The sim handles outliers itself - by performing those tens of thousands of iterations, and then providing the mean - as mentioned above.
Just as you will have some outliers near the top (where lucky crits, etc happened), you will also have outliers near the bottom. Over those iterations, these outliers work themselves out. You can view the graph by clicking âView Full HTML Reportâ after you run a sim.
The only comparison I gave - that involved sims - were imperfect play ret compared to perfect play ret (or at least perfect play based on the APL that theorycrafters have determined to be that âperfect playâ).
Youâre assuming that each iteration is not nested by configurations of variables. Again it is an estimate that assumes best-case scenarios even if it is looking at the value that falls within the 50th percentile.
Yeah but the results of those tens of thousands of iterations are not a normal distribution. I donât honestly know where youâre going with this.
I donât really want to go on and on about potential faults in the sim (we all know they exist). I donât want to argue about the difference between a mean or a median. We are all smart people - we understand the difference.
Youâre making a faulty argumentâŚas you have this entire thread.
If its not normally distributed after 1000s of iterations (which is why you run so many iterations) then itâs lost meaning at that point. If I have 1000 iterations run but there is a negative skew in the distribution, are you able to determine that the mean of said sample is accurate to what you are asking it?
Again as many people stated, numeric values such as representation numbers or sim values are not meant to be taken in a vacuum but many of your arguments have been doing so.
No, it wonât be normally distributed, because thatâs not how randomness (or in this case pseudo-randomness) works. In fact, if it were normally distributed, there would be a problem.
And it runs so many iterations to get rid of those outliers we talked about earlier, when delivering the results to the user.
And yes - we can assume that it is accurate to within 0.05%. This is given to us from the developers of the app itself, who were the ones to determine the proper number of iterations needed to get us to that accuracy level. I tend to trust them (they are probably a lot smarter than me).
Can you point out the faulty argument, and explain whatâs exactly wrong with it? If Iâm wrong, I have no problem admitting it and rethinking things.
If sim iterations values are not normally distributed after 1000s of iterations (which is why you run so many iterations in the first place) then itâs lost meaning at that point. If I have 1000 iterations run but the distribution is negatively skewed, are you able to determine that the mean of said sample is accurate to what you are asking it?
Again as many people stated, numeric values such as representation numbers or sim values are not meant to be taken in a vacuum but many of your arguments have been doing so. Not to say the representation numbers are meaningless persay but a lot of the assumptions you make based on that data is a bigger stretch.
An example of this I can give is examining representation numbers of keys done higher than a 20 is more a statement about over-game designing. Enhanceâs lack of defense is meaningful because the game is designed to have a large number of 1 shots and high single-damage events instead of rot damage due to how healing kits function (further blame can be pointed towards how talents work now as well). But representation data in general is going to be skewed towards ret anyways because people just like to play Ret regardless of performance and Enhance is historically not popular. To give you an example of this in Mythic last season, Enhances overall was performing better than Ret yet they still had a similar number of parses. Looking at it in a more targeted manner, there were 877 Enhance parses for Mythic Sark vs 1275 Ret Parses looking at 10.1.5 data.
The logic that you have been presenting so far in this thread would say that Ret is somehow better spec in multiple facets on Mythic Sark and there are huge problems with Enhance Shaman when in reality Enhnace was 2nd from the top and Ret was 4th from the bottom in damage. Yet Ret had about 400 more parses for the fight. Turns out Deus Vult lives on regardless of performance.
As for the top two paragraphs, they are the same as your last post - weâve covered this.
As for the rest, I mean no argument from me. If thatâs what you gathered from any of the data presented here, then good for you. Fortunately to me, we donât have to guess about last tier regarding what was more popular or more effective there - because the data already exists, as itâs in the past. We can just go look it up.
My main concern is the present and the future.
But I appreciated reading your thoughts either way.
But understanding this is important for determining steps in the future. And the thoughts Iâve gathered have very little to do with the representation data youâve presented. Which is the point. Again, the data you presented and your inferences amounted to âIce Cream Sales and Murder rates are relatedâ. Others and I have been saying âBoth Ice Cream Sales and the Murder Rate tend to increase in Summerâ.
