percentage based upgrades versus addition based upgrades

[BreakfastBurrito_007]

I have been considering unit upgrades in 0ad for a while now and I have come across a variety of ways that they are problematic. 

Issues with percentage based upgrades:

• boring: percentages are nearly the exclusive method for upgrades to units and there are interesting cases for addition based upgrades
• imbalance: units that do higher damage per second are benefitted more by the same 1.2x damage upgrade than lower-damage units. Even more concerning is that the gains in damage per upgrade grows with each upgrade, which means the discrepancy between lower damage units and higher damage units increases at an increasing rate. This certainly is one of the contributing factors for champion cavalry skirmishers (firecav/brit chariots) being so OP.
• Another example is indibil's cost reduction percentage giving a bigger savings for champions and cavalry more than citizens and infantry.

I know that eons of balancing efforts have been done with these percentage based upgrades in mind, which is evident in the quite good balance between skirmisher infantry, slinger infantry and archer infantry. Because of this, I think that rather than making a drastic overhaul of the blacksmith, we need to consider addition based upgrades in certain problematic spots or for future creative upgrades.

One option for the blacksmith could be to make the final attack upgrades a +1 rather than a percentage change, and then also addition based upgrades could be considered for certain civ bonuses, team bonuses and heros. I think for the development of the game it would make a lot of sense if addition based upgrades could work in sync with percentage upgrades by the use of order of operations from math.

Would you be interested in seeing more addition based upgrades in the game? please tell me what you think!

 

[real_tabasco_sauce]

I don't know if having a mixture of percentage base upgrades and addition based upgrades would be problematic.

[BreakfastBurrito_007]

I don't know either, but ideally it would not be problematic if implemented. There is a known framework for doing these calculations called order of operations, the current system maybe doesn't use it since nearly everything is percentage-based, but I think the capability to handle addition would be a major win for making interesting and balanced upgrades.

[Fabius]

Armour already functions like this adding +1 and +2 respectively. Attack would work just fine, the issue however is that raw attack point bonuses will interact with the percentage attack bonuses, eg if you add attack points and then do percentage bonus you will have a bigger final total than with just a percentage bonus. So that would need to be considered prior to tweaking.

[AIEND]

The percentage enhancement from experience points should be reduced. Now it is 20%. It should be reduced to 10%. At the same time, I think the data correction should be carried out for the ranged units. I can't understand why the javelinmen have so much higher damage than the pikemen.

[BreakfastBurrito_007]

30 minutes ago, Fabius said:

Armour already functions like this adding +1 and +2 respectively. Attack would work just fine, the issue however is that raw attack point bonuses will interact with the percentage attack bonuses, eg if you add attack points and then do percentage bonus you will have a bigger final total than with just a percentage bonus. So that would need to be considered prior to tweaking.

I think the best way would be to do the multiplication first, and then add to the result. This way, the added damage does not get grown by the percentage bonuses.

[Fabius]

53 minutes ago, BreakfastBurrito_007 said:

I think the best way would be to do the multiplication first, and then add to the result. This way, the added damage does not get grown by the percentage bonuses.

That seems a reasonable application. I like units gaining experience more than blacksmith upgrades since it provides a trade off for extra stats, which is more interesting than a blacksmith, so I would prefer that it is not nerfed into the ground. Potentially raw points could be used for the experience levels instead of damage percentages.

[Fabius]

1 minute ago, Fabius said:

That seems a reasonable application. I like units gaining experience more than blacksmith upgrades since it provides a trade off for extra stats, which is more interesting than a blacksmith, so I would prefer that it is not nerfed into the ground. Potentially raw points could be used for the experience levels instead of damage percentages.

This way you can tone down unit scaling for those that have high attack and you no longer treating units with a blanket 20% or whatever. Now you can tweak individual units without busting the others too. Overall it seems a good solution to your balance problem. The big question now is how much effort would it require to implement.

[wowgetoffyourcellphone]

Thing is, percentages are more intuitive to the player. Is +3 attack a lot? Is +1 speed a lot? Dunno. +20%, now that gives you a great idea.

