(7-72) Change How "Purchase Cost Reduction" Abilities Work

[azum4roll]

Part of the "stacking discounts is evil" series.

Currently, there are multiple sources of purchase cost reduction that reduces the Gold cost of investing in buildings and buying units (and projects, if enabled):

• Forbidden Palace: -15% global
• Industry: -30% global in total
• Stock Exchange: -20% local
• Rialto District: -10% global, -5% local

Military-Industrial Complex and Communism only affect cost of purchasing units/investing in buildings respectively and stack multiplicatively with the above, and thus not considered in this proposal.

A Venice player can get a whooping -80% off of all gold purchases in the capital and -75% in the rest. You may say this is synergy, but I can only see imbalance when a single component (Industry scaler or Stock Exchange) may contribute in doubling the value of Gold used for purchasing (which is ~1/4 of gold spent for the AI, according to @L. Vern; very likely higher for humans).

EDITED: @axatin clarified that the local and global modifiers actually stack multiplicatively with each other. Correcting the above paragraph (point still stands).
A Venice player can get a whooping -66.25% off of all gold purchases in the capital and -64% in the rest. You may say this is synergy, but I can only see imbalance when a single component (taking the Industry tree) may contribute in raising the value of Gold by ~50% when used for purchasing (which is ~1/4 of gold spent for the AI, according to @L. Vern; very likely higher for humans).

We already have precedence of changing stuff like this in Border Growth. We can do something similar here:

• Forbidden Palace: Gold is +20% more effective on building investment and unit purchase in all Cities.
  
  • Slightly better compared to now if not stacked with other sources, worse after Stock Exchange
• Industry Opener and Scaler: Gold is +7% more effective on building investment and unit purchases.
  
  • Around the same as now, if not stacked with other sources
• Stock Exchange: Gold is +25% more effective on building investment and unit purchase in this City.
  
  • Same as now, if not stacked with other sources
• Rialto District: Gold is +15% more effective on building investment and unit purchase in all Cities (+25% in Capital).
  
  • Better numbers to compensate for some of the lost "synergy". Obviously nowhere as close if you grab everything.

Cost formula:
Discounted cost = Base cost / (1 + Sum of effectiveness modifiers)

Implementation:
Add new column HurryEffectivenessModifier to Building_HurryModifiers, Building_HurryModifiersLocal, and Policy_HurryModifiers tables, default 0.

Expected bugfix (not part of the proposal):
Make the local and global modifiers stack additively as expected.

Conveying to players:
Should be no different, since currently the formula on purchase cost isn't shown on the UI.

Edit 1: Changed Industry scaler to +7% (was +5%), still a nerf when stacked with other sources
Edit 2: Corrected some text

 

[pineappledan]

Rialto is -10 global and -5 local. You got them reversed.

You didn't mention the Communism Order tenet, which lowers building purchases by 25%. Though I suppose you would keep that as a cost reduction, since you did the same with MIC?

This is similar to what we did with border growth, which was to keep 1 cost reduction alive, which could multiply the power of all the others (Russia)

"more effective on building investment" sounds too ambiguous relative to what the Babylon UA and Protectionism do "investments in Buildings reduce their Production cost by an additional 10%."

re: balance, this is a major hit to Venice and Babylon, two civs that focus heavily on gold purchases. Otherwise it looks good.

 

[Stalker0]

didn't we spend a lot of time on the border growth change though. really dug into the data, tried to ensure the numbers worked out similarly. If we put in that work I'm on board, but I think just putting in some numbers and going with it is not the right plan. might be better for the next congress after we put some data toghet.er

 

[pineappledan]

didn't we spend a lot of time on the border growth change though. really dug into the data, tried to ensure the numbers worked out similarly. If we put in that work I'm on board, but I think just putting in some numbers and going with it is not the right plan. might be better for the next congress after we put some data toghet.er

Taking all current cost reductions in a vacuum:
Forbidden palace = 1/0.85 = 18% gold efficiency.
Industry scaler = 1/0.7 = 43% gold efficiency.
Stock Exchange = 1/0.8 = 25% gold efficiency.
Rialto (capital) = 1/0.85 = 18% gold efficiency.

