keep modifications similar item level

"

DarkenDragon wrote:

as for your "not a formula" sorry to burst your bubble but everything is based on a formula. what you were doing is just using statistics to figure out what the ratios are. which of course is based on a formula the game uses to determine what is the chance of getting said value.

Ah right. When I said "Those numbers aren't from a formula", I also explicitly stated that I ran a simulation to get them. A simulation is not the same thing as deriving a formula. I didn't use statistics at all aside from counting up the ones that matched the criteria I specified.

"

DarkenDragon wrote:

which again is a formula. and you used a formula to scale those values up as well.

I didn't scale the numbers, I plugged in new values for the sim and then ran it again. The reason they correspond very well to each other is because I had an incredibly large sample size (1 million items with prefixes, suffixes, and values attached to each of them). Think of it like using one million Chaos Orbs.

"

DarkenDragon wrote:

either way they arent calculated right because your assuming all modifcations have the same amount of ranks, which isnt true. some have less some have more. you cant just make it an exponent of 6, especially since only 3 of the slots can be any said mod because you can only have 3 prefixes and 3 suffixes.

I didn't assume an even mod distribution at all. I took into account ALL possible mods that could roll on the items which was parsed by some clever chaps who data-mined the client.

I literally did nothing that involved probability. Please at least read my post before telling me my methods are wrong. Thusfar, I haven't seen any statistics from you at all aside from your invalid formulas.

===============

The data that I've provided shows that reducing the number of possible affixes that an item can roll GREATLY increases the probability of obtaining high level gear.

im sorry but using a simulation is no different than using a formula. and its no different than figuring out probabilities and statistics, I dont think you understand these terms. because everything your saying your not doing, is exactly what you've been doing. this is university level statistics 101

also how do you even think a simulator gets its number? it uses a formula to figure it out.

"

DarkenDragon wrote:

im sorry but using a simulation is no different than using a formula. and its no different than figuring out probabilities and statistics, I dont think you understand these terms. because everything your saying your not doing, is exactly what you've been doing. this is university level statistics 101

also how do you even think a simulator gets its number? it uses a formula to figure it out.

Are you high? Simulation and formulation are entirely different things. Let's take a basic example: rolling a die. Say we want to find the probably of getting a number equal to or greater than x on a single standard 6-sided die. The formula for this is simple:

(7 - x) / 6

So clearly we can see that if our target is 5, the probability is 2 / 6, or 0.333....

Now then, here's a simulation to find that same probability. I'm going to write it in Python, but I think it'll be simple enough for anyone to follow.

trials = 1000000
target = 5
positive = 0

for i in range(0, trials):
_ roll = random.randint(1, 6)
_ if roll >= target:
___ positive = positive + 1

estimated_probablity = float(positive) / trials

What that does is roll a virtual die one million times. And it estimates the probability based on a tally of the outcomes. Executing that will give you a number like 0.333145, which is unlikely to be exactly what the actual probability is, but is more likely to be more accurate as you increase the number of trials. In no form does the actual formula for the probability of achieving a desired outcome manifest itself in the simulation.

vmrob is correct here, and I'd be very curious to hear where you took university level statistics (and whether you passed it). Though I will say this: Simulation wasn't part of my level 101 statistics class. It was all formulation in there, so if you've only taken 101, I can see why you might be confused.

with a virtual die, how do you think the computer determines what the outcome is? how do you think the system simulates you rolling the die a million times?

oh right, a FORMULA. no matter how you look at it, its a formula that determines the outcome.

and if you did do it that many times or more, the value will end up going more towards 0.33333333 unless your die is not properly balanced. but in a system simulating it, it will assume its perfectly balanced. same with this game, its all perfectly balanced there for we can assume the formula will stay true. so if you keep going, increasing the amount of trials, you'll find that it'll be very close to the formula.

formulas are used to create rules, simulations are used to prove said rules. nothing more nothing less. therefore formulas are created based on simulation's outcomes.

and guess what all this is, statistics and probability. and my education came from ryerson university with a bachelors of science in computer science.