Iâve been polite, humble, and helpful this entire time. But come on man - like where are you even getting this from. How does one surmise this?
Look, at the end of the day - i cannot determine what you (or anyone else) infers from dataâŚbut if thatâs where it took you, then so be it. Again, no argument from me. Iâm just lost is all.
I honestly canât argue with it. I think Iâve said my piece. Good day.
You keep pointing to representation numbers of Ret vs Shaman as a statement that something is wrong with Shaman or as you said, shaman being âa second-class citizenâ. Weâve been saying itâs meaningless to the discussion of difficulties that Shaman deals with. Representation numbers are a faulty metric. The ease of play argument is also very irrelevant.
But thereâs a reason 38x more people are playing Ret right now in season 3 at the highest level than are playing Enhancement right?
whereas in DF season 2, more people played Enhance at the highest level than played Ret
The argument I keep hearing is âoh well Ret has just always been more popularâ. But it wasnât last season at the highest levels. Enhance gained representation as the difficulty went up.
And no. Representation numbers are not a faulty metric. They show, literally, representation numbers. I cannot control what you surmise from those numbers - but the metric itself is certainly not faulty by literally every definition possibly imaginable.
Then please, letâs discuss why representation is down.
If the problem is, like many people hinted to earlier - defensives - then letâs spend time lobbying for defensive changes - as opposed to arguing with eachother about how accurate some data might be, etc.
Yes because Ret is broken in this season and garbage last season. Has no bearing on Shaman whatsoever. Ret is also way more represented than other classes as well.
That statement is pretty faulty as well. Looking at this for reference: https://raider.io/mythic-plus-rankings/season-df-2/all/world/leaderboards-strict#role=all:mode=unique:minMythicLevel=25:maxMythicLevel=99
Your statement is correct: there are more unique Enhance shams than Rets in Season 2 at the highest bracket. 50 Enhance vs 41 Rets at +25 or higher (no data for +30 and up) looking at unique DPS class population. Hereâs my question: Is that difference meaningful? Is looking at the âhighest levelâ actually meaningful? With the sample being 598, Retâs makeup 6.8% of the sample and Enahnce make up 8.4%.
Now consider this: https://raider.io/mythic-plus-rankings/season-df-2/all/world/leaderboards-strict#role=all:mode=all:minMythicLevel=20:maxMythicLevel=99
Same thing but looking at +20 and higher bracket theres 23,379 (12.3% of the sample) Rets vs 7728 Enhance Shamans (4.1%).
Huh? My statement is faulty. But then you link people to a graph that shows 50 Enhance and 41 RetsâŚwhich is exactly what my statement just was.
Yes! Which is what I have been saying. Enhance representation as the difficulty went up increased in prior seasons. This data was only brought into this discussion because people said âwell Ret is just always more popular than Enhanceâ.
We just agreed on thisâŚbut somehow you agree with me, call my statement faulty, then provide data that supports my statement.
My opinion is: This season, brought to light, certain fundamental flaws that exist in certain specs - relative to their defensives. When you canât count on one hand certain specâs defensives, but other specs have only one, thatâs an issue. Itâs called defensive creep. That said - Hero talents were released - and with some classes (who do not suffer defensively) received more defensives in their Hero trees. That, again, is an issueâŚthat needs to be talked about and addressed now before another season sidelines the class that I love.
Iâm honestly bowing out. Itâs been fun, but thereâs really not a lot more to discuss here. Iâve said my piece. I canât do it anymore. Iâve been called dumb, Iâve been told I should get fired from my job. Iâve been repeatedly mocked. And Iâve done nothing but try to help. But at the end of the day, I donât know how âworth itâ it has been. Iâm out.
It is but the implication you are attempting to make is meaningless when you actually look at the data and the other external factors that lead into the data.
Also gonna throw this out there: Looking at the very best of the best of the best is generally a terrible way to attempt to determine how to balance a game in some respects. Quite literally balancing the game around the top 1%.
Weâve been saying that and because its broken.
Itâs not, youâve again just been posting numbers and doomsaying about something that is largely irrelevant. We told you why the numbers are like that and that ultimately its meaningless in the grand scheme of things but that doesnât validate your hatred of Ret paladin for some reason.