Thinking about next server, I decided to check the rumour that dwarfs are stronger than elves. What I found is so broken, that I think it should be shared.
TLDR (Summary) if you are lazy to read:
19.07.2020 UPD: there have been changes introduced to the game since this post was published. I believe that both races are now balanced
Dwarfs have 2% better units by power (effect of inherent higher initiative) [Need to confirm, but effect seems to be reduced by units changes in S3 patch] Dwarfs have preferable matching mechanics early-mid game, and equal in endgame. [does not seem to be the case any more with S3 patch] Dwarfs have PVP advantage from 9% (early game) to 3.5% (endgame) due to masteries. [found error in calculations. difference is 1% only in dwarfs favour] Dwarfs have 11% better economy bonus due to masteries (construction time) [balanced by updates to patch S3]
Here are some conditions I set, to make the comparison manageable:
* PVP strength is determined by slot size multiplied by army strength plus any modifiers, that is why I analysed 3 aspects: 1- Warlords (Speed, morale, slot size); 2 – units; 3 - mastery points.
* Anti-goblin troops are ignored since they don’t affect pvp.
* production cost and time are ignored (it is important to balance them. But they do not impact whether someone wins or loses a fight)
* I used new values for units proposed here. (what we played with on this server was more imbalanced) 💬 Unit and Race balancing - Prepare for DevTalk
1. Warlords analysis (WLs).
Dwarfs have more spread out speeds, but on average they are equal – 36.
Attached graph to see how it changes from early to later WLs.
On average it is 48.
SLOT size and base POWER: equal.
Slot size has to be adjusted for units strength, because 500 infantry is not the same as 500 cavalry. To calculate power, I multiplied slot size by average unit strength for each race. WL strength is higher for high-level WLs, but comparing Dwarfs and Elves, they are equal. 122.8k per WL average.
I have much more data, but it is not useful to share all. E.g. I analysed slots preference over time, units’ preference over time, but it is insignificant in this topic so I am excluding them for the sake of clarity.
== CONTINUE ==