Weights are not uniform between test cases. May or may not have been confirmation bias.
- 1
Advisors + RNG = Headaches
- Thread starter Abnormalmind
- Start date
-
We have updated our Community Code of Conduct. Please read through the new rules for the forum that are an integral part of Paradox Interactive’s User Agreement.
Did you fire the advisor and not retire him? because of course that gives you the same guy. You are returning him to the pool.
Ok so I just ran a second test with a coastal OPM. I retired one advisor of each category and obtained different advisors next month. If the weights were the same I should have obtained the same advisors.
However I believe the person I quoted didn't retire but only returned the advisors to the pool (see previous comment). So yeah thats not gonna work because its not intended to work that way.
- 3
Oh yeah. If that's what's happening, it's silly. But you could repeatedly hire and fire the same guy. That's not really confirmation bias either though. That's...something else entirely, lol.
- 1
I mean you can't even really call EU4's advisor mechanics "a system", given how underdeveloped it is.
It's funny though because cultural tradition was very much a mana-like construct of EU3, now in EU4 practically every DLC has added some kind of "bar go up till you press a button to get a reward" mechanic, but needing to get specific advisors is still a one way ticket to pain and despair. It could be a good money sink to expand the system and have more rare advisors that require infrastructure investment to generate.
- 1
- 1
The lack of statistics understanding this comment betrays just further underscores the likelihood that what you experienced was confirmation bias.
It's rarely worth it to cycle advisors for a specific type, unless there are 2-3 different ones that would all be great.
For a fully coastal nation pre-colonists, you'll have almost a 50-50 to get a specific type by your third try and almost 80% chance after 7 attempts. That's a lot of gold though, for something that still may not work out. I personally just pick the best available option instead.
For a fully coastal nation pre-colonists, you'll have almost a 50-50 to get a specific type by your third try and almost 80% chance after 7 attempts. That's a lot of gold though, for something that still may not work out. I personally just pick the best available option instead.
- 2
You know the answer about priority advisors and the game's advisor-to-court mechanics? I'd be happy to learn. Please point me to the best resource to educate myself.
P1 A modifier exists, therefore
C1 it is not truly random, therefore
C2 I am not experiencing confirmation bias
Does not follow. It is entirely possible for the premises to be true and the final conclusion false: a system could be semirandom or even not-at-all random and you could still experience confirmation bias. What you have presented is, as a high school graduate in any country with a functional education system could tell you, an invalid argument.
There may well be a modifier. Nonetheless, as independent experiments in this thread have demonstrated, it is wholly possible to get different advisors when firing, adjusted by weighting. The existence of a modifier has no bearing on whether or not you’re experiencing confirmation bias, QED.
- 2
- 1
Okay thanks. Given that there exists at least one unknown variable, concealed to players, we can conclude that any forthcoming answer is a theoretical answer at best. Given that the "independent experiments" did not display any statistical evaluation methodology, did not disclose the value (or game-state) of the (at least one) hidden advisor-to-court value, and did not include any potential other game-state conditions -> C1 and C2 may or may not be valid. Unless, you have a resource which clearly explains how the advisor system is calculated?
That is not how validity works. Things can’t “may or may not be valid”. They are or they aren’t. Validity is when the premises and conclusion are related in such a way that if the premises are true, the conclusion cannot be untrue.
C1 it is not truly random therefore C2 I am not experiencing confirmation bias is an invalid argument. The premises are not related to one another.
Advisor pools could be 100% scripted and you could still experience confirmation bias. You could be right and experiencing confirmation bias. C1 is irrelevant.
I only point this out because understanding the proclivity for confirmation bias is important in the modern information environment in general. I don’t really care about advisor pools.
- 3
I absolutely agree that it's best to be aware of any potential confirmation bias. Nevertheless, we shan't arrive at a statically significant result without statistically relevant data. The best arrival point: "observation" of <behavior> at [given current game-state]
Given this game-example, confirmation bias is purely subjective. Thanks
How are you still dying on this hill?
Confirmation bias can’t be subjectively occurring. Someone either is or is not experiencing confirmation bias.
The argument you’re presenting is:
P1 there is some unknown modifier or modifiers, therefore
C1 it is possible that said modifiers were coinciding such that only one possible advisor type could be available, therefore
C2 it is possible that a person who reports firing and receiving the same advisor type over and over is not experiencing confirmation bias.
That’s a frivolous claim to argue to this extent: possibility isn’t interesting or meaningful; likelihood is worth discussing. Notwithstanding it’s meaningless, I’m happy to concede it’s perfectly true.
But the original claim is that “it seems like you always” get the same advisor you fired “over and over”. Not in any specific case, but always. Not “you might possibly sometimes”, always. This is plainly untrue (and I’m sure conscious hyperbole on the part of the author), and two tests have now been run demonstrating that it’s untrue.
The possibility that as you’ve exhaustively and pointlessly argued it might sometimes occur that this happens does not imply that you “always” get the same advisor you just fired “over and over”. It implies that this might (but also might not) sometimes occur. We don’t need a statistically significant event record to demonstrate that the user claiming an event “always” happens is wrong (and probably not being serious); they are experiencing confirmation bias.
- 2
Its not hard to be rich enough where cycling a level 1 advisor is pretty cheap. Essential advisors include:
+improve relations (basically always desired because AE is a bitch)
+trade efficiency and +inflation loss (get the +200/+200 mana event for this combo early)
+morale (best general combat advisor by far)
+siege defense (if you're in some tough start where you really need to siege race or do something tricky with forts).
+conversion rate (converting provinces at 2% per month is infinitely faster than 0% per month)
- unrest (in the right circumstances this +army patrolling can entirely stop rebels from spawning, thousands of manpower saved).
The other good thing is that when you hire immediately after spawning them they tend to be pretty young and will last a long time.
Last edited:
- 1