[LetswaveaBook]

On 16/09/2022 at 8:11 AM, BreakfastBurrito_007 said:

imbalance: units that do higher damage per second are benefitted more by the same 1.2x damage upgrade than lower-damage units. Even more concerning is that the gains in damage per upgrade grows with each upgrade, which means the discrepancy between lower damage units and higher damage units increases at an increasing rate. This certainly is one of the contributing factors for champion cavalry skirmishers (firecav/brit chariots) being so OP.

I don't see it as an imbalance. If without upgrades unit A can beat unit B with 66% health remaining, then if both units get the same upgrades then unit A still wins with 66% HP remaining. So it stays the same.

 

[BreakfastBurrito_007]

@LetswaveaBook I was thinking within the context of ranged units killing other units. Doesn’t really matter what the other unit is, just that it’s hp is subtracted by enemy attacks. The rate at which skirmishers outpace slingers in the killing of the unit grows with each upgrade. And of course the rate at which firecav/Brit chariots damage grows is much greater than any other unit. 
 

Of course there are other things that affect balance like range, accuracy move speed, and acceleration. These values for the most part, do not increase with blacksmith upgrades. Those balance factors remain constants, so the increasing damage of the higher damage units swings balance in their favor. 
 

For the 3 citizen soldier ranged infantry units this is not a huge problem since balance has been refined each alpha primarily focusing on those units. But now we need to keep in mind the 1.2x damage does not provide an even boost in damage. To be frank, the balance state before and after upgrades is different.

Edited September 18, 2022 by BreakfastBurrito_007

[BreakfastBurrito_007]

7 hours ago, wowgetoffyourcellphone said:

Thing is, percentages are more intuitive to the player. Is +3 attack a lot? Is +1 speed a lot? Dunno. +20%, now that gives you a great idea.

for this you just need to look at the damage values in the tooltip. I would argue 1.2x seems simpler, but creates some balance issues. 

 

5 hours ago, LetswaveaBook said:

If without upgrades unit A can beat unit B with 66% health remaining, then if both units get the same upgrades then unit A still wins with 66% HP remaining

the hp percentage changes before and after upgrades.

[LetswaveaBook]

1 hour ago, BreakfastBurrito_007 said:

the hp percentage changes before and after upgrades.

Can you try to convince me that it is actually true? If the winning unit deals 2x as much damage per second, then the weaker unit dies in half the time. However in half the time, the weaker unit can deal 2*0.5 the original damage if the weaker unit has the same upgrades.

[BreakfastBurrito_007]

33 minutes ago, LetswaveaBook said:

Can you try to convince me that it is actually true? If the winning unit deals 2x as much damage per second, then the weaker unit dies in half the time. However in half the time, the weaker unit can deal 2*0.5 the original damage if the weaker unit has the same upgrades.

yes indeed, an increasing damage advantage will cause the winning unit to win with more remaining hp. The damage advantage increases due to the upgrades being percentage based.

Edited September 18, 2022 by BreakfastBurrito_007

[LetswaveaBook]

huh? 2*0,5=1?

[BreakfastBurrito_007]

46 minutes ago, LetswaveaBook said:

huh? 2*0,5=1?

yes that is correct, but I was confused about the situation you are applying that math to.

[real_tabasco_sauce]

@LetswaveaBook 

@BreakfastBurrito_007 is just talking about how percentage upgrade give more dps to units that start out with more DPS.

1.20 * 5 is less than 1.20 * 10

It is a difference, and it is something to consider when we balance units, but I don't think the forge techs should be changed and I don't think we should use +/- damage upgrades instead. We should definitely not use a combination of percentage and addition upgrades.

[chrstgtr]

6 hours ago, BreakfastBurrito_007 said:

@LetswaveaBook I was thinking within the context of ranged units killing other units. Doesn’t really matter what the other unit is, just that it’s hp is subtracted by enemy attacks. The rate at which skirmishers outpace slingers in the killing of the unit grows with each upgrade. And of course the rate at which firecav/Brit chariots damage grows is much greater than any other unit

That’s only true if no armor or health upgrades are taken to offset the attack dmg.

To be honest, this is really just a problem of units dying to quickly in late game, which leads to spam fights. I think the easiest way to fix this is to undo another of the a24 changes: health phases bonuses. Reinstating health phase bonuses would also offer the pure importance of forge upgrades and make phasing more important.  