"

DarkenDragon wrote:

with a virtual die, how do you think the computer determines what the outcome is? how do you think the system simulates you rolling the die a million times?

oh right, a FORMULA. no matter how you look at it, its a formula that determines the outcome.

and if you did do it that many times or more, the value will end up going more towards 0.33333333 unless your die is not properly balanced. but in a system simulating it, it will assume its perfectly balanced. same with this game, its all perfectly balanced there for we can assume the formula will stay true. so if you keep going, increasing the amount of trials, you'll find that it'll be very close to the formula.

formulas are used to create rules, simulations are used to prove said rules. nothing more nothing less. therefore formulas are created based on simulation's outcomes.

and guess what all this is, statistics and probability. and my education came from ryerson university with a bachelors of science in computer science.

Generating a random number between 1 and 6 is not a probability formula.
Doing that a million times isn't a formula either.

if you taken computer science or any programming language, you'll know theres no such thing as "random". what we have is pseudo random which means kind of random, its a program that will simulate random through a formula.

so yes ITS A FORMULA!

"

DarkenDragon wrote:

with a virtual die, how do you think the computer determines what the outcome is? how do you think the system simulates you rolling the die a million times?

oh right, a FORMULA. no matter how you look at it, its a formula that determines the outcome.

If your point is to prove that a simulation uses at least one arbitrary formula, you're right. When generating a pseudo random number it does use a formula (usually something like (w + x) * y % z) and when we divide the positive outcomes with the number of trials, that division is technically a formula. But what does that have to do with anything? o.O It's a simulation model which happens to use very, very generic formulas as building blocks to estimate the probability. It's a very fundamental and enormous difference between simulation and formulation.

I really think we're just being trolled right now, because this guy, while right in the most trivial and unrelated of aspects (and wrong in most others) has no applicable point.

Edit: Actually, I wouldn't consider the random number generation to be a formula at all. It really is just an algorithm instead. The division to get the estimated probability from the number of trials and desired outcomes is a formula though.

"

DarkenDragon wrote:

if you taken computer science or any programming language, you'll know theres no such thing as "random". what we have is pseudo random which means kind of random, its a program that will simulate random through a formula.

so yes ITS A FORMULA!

Not all functions/algorithms are formulas... you're using the word way too broadly.

In either case, it's not the formula for the odds of getting a 4 or higher on the die.
You can calculate that probability by simulation without using the formula, just like you can use a simulation to find the odds of getting 500+ es on a chest armor without figuring out the complex formula for it... which is what vmrob did & said.

"

DarkenDragon wrote:

formulas are used to create rules, simulations are used to prove said rules. nothing more nothing less. therefore formulas are created based on simulation's outcomes.

What do you think the "formula" for determining the roll on a piece of gear is? I'll give you a hint: it has nothing to do with combinations and permutations.

In the computer science world, as I'm sure you know, we call step-by-step procedures algorithms. I would say that there is an algorithm for producing a piece of gear, but certainly not a probabilistic formula.

Just for my own sanity: What do you consider a formula? Can you give me a concrete example of whatever formula you think the game uses to roll gear is?

"

vmrob wrote:

"

DarkenDragon wrote:

formulas are used to create rules, simulations are used to prove said rules. nothing more nothing less. therefore formulas are created based on simulation's outcomes.

What do you think the "formula" for determining the roll on a piece of gear is? I'll give you a hint: it has nothing to do with combinations and permutations.

In the computer science world, as I'm sure you know, we call step-by-step procedures algorithms. I would say that there is an algorithm for producing a piece of gear, but certainly not a probabilistic formula.

Just for my own clarity: What do you consider a formula? Can you give me a concrete example of whatever formula you think the game uses to roll gear is?

I'm pretty sure the formula is something like "player = taekvideo ? Sockets = 1 : sockets=rand(...)"
At least that's what it feels like QQ