Showing "Advantage" In-Game/End-Game

[Sapient]

Here's a possible idea for more in-game Statistics. The original idea came from eddieballgame in this thread (although I dislike his Performance Rating idea).

The equation would look something like:

Code: Select all

Advantage = Income Factor + Holdings + Other Advantage Factors
Income Factor = Net Income * INCOME_WEIGHT
Holdings = Gold + ALL( Unit Value )
Unit Value = MAX( price*(%hp), Next Level Value*(%exp) )
Other Advantage Factors = {undefined}

Essentially, it would be the computer's idea of "Who's winning?"

It might be nice to display Ending Advantage if turns run out in MP.

[Temuchin Khan]

Here's a possible idea for more in-game Statistics. The original idea came from eddieballgame in this thread (although I dislike his Performance Rating idea).

The equation would look something like:

Code: Select all

Advantage = Income Factor + Holdings + Other Advantage Factors
Income Factor = Net Income * INCOME_WEIGHT
Holdings = Gold + ALL( Unit Value )
Unit Value = MAX( price*(%hp), Next Level Value*(%exp) )
Other Advantage Factors = {undefined}

Essentially, it would be the computer's idea of "Who's winning?"

It might be nice to display Ending Advantage if turns run out in MP.

I like this idea. It has always seemed a little odd to me that someone can be trampling his opponent's units in multiplayer, but still be counted a loser if time runs out.

[AT]

I've always wanted a more comprehensive set of statistics to be generated (at least) after the game is over. I'll see if I can come up with something in a patch once I get back up to speed.

[AI]

I have started implementing this feature. It can be found in trunk using the default era.

Income currently has a weight of 5 and unit value is simply the cost of the unit. A more complex (and configurable) unit worth formula can be found in the test scenario. (right click any unit)
The current formula there is:
(BASE + (hp/max_hp) * HEALTH_WEIGHT + (xp/max_xp) * XP_WEIGHT)
where the default values are 0.5, 0.5 and 2.

If anyone can come up with a better formula or better (default) weights, I'd love to hear it.

[Sapient]

ummm... I like the formula that was proposed a lot better:

Code: Select all

Unit Value = MAX( price*(%hp), Next Level Value*(%exp) )

(NextLevelValue would be the highest price from the available advanceto choices)

[AI]

I considered that, but that would require accessing the cost of a unit type by its name. Not an easy task from WML. (I'm already abusing the fact that a [unit_type]'s variables are carried over into it's [unit] instances)

This would require something along the lines of:
-parsing advances_to
-creating a unit of each type
-accessing its cost

[AI]

I've moved the implementation to C++ and changed to your formula, the formula's innards are exposed in the test scenario.

The UI has also been polished (mostly Soliton), so I think it's about done.

[Sapient]

I've moved the implementation to C++ and changed to your formula, the formula's innards are exposed in the test scenario.

The UI has also been polished (mostly Soliton), so I think it's about done.

Very Nice!

thanks