Edited September 18, 2022 by chrstgtr

[BreakfastBurrito_007]

2 hours ago, chrstgtr said:

That’s only true if no armor or health upgrades are taken to offset the attack dmg.

To be honest, this is really just a problem of units dying to quickly in late game, which leads to spam fights. I think the easiest way to fix this is to undo another of the a24 changes: health phases bonuses. Reinstating health phase bonuses would also offer the pure importance of forge upgrades and make phasing more important.  

:     o

lmaoo I feel a bit embarrassed now and I hope all of you can forgive me. sure enough just as attack upgrades increase damage "unfairly" for higher damage units, armor upgrades remove damage as a percentage. For example archers: 6.7 x1.7(attack upgrades) times (1-.41)= 6.7. Firecav 35*1.7*(1-.41)=35. 

I suppose then only hero bonuses, wtf, and relics are sources of damage that don't always have armor to compensate.

 

Edited September 19, 2022 by BreakfastBurrito_007

[LetswaveaBook]

17 hours ago, BreakfastBurrito_007 said:

sure enough just as attack upgrades increase damage "unfairly" for higher damage units,

Attack upgrades do not increase the damage unfairly for higher damage units, regardless of armor. Obviously+1 pierce attack has more effect on an archer than on a Briton Chariot. For the chariot +1 attack is negligible. @BreakfastBurrito_007, What would in your opinion be a "fair" (such that comparative strengths remain equal) increase for the chariot if the archer had an upgrade providing+1 pierce damage?

[chrstgtr]

1 hour ago, LetswaveaBook said:

Attack upgrades do not increase the damage unfairly for higher damage units, regardless of armor. Obviously+1 pierce attack has more effect on an archer than on a Briton Chariot. For the chariot +1 attack is negligible. @BreakfastBurrito_007, What would in your opinion be a "fair" (such that comparative strengths remain equal) increase for the chariot if the archer had an upgrade providing+1 pierce damage?

You guys are talking past each other. @BreakfastBurrito_007 had a problem with % increases because a higher base dmg will benefit more from a percent-based system. Here, you are discussing an addition based system, which comparatively favors lower base dmg units. 

Armor % increases offset attack % increases.

Without armor increases % based attack bonuses can lead to large differences.

These numbers are simplified, but they show the effect discussed. Imagine two units with 100 health and no armor. Unit A deals 50 dmg. Unit B deal 33.3 dmg. With no upgrades unit A kills in 2 hits while and unit B kills in 3 hits. In other words, unit A kills 50% faster. But with a 100% attack upgrade, unit A kills in one hit while while unit B kills in 2 hits. In other words, with upgrades, unit A is kills 100% faster. Now, imagine both units have armor/health upgrades that doubles their increase resistance. With all upgrades, unit A again kills in 2 hits while unit B kills in 3 hits. In other words, armor entirely offsets the attack upgrades and unit A, again, kills just 50% fast. 

Regardless, even the proposer, here, seems to have moved on, so this potential change is moot. 

[Dakara]

On 19/09/2022 at 12:51 AM, chrstgtr said:

Cela n'est vrai que si aucune amélioration d'armure ou de santé n'est prise pour compenser les dégâts d'attaque.

Pour être honnête, ce n'est vraiment qu'un problème d'unités qui meurent rapidement en fin de partie, ce qui conduit à des combats de spam. Je pense que le moyen le plus simple de résoudre ce problème est d'annuler un autre des changements a24 : les bonus de phases de santé. Le rétablissement des bonus de phase de santé offrirait également l'importance pure des améliorations de forge et rendrait la phase plus importante.  

 

I also sometimes feel like the fight is way too fast. it the result to a race for technologies that are too powerful in relation to the health of the units.

Of course we must reward the boom of a player, the fact that he is advantaged on the technologies obtained.
But maybe we can do better than that. It's a very simplistic view of just increasing damage and armor.

My idea would be to completely reform the technologies of the forge. Leave the facility improvements found in all the str (armor / life / attack)  Of course we keep in the castle the +25% damage tech. Interesting to me is to link technologies to unit classes.

