF using homam corazza.

The fact that theres a 2% difference between each set of data with the same setup means that you have no way of measuring the 1-2.5% variation due to the Corazza. The sample size just isn't large enough to reduce the error in the innate TA.

You can't just average 2 inaccurate results and expect to get a true value...

Gotta love Allah

Good job only reading the first quote.

Actually no,

Actually yes. You just did the math for ONE THOUSAND data points. He had EIGHT THOUSAND.

At 10.07% with 8000 data points and 99% confidence level the range for the true proportion is 9.14 - 10.86. This means that either a. Triple attack caps at 10%, b. Thief doesn't start out with 5% triple attack, c. Homam Corazza adds less than 1%, d. Homam Corazza is broken, or e. Triple Attack caps at 10%

I may be slightly wrong as I'm just using that calculator linked above so please correct me if I am, but everything said in regards to the last post I made has been just so much more wrong..

[Redshift]

guys, all of this arguing is breaking our friendships apart