I've been using the phrasing "theoretical math" because that is the phrasing that
@Silly Bubbles has been using in expressing his position, and I've been objecting to that position. A gracious assumption is that he refers to it as "theoretical" math because the formula that is generally understood to account for the impact on fighting ability accounting for research progress and AW levels, etc, has been derived through experimentation, and so is a best guess and not definitive based on knowing what the actual coding implementation is. I admit I don't know the details, but I believe that the effort that has gone into attempting to determine the fighting formula correlates with the observations fairly well. To repeat your phrasing, I do believe that the formula currently, the "theoretical predictions" have been "tested against reality to help verify those assumptions." If my gracious interpretation does agree with his usage, I disagree with his belief that the formula is not accurate or precise; I think it has been worked out to a pretty high agreement with reality. A pessimistic assumption is that he refers to it as "theoretical" math because he doesn't like the approach in general, possibly because he doesn't agree with the conclusions, and so is trying to express derision at the approach. Personally, I think it is at least partially the later, and represents a libel or slander at attempting to analyze the question with math; this will possibly make some people think I'm trying to be personally insulting again when I am trying to describe the argument and not the person, but I view the above use of the phrase "theoretical math" as being in the same vein as creationists referring to "evolutionism" and "scienters" or flat earthers referring to "globists"." Theoretical math" is being used in a negative way similar to references to gravity being "only a theory". If my pessimistic interpretation does agree with his usage, I object this characterization.
As for Godel applying in this situation, I again express my opinion that it does not apply in this situation. Whatever the actual implementation is, whatever the coding that is being executed, it is undoubtedly an implemetation of an overglorified game of war, a collectible card game conflict, a role playing game combat or other similar activity, a complex but fixed set of rules. At it's core, like war, it will boil down to who has the bigger number, expressed as a combination of squad size, unit health, damage, defense, etc, possibly requiring multiple rounds before determining the final outcome.. Like a CCG, there are modifiers on top of those base numbers; particular troop proficiencies, military improvements, AW levels, special buildings, fed pet effects, etc. And like an RPG, there is an involvement of an RNG. It's a complex process; there are many pieces. But, each piece is a fixed piece. Each piece can be explicitly and exactly stated. If known what the code is, each piece can be evaluated for what it's impact is, relative to all the other pieces in the algorithm. This produces knowable results. Understanding the Elvenar Fighting formula is not a Godel Incomplete problem.
In college, I played a card game called Mao. The dealer would secretly make up a set of rules, players would take turns playing cards and trying to make the correct reaction to the played card, and if they made the wrong reaction, they would be penalized by being given additional cards. The rule might be simple, like "state whether the played card is odd or even" or "state how many cards are now in your hand". The rule might be more complex, as in "state the opposite of the color of the previously played card" or "state three more than the number of cards in the hand of the player to your right". The rules might be really complicated, like "touch your nose with your left hand if you played a red card, give a card to the person on your left if you played an even card, take a card from the player on the right if you matched suit with the previously played card". If you demonstrated successful following of the rules for several rounds, you could secretly tell the dealer what you think the rule(s) is/are to see if you are right. The game would go to who guessed correctly first. Our situation with working out the Elvenar Fighting algorithm is a game of Mao with a complex set of rules, and the inability to ask the dealer if our guess at understanding those rules is correct. We can't know if we have gotten the correct answer, even if we are reasonably reliable in predicting average outcomes over many encounters after changing a game state, such as upgrading an AW or completing a Research. In this way, the answer is ultimately unknowable. But, it is not unknowable because it is Godelian Incomplete. The Elvenar fighting process does have a fixed set of rules. There are a fixed number of inputs, being combined in a fixed process, and incorporating random elements in a fixed way; if we could see the code, we could select a set of values for all the inputs and do all the calculations ourselves, and with those selected set of values we could compute the impact of the RNG elements, and thus we could get an exact knowledge of the entire process. If we knew the code, we could know the answer. That we don't know the code, and thus can't be certain we know the answer, does not make this a Godel Incomplete situation. For a game of Mao to be Godel Incomplete, the dealer would have to choose a set of rules that, because of very specific circumstances, the dealer would not know what the proper response is under for at least some play . The rules would result in a particular type of unresolvable question. To the best of my understanding, the Elvenar Fighting algorithm does not compute itself into a mathematical Godelian dead end.