Trotterdam wrote:Worth noting is that my site is designed to be "fail-broad": while glitches might cause it to report something can happen when in fact it can't happen or is only very rare, it is much less likely for it to report that something can't happen when in fact it can happen, even if only in relatively outlying nations. Most issue options have hundreds of data points, meaning that the probability of any nation getting a result that is outside of the listed range is extremely low, and the probability of being more than a tiny fraction of a point outside of the listed range is even lower.The New California Republic wrote:But then again the issue effects often can and do vary depending on the nation that has the issue, so the issue effect on the third party sites are there as a rough guide, they aren't intended to be totally accurate in terms of what will happen stats-wise.
Most claims of "inaccuracy" are from people who don't understand the concept of minimum and maximum changes, and conclude that getting a result which is in the opposite direction from what I list as the average change means my data is wrong, even when it explicitly said that could happen.
Yes, I was more meaning that each issue isn't going to give a fixed precise number for stat changes when answering an issue, a fact which can sometimes confuse the newbies. For example option 2 of 1,202 isn't going to increase culture by, say, 5% exactly each and every time. That's more what I was meaning when I said that the third party guides can't be totally accurate.
Trotterdam wrote:Not always. Some of the listed policies are psuedo-policies that you get for having a numerical stat within a certain range, so an issue option will only sometimes assign/remove those policies depending on how close you were to the threshold before. My site also aims to accuracy list this variance by describing the policy changes in question as "sometimes" happening, and still tries to apply the fail-broad principle here, although there is somewhat more risk of it failing by something that can theoretically happen not being reported due to being rare enough to have never been sighted (for example, the site does not currently report #6 3 as removing the Theocracy policy even though it is known to do that, because it is nearly unheard of for theocratic nations to choose this option to begin with). Stats, which have a range of effects rather than being all-or-nothing, suffer less from this.The New California Republic wrote:The only thing that can be said with any reliability is what policies an option will enable or disable, as those are set in stone.
Yes there are some that can, I was meaning the ones that are hardcoded into the issue itself as absolutes, rather than the range-based ones that tip over.
