Balance Mod to test changes.

[Boudica]

23 minutes ago, Nescio said:

[EDIT] Oops, wrong thread. @Stan`, please move.

For future reference:

armour level , damage taken , health equivalent
 x           , 0.9^x        , 1 / (0.9^x)
 0           , 1.000000     ,   100%
 1           , 0.900000     ,   111%
 2           , 0.810000     ,   123%
 3           , 0.729000     ,   137%
 4           , 0.656100     ,   152%
 5           , 0.590490     ,   169%
 6           , 0.531441     ,   188%
 7           , 0.478297     ,   209%
 8           , 0.430467     ,   232%
 9           , 0.387420     ,   258%
10           , 0.348678     ,   287%
11           , 0.313811     ,   319%
12           , 0.282430     ,   354%
13           , 0.254187     ,   393%
14           , 0.228768     ,   437%
15           , 0.205891     ,   486%
16           , 0.185302     ,   540%
17           , 0.166772     ,   600%
18           , 0.150095     ,   666%
19           , 0.135085     ,   740%
20           , 0.121577     ,   823%
21           , 0.109419     ,   914%
22           , 0.098477     ,  1015%
23           , 0.088629     ,  1128%
24           , 0.079766     ,  1254%
25           , 0.071790     ,  1393%
26           , 0.064611     ,  1548%
27           , 0.058150     ,  1720%
28           , 0.052335     ,  1911%
29           , 0.047101     ,  2123%
30           , 0.042391     ,  2359%
31           , 0.038152     ,  2621%
32           , 0.034337     ,  2912%
33           , 0.030903     ,  3236%
34           , 0.027813     ,  3595%
35           , 0.025032     ,  3995%
36           , 0.022528     ,  4439%
37           , 0.020276     ,  4932%
38           , 0.018248     ,  5480%
39           , 0.016423     ,  6089%
40           , 0.014781     ,  6765%
41           , 0.013303     ,  7517%
42           , 0.011973     ,  8352%
43           , 0.010775     ,  9281%
44           , 0.009698     , 10312%
45           , 0.008728     , 11457%
46           , 0.007855     , 12730%
47           , 0.007070     , 14145%
48           , 0.006363     , 15717%
49           , 0.005726     , 17463%
50           , 0.005154     , 19403%

Because armour is a power law, differences in armour levels are more consequential then they might seem. Take, for example, the house; the tooltip shows us: Armour: 20 hack (88%), 30 pierce (96%), 3 crush (27%), which although true is somewhat misleading. The difference in hack and pierce is 10 levels, between hack and crush 17 levels, and between pierce and crush 27 levels. Another way of phrasing this is that 1 crush damage is equivalent to 6 hack damage or 17.2 pierce damage.

I doubt many people would have guessed this from the current tooltip, so maybe we should have a tooltip that displays armour-health equivalents (i.e. the total amount of attack damage it can take).

I don't know what thread this should have been in but thanks for making this table anyway. I've done these computations many times in my head. Also I doubt that there are many people that have an idea what the armor levels means. Even the basic explanation that an extra armor level blocks 10% of the damage could help. This system is much better than simple subtraction, which forces you to make a minimum damage of 1 (I think that's how it was in AoE). However, it's harder to present it clearly to the player.

[ValihrAnt]

New version released, available on the first post.

Changes in this version:

Reverted all champion changes, due to extra metal mine spawns being entirely random. This means that very often 1 team will have a considerable advantage, due to having access to extra metal. I've even played multiple matches just in this balancing mod where one team had only 1 or no extra metal mines at all, while the other team had safe access to all other metal spawns. Also removed the champion tech unlock change.

Additionally I've lowered archer damage by 0.1P from 6.8 to 6.7, changed around War elephant cost to have higher food than metal cost and also slightly lowered catapult aoe damage.

Version 5.0 changes

Champion cost and unlock tech back to old cost
Citizen archers: 6.8P to 6.7P
War elephants: 225F and 225M to 300F and 200M
Catapults: Splash damage radius: 10m to 7m

 

[Nescio]

15 hours ago, ValihrAnt said:

also slightly lowered catapult aoe damage.

Catapults: Splash damage radius: 10m to 7m

If I recall correctly (@fatherbushido?), splash damage is spread out over the splash area, so a lower radius means the same total amount of damage is concentrated in a smaller area. From 10 to 7 means that the area is half as large, so units inside take double the damage (π*10^2 / π*7^2 = 2.04).

[ValihrAnt]

Quote

If I recall correctly (@fatherbushido?), splash damage is spread out over the splash area, so a lower radius means the same total amount of damage is concentrated in a smaller area. From 10 to 7 means that the area is half as large, so units inside take double the damage (π*10^2 / π*7^2 = 2.04).

Thank you for pointing it out. Looks like I'll have to experiment with the numbers more to see what works best.

 

[fatherbushido]

5 hours ago, Nescio said:

If I recall correctly (@fatherbushido?), splash damage is spread out over the splash area, so a lower radius means the same total amount of damage is concentrated in a smaller area. From 10 to 7 means that the area is half as large, so units inside take double the damage (π*10^2 / π*7^2 = 2.04).

I don't know what's the current state of the svn, so I will refer to the 0AD a23b state.

It's always pleasing to see people who don't do random things!

Indeed it would be more realistic to distribute the effect. What the code do is slightly different.

EDIT: I didn't see it like that, but that's the same I guess.

There is a multiplier of the damage with a quadratic falloff in the distance from the impact.

Looking from above, the damage multiplier reduction looks like that:

 

Again, those numbers are mainly for emergent behavior, not for direct gameplay control.

