I did open it.
Thx - at least you opened something before forming an opinion. This is a topic close to heart of any player and often argued, in the past, without scientfiic basis, so I can understand the initial animocity.
... The original thread's data was ultimately reinforced by confirmation in the game's code...
Someone else also remarked on this before.
It is one thing to take a single variable from the code. It is another to explain why we are getting Pack (
note - I am not saying "slot") drop rates like we do.
It's like saying - Black magic happens on a Pack Level, but at least we know that 7.5% have something to do with it on Slot level.
but you seem to be suggesting that the thread failed to work out how the theory actually works.
No. I used the word "threads". It's all other "threads" that refers to the work of ComradeShepard7 without understanding the science behind it. The latest example of such threads is the discussion crusader thread. Some people made valid statements, others just made popular assumptions.
ComradeShephard7's thread successfully uses statistical measures to prove that a specific variable (7.5%) plays a role (on slot level). That paper is, however, not sufficicient in itself to predict
Pack drop rates, especially when multiple slots have different probabilities to drop the same rarity item/equipment.
There is no theory derived in the paper from Comradeshephard7. We don't buy Packs on slot level. Hence it doesn't show the droprate of a AP.
I encourage you to expand on information as much as you like, but I'm having a hard time seeing how re-presenting the same information with a different author is going to expand on what was already established - a 7.5% UR drop rate, with each potential UR slot calculated individually.
Maybe if you clarified what it is you hope your document will show?
Consider reading it now, after opening it.
- It introduces a generic formula to predict the PACK drop rate(without using the 7.5% variable or the raw data), by simply using store mechanics (2 slots per UR, mutually exclusive) as basis. That formula holds true for any probability setting. We didn't use such formula in ComradeShepard7's paper.
- thereafter we can use the 7.5% in that formula and calculate the predicted drops on pack level.
- It goes on to show that the predicted PACK drop rate matches very well for large samples. ComradeShepard7 document doesn't predict based on Combinational theory. It predicts based on overall samples. Some samples were very dirty.
- The new document that was submitted, now provides a way to verify the accuracy of some samples and to cross verify the accuracy of the comradeShephards' findings. So it provides the ability to say whether a submitted sample is "dirty" (not matching the theory) and should be ignored in the statistical data. The findings of ComradeShepard's paper were skewed slightly because of dirty samples. Up until now we had no measure to see how far it was skewed compared to most likely outcomes.
- It shows why samples < 30 is not useful
- why 2 simultaneuous UR's in AP happens very seldom and what that probability is.
- ... (and more)
The only part of this that hasn't already been exhaustively covered in ComradeShepard7's thread is E) time required for maxed manifest. It's simple enough to calculate that the average cost to unlock and fully level every UR weapon is 138,600,000 credits if one buys Arsenal Packs exclusively. However RNG will play a huge role, any purchases other than Arsenal Packs will slow your progress, and of course how long it takes you to earn those credits depends on a number of variables.
I'm glad some people are still sufficiently invested in the this game to take the time and effort required to put this kind of thing together, but there's no need to re-invent the wheel.
Thx. Kindly show me in that paper where I can find the UR Pack Drop rate for an Arsenal Pack?
U used 15% to calculate those 138,600,000 and assumed that one can simply add 7.5% with 7.5%. That is an incorrect assumption for many reasons. These values becomes crititical when 3 mutually exclusive slots drops the same type of inventory. Probability Theory predicts 14.4375% and that matches to 0.0415% with specific (large) samples in the raw data. The more likely expenditure is 124m, and means a different average AP drop rate.
As someone with three ME 3 MP accounts and nearly 1200 hours on my main account that will, realistically, never have a maxed manifest, I honestly don't believe that it would be possible to calculate that information in any practical way.
I fail to see, how the fact that you have 3 accounts, or played a certain amount of hours, serves as a scientific way to calculate probabilities better than the binomial distribution.
The sheer number of variables involved precludes it, at least from our perspective. ...
Surprise. There is just 1 variable for UR weapon slots, and combinational theory explains how to combine that for multiple slots.
If within 0.04% margin of accurate prediction over 903 Purchases is not accurate enough...
-->The Store is much less chaotic than some think. One just have to know the rules and boundaries. The submitted document explains those boundaries.
Nature does the rest to always ensure that probabilities are met towards infinity.
Hence, with a certain probability, we can now show when you can expect another UR drop over a longer row of Credits (based on a ceiling certainty).There's now even a lookup table for that.
...rudimentary understanding of math is enough to know that there would never really be such thing as "average time to max manifest" unless one eliminated a massive portion of the total data and focused only on people with maxed manifests who could somehow accurately report the exact number of credits they spent before maxing it.
Unless I've misunderstood the mathematical definition of "average," if you have to eliminate the vast majority of the data to calculate it, then "average" tends to lose its practical meaning.
You misunderstood the mathematical definition of "average".
Retour en haut
if u open the document u will notice that I'm using the same raw data as comradeshepard7. Difference is that I talk about Packs and not Slots. We buy packs, not slots. Hence the probabilities change when slots are viewed together. I've even referenced those links u tried 2 find... In there
Use that number in your time to maxed manifest calculation.
