Balanced win rate in Civilization over time

Age of Empires III: DE III - Discussion
1

Title: Hidden Balance Issues May Be Missed by Global Win Rate Statistics

I would like to suggest a more detailed approach to civilization balance analysis.

A civilization can appear balanced when looking only at its overall win rate, while still being significantly overpowered or underpowered in specific game states.

For example, instead of analyzing only total win rates, it may be useful to segment statistics by:

  • Civilization

  • ELO bracket

  • Match duration

  • Highest Age reached

  • Game mode

A civilization with a 50% overall win rate could still have:

  • 65% win rate in games that reach Age V

  • 35% win rate in games ending before Age III

The overall average would hide this imbalance.

From my personal experience, many matches feel competitive and interesting while players remain in earlier Ages. However, once games reach the later Ages, certain civilizations seem to become disproportionately strong.

I believe balance patches could benefit from tracking win rates separately for different game durations and Age progressions. This would allow hidden late-game or early-game imbalances to be identified more accurately.

My suggestion is not necessarily to target a perfect 50% global win rate, but to reduce extreme deviations within different stages of the game.

Thank you for continuing to support Age of Empires III.

2

Firstly, welcome! I see this is your first post on here.

I do have a couple disagreements with your post however. Firstly the game was built with asymmetry in mind, different civs are supposed to have different stronger and weaker points during the game. Except for DE ottomans who for some reason are basically op all the way through :confused:. A civ like Portuguese, for example, is weaker, even trashy, in the early game and stronger later. It’s not really fair for civs that are strong in the early-mid game to not be meaningful weaker in the late game.

Secondary and more importantly, WR is an iffy way to balance things. The reason for this is that the elo system is designed to push players towards a 50% WR. Things have to be really bad to overcome this by a significant amount. What should most often happen is a player with (lets say) 1300 elo might gain ~50-100 elo playing an op civ or lose the same playing an underpowered civ. Actual differences will vary of course. That 1300 skill player will eventually reach a point where they level back to a 50% WR vs players who are better or worse than they are based on the lost or gained total of skill + civ strength.

2 Likes
3

As well as what kingfisher said, the sunbros attempted this at some point and at the time we just didn’t have enough data to really get good power out of various analysis. I believe some graphs are still updated when data is entered but generally, sans the top5 played civs, too many matchups don’t have enough matches to make great data. I’d imagine length of time to win compared independently is the only metric we could really parse out, and again only for the more popular civs

3 Likes