For example:

Damage multiplier in function of the distance between the impact and the target (in blue for radius=10, in red for radius=7). I plotted the left side because it's nicer:

 

It's also surprising to see attempts of fine tuning such a number whereas there are many broken things in the splash feature code. @ValihrAnt is doing nice though (and he makes awesome videos)!

 

Edited January 18, 2020 by fatherbushido
typos, always typos + edit

[Nescio]

45 minutes ago, fatherbushido said:

I don't know what's the current state of the svn, so I will refer to the 0AD a23b state.

Now suppose an unit is located at a distance of 5 from the point where the splash attack hits. How much damage would it take if the splash radius is 10? And how much if it's 7?

[fatherbushido]

In theory:

The damage is multipilied by

1-(5/10)**2 when the radius is 10

1-(5/7)**2 when the radius is 7

In every case:

damage is full at origin

damage is multiplied by 3/4 when at the half radius

damage is 0 at full range

 

Edited January 18, 2020 by fatherbushido
(I need english lessons)

[Nescio]

25 minutes ago, fatherbushido said:

In theory:

The damage is multipilied by

1-(5/10)**2 when the radius is 10

1-(5/7)**2 when the radius is 7

In every case:

damage is full at origin

damage is multiplied by 3/4 when at the half radius

damage is 0 at full range

 

Thanks for the clarification. Basically damage to an unit is 1 - (d/r)^2, with d the distance to the splash centre and r the splash radius?

How does that work out for the total damage over the entire splash area?

Also, only the unit centre matters, the footprint radius doesn't, right?

[fatherbushido]

58 minutes ago, Nescio said:

Basically damage to an unit is 1 - (d/r)^2, with d the distance to the splash centre and r the splash radius?

Exactly that.

58 minutes ago, Nescio said:

How does that work out for the total damage over the entire splash area?

I am not sure how we have to interpret it in the game context. Units and structures are rectangle and apart units in formation, they don't overlap. So you will have only a finite numbers of damages in that area in the game context. Out of the game we want the (continuous) sum over the whole disk, but really I think it's not relevant.

EDIT: writing that, I edited one post above.

58 minutes ago, Nescio said:

Also, only the unit centre matters, the footprint radius doesn't, right?

Yes, for the splash part.

In fact it's even worse than you can expect. But not enough place in that margin.

Edited January 18, 2020 by fatherbushido

[Nescio]

16 hours ago, fatherbushido said:

Yes, for the splash part.

In fact it's even worse than you can expect. But not enough place in that margin.

Attack ranges are from centre to edge, thus that means that if an unit with no base damage and only splash damage would attack a structure which has a footprint radius larger than the unit's splash radius, the structure won't be damaged by the unit's splash attack?

[fatherbushido]

14 minutes ago, Nescio said:

thus that means that if an unit with no base damage and only splash damage would attack a structure which has a footprint radius larger than the unit's splash radius, the structure won't be damaged by the unit's splash attack?

Indeed.

Edited January 19, 2020 by fatherbushido
I am speaking of a23b, I didn't follow what happened since

[ValihrAnt]

Updated the mod to most likely the last version.

VERSION 6.0 CHANGES

Actually changed archer damage to 6.7P from 6.8P
Fixed elephant archer speed, as I originally didn't realise they had cav archers as parent template.
Carthaginian temple cost from 400S to 300S
Kushite small pyramid cost reduced from 300F 300S to 300S

The Carthaginian temple is the only temple not to have a resource cost of 300, I have changed it to fit with the rest after some requests. Is the higher cost intended or is it an oversight that should also be fixed for a24? Additionally I have changed the small pyramid cost to just 300 stone from 300 food and 300 stone. Should make them a bit more accessible and hopefully more used.

[Nescio]

2 hours ago, ValihrAnt said:

Carthaginian temple cost from 400S to 300S

The Carthaginian temple is the only temple not to have a resource cost of 300, I have changed it to fit with the rest after some requests. Is the higher cost intended or is it an oversight that should also be fixed for a24? Additionally I have changed the small pyramid cost to just 300 stone from 300 food and 300 stone. Should make them a bit more accessible and hopefully more used.

It's probably because Carthage trains their champions at their temple. (Whether that's a good thing is a different discussion.)

2 hours ago, ValihrAnt said:

Updated the mod to most likely the last version.

Don't be so sure.

[ValihrAnt]

Quote

It's probably because Carthage trains their champions at their temple. (Whether that's a good thing is a different discussion.)

Kushites also train champions in their temple, but the temples still cost 300 stone. Though, I suppose it is 1 champion vs 2 champions.

[wowgetoffyourcellphone]

It's so goofy you guys getting so granular as to change a stat from 6.7 to 6.8.

[sarcoma]

I hope for a24 kushite heros are moved before champions in fort. They used to be from cc and when they moved them to fort they got the order wrong.

[Feldfeld]

5 hours ago, wowgetoffyourcellphone said:

It's so goofy you guys getting so granular as to change a stat from 6.7 to 6.8.

Well, in a game like this, a very small change can have big consequences, especially when there is no counter system, the stats have to be near perfect.

[Nescio]

3 hours ago, sarcoma said:

I hope for a24 kushite heros are moved before champions in fort. They used to be from cc and when they moved them to fort they got the order wrong.

In A24 heroes are actually always listed last, regardless where they are trained (centre, fortress, apadana, prytaneion, syssition).

Speaking of which, any opinions on D2580 (display centre before house)?

[ValihrAnt]

New version. Changes include:

Removed Promotion animations
Spear cavalry: 6H + 5P / 2s to 4H + 3P / 1.25s, champ and hero spear cav adjusted in same manner.
Rams: 0.9 speed to 0.8 and can't attack organic units + fields anymore
Allow building Kushite merc camps in neutral territory

Mostly made this to play around with some proposed changes in multiplayer more easily.

BalanceTestModV7.zip