We have sevaral classes : Spear, Pique, Archers, Swordmen, Slingers, Skirmish, Axe men, Cavalerie and more generalist : Infantery

 

 

Basic technlogie : we can adapt for each civ if neccesary. 

Phase 2 :

Range javelin (insert roleplay name  of technlogie for improve javelin) Skirmish : +3 range 

Shild of skirmish : +1 piercing armor (because skirmish have little shield we can imagine we improve with some part of metal their shield)

Bow (insert roleplay name ): +5 portée 

insert name : better gear of cavalerie : +5% speed 

Sword : +10rate attack speed 

We can imagine 2 or 3 more tech in P2

Phase 3 :

Pique : best efficient of armor, same armor for same speed as spear unit( it 9 i guess)

Axe and sword men : deal 30% dammage to area unit of target (buff melee unit without up imbeatable in little army)

crossbow : ignore all percing armor

choice techlogie : Low cost. 100 Metal, give advantage and disaventage. 30 seconds Reversible click.

choice of bow: Long range bow +20 meters but attack time +50%

pike armor choice: +50% armor and 50% reduced speed

spear cavalry armor choice: +50% armor and 30% reduced speed

Orther choice of spear cav armor : -70% armor but +20% speed and 30% damage.

Metal balls for slingers with little piques: damage +5% and causes the target to bleed 1 damage per second for 10 seconds but slinger cost +10 metal.

We can imagine 2 tech more tech in P3, and maybe some tech for champ of each civilisation. If not enought maybe reduce power of phase 2 tech for give level 2 tech in phase 3?

 

To conclude :

In P3 full technology a player would have at his disposal slightly stronger units in terms of armor and damage inflicted without this creating a snowball effect but thanks to technological progress he would have notable improvements in his movement speed, its reach, utilitarian things. But it also makes it possible to implement real player choices, which will certainly be little used because of the penalties they confer, but as it is situational, some players will have fun using its so-called "choice" technologies.

 

 

 

 

 

Edited September 20, 2022 by Dakara

[real_tabasco_sauce]

8 hours ago, Dakara said:

Basic technlogie : we can adapt for each civ if neccesary. 

Phase 2 :

Range javelin (insert roleplay name  of technlogie for improve javelin) Skirmish : +3 range 

Shild of skirmish : +1 piercing armor (because skirmish have little shield we can imagine we improve with some part of metal their shield)

Bow (insert roleplay name ): +5 portée 

insert name : better gear of cavalerie : +5% speed 

Sword : +10rate attack speed 

We can imagine 2 or 3 more tech in P2

Phase 3 :

Pique : best efficient of armor, same armor for same speed as spear unit( it 9 i guess)

Axe and sword men : deal 30% dammage to area unit of target (buff melee unit without up imbeatable in little army)

crossbow : ignore all percing armor

choice techlogie : Low cost. 100 Metal, give advantage and disaventage. 30 seconds Reversible click.

choice of bow: Long range bow +20 meters but attack time +50%

pike armor choice: +50% armor and 50% reduced speed

spear cavalry armor choice: +50% armor and 30% reduced speed

Orther choice of spear cav armor : -70% armor but +20% speed and 30% damage.

Metal balls for slingers with little piques: damage +5% and causes the target to bleed 1 damage per second for 10 seconds but slinger cost +10 metal.

We can imagine 2 tech more tech in P3, and maybe some tech for champ of each civilisation. If not enought maybe reduce power of phase 2 tech for give level 2 tech in phase 3?

 

To conclude :

In P3 full technology a player would have at his disposal slightly stronger units in terms of armor and damage inflicted without this creating a snowball effect but thanks to technological progress he would have notable improvements in his movement speed, its reach, utilitarian things. But it also makes it possible to implement real player choices, which will certainly be little used because of the penalties they confer, but as it is situational, some players will have fun using its so-called "choice" technologies.

 

https://code.wildfiregames.com/D4788

mod (not identical to patch atm) :

unit_upgrades.zip

[BreakfastBurrito_007]

Sorry for the confusion @chrstgtr and @LetswaveaBook. I added quotations around "unfair" on that post because I was acknowledging that my previous idea about attack upgrades being unfair was wrong. I hope that makes sense for everybody.