The following applies to the game played on PC/Steam only. I can’t attest for the numbers on other versions of the game. Tests were run on August 3 at about 12 p.m. PST over the course of a < 10 minute span.
I used this team on purple day to run some tests:
I had a good starting board every single game, so I was able to use the Mercy opener, which is why this data is so consistent. I won every single game easily. I took three damage in one game from a Famine trait, and zero in all others - my damage ratio is maxed out. I allowed 7 mana to be collected one game, and 0 every other, while collecting at least four times that in every game - my mana ratio was most certainly maxed out. My troops weren’t scratched except from a single famine hit on life, so I didn’t lose or summon any troops, so my troop bonus had to be maxed out. Total actions taken were 7, 10, 9, 11, and 10, respectively. Total scores were 1329, 1320, 1323, 1317, and 1320. If the formula is correct as given, we should be able to reasonably conclude that B% is 1% - in other words, every action less than 30 is a bonus of 1% to your base score. We know that all the bonuses are additive now, because it is spelled out in the original post. Therefore, we can stop thinking about bonuses in percentage terms and start thinking about them in terms of total points. Each action under 30 actions taken is worth 3 points each. This is a potential of 87 total points, (since it would be impossible to win without taking an action).
We have also been told that color bonus calculation has not changed. We know from before the changes that each correctly colored troop is worth 50% of the base score (with level 100+ statues), therefore each correctly colored troop is worth 150 points, for a potential of 600 points.
We are given 300 points for a win, displayed right there in the GW screen, so we already know that.
Extrapolating from my data set and assuming the formulas given in the original post currently are correct, on the battles where I earned 1320 points, I earned 450 points for correct color troop colors, 300 for winning, and 60 for speed. That is a total of 810 points accounted for. For the other bonuses, which according to the criteria given I should have been given perfect marks according to the criteria as outlined, that leaves 510 points for remaining troops, damage ratio, and mana ratio. Given the consistency of my scores, I’d say I did earn perfect marks here.
While everyone else is trying to figure the ratio between those, I’d like to point out that under this formula with the information given, the maximum possible achievable score for any given battle should be 1497 points (300 base + 600 colors + 510 survival/damage/mana + 87 speed). Since every bonus is additive, and every bonus has a cap, every score should fall completely within this model. With the formulas and criteria all given, its just basic algebra at that point. Anything above 1500 points does not fit the model given and is anomalous.
Given that we’ve seen scores higher than that reported after Tuesday, either my math is incorrect (could be, so correct me if I’m wrong) my data is incorrect (I recorded four of five of the games I got this data from and watched them several times, there were almost no variables given I barely let the AI do anything, highly unlikely), something in the initial data is incorrect (the formula itself, the assumption that color bonus calculation are unchanged, etc.), another assumption is incorrect (there are different calculations in different versions of the game and the scores we reported are from other versions, unbounded scoring or min/max reversal bugs still exist in other versions, etc), or there is still a bug in the calculation somewhere.
I’d be interested in any recordings that could be provided of games where the user scored over 1500 points and/or what platform these users played those games on.
Overall, what I have to say about this whole thing that I’ve been mostly silent on so far is that all these tweaks and data collecting are costing dev time to get people to like a feature that they may not fundamentally like in the first place just because they feel obligated to play as to not “let down” their guild and because an exclusive reward is locked behind it. If Guild Wars troops can be reasonably mythiced through crafting and when it is no longer basically required to get a mythic on exclusive week to have an even close to reasonable chance to obtain it, many people wont care what the scoring criteria is because Guild Wars will be moved into the optional category and played only by people that actually enjoy it and not because it is tied heavily into the main collection mechanic of the game. I know a lot of time was spent on guild wars in the first place, but some people are just not going to like it - fundamentally. It was never going to be a something for everyone mode. Attempts at greater mass appeal and trying to stir up some variety by obfuscating the scoring system seem to instead be frustrating people on both sides of the fence while doing absolutely nothing for variety (and certainly not long term variety). I’m hoping we can just fix up all the actual bugs here and move on to crafting, hopefully getting those that like guild wars to play it, and giving those that don’t other options to reach the same collection goals (albeit with probably considerably more effort).