So taken in isolation:
the Rialto District and Forbidden palace changes are modest buffs
The stock exchange is unchanged
The industry scaler change will receive a MASSIVE nerf

Then of course you will combine at least 2 of these, because everyone has access to stock exchanges.
If you compare the other 3 in combination with stock exchange, Azum's changes will give
Forbidden Palace will go from +54% gold efficiency (1/[1-0.35]) to 45% gold efficiency (1+0.25+0.2)
Industry Scaler will go from +100% gold efficiency (1/[1-0.5]) to 55% gold efficiency (1+0.25+0.30)
Rialto (Capital) will go from +54% gold efficiency (1/[1-0.35]) to 50% gold efficiency (1+0.25+0.25)

If you combine any of the above with Autocracy or order, you will of course add an additional -33% to unit costs and -25% to buildings on top.

By my reckoning, the decision to change 5% cost reduction to 5% gold efficiency on the Industry scaler is a gargantuan nerf. Given Industry's other bonuses to Stock exchanges, which raise the priority of that building in all cities, and the lateness of when you will be adopting these tenets relative to the other cost reducers, you should only consider the scaler's power in combination with Stock exchanges. You could double the scaler to 10% gold efficiency per level (60% total with opener; 85% with Stock exchange) and still not be at the same power as the current bonus. Even without considering the power of Industry's scaler in combination with the Forbidden Palace and ideologies, this is nothing short of a massacre.

TL;DR, I think Azum's numbers are fair, EXCEPT for his decision not to increase the industry scaler. I recommend increasing gold efficiency to 10% on opener and scaler.

 

[pineappledan]

I should also note that in the current paradigm, you can get more than 100% cost reduction.
The maximum attainable cost reduction is currently:
-20% on stock exchange
-30% on industry scaler
-15% on Forbidden Palace (progress)
-15% on Rialto District (Venice capital)
-33% on Military Industrial Complex (Autocracy)
== 113%
The game will charge you 1 for buying units if you hit 100% cost reduction (though it is funny to imagine the game paying YOU for 13% of the bought unit's cost)

Venice would probably not go Progress, but even without considering Rialto, any civ can attain -98% cost reduction on unit purchases if they manage to build 1 semi-competitive wonder, and then open up 2 different policy trees. That is equivalent to 5000% gold efficiency.

Changing the subtractive cost reductions to additive gold efficiency leaves the door open for modmods etc to also add cost reduction or gold efficiency abilities. Currently the system is already capped at a limit where adding any more could just not yield any benefit.

 

[Stalker0]

we are discussing in discord, but I challenge these numbers from an actual game I'm in. I do not think the cost reductions are as steep as these numbers would indicate, we aren't accounting for somethign.

 

[axatin]

@azum4roll The assumptions of this proposal are not correct. Player-level modifiers and city-level modifiers already stack multiplicatively. The combined effect of Industry and Stock Exchange is -44%, not -50%.

Spoiler :

 

[azum4roll]

Which makes it even weirder.

Global and local modifiers in most (if not all) places stack additively. This one would be the exception, and not desirable because Venice gets both on one building.

I'll correct the OP, but the proposal and rationale still stand.

Also buffing Industry scaler to +7% effectiveness.

 

[azum4roll]

Sponsored.

 

[usadefcon1]

Can someone post a TLDR on this one for us common folk to understand please?

 

[Anarcomu]

Can someone post a TLDR on this one for us common folk to understand please?

It works as an "effectiveness" divisor factor.
You have a base 100% effectiveness.
If you get 100% increased effectiveness, you have a total 200%, cutting costs in a half (-50%)
If you get 200% increased effectiveness, you have a total 300%, cutting costs in a third (-67%)
If you get 300% increased effectiveness, you have a total 400%, cutting costs in a quarter (-75%)

Any other value gives a result in between. Those are example values that should be easier to grasp the logic on.

 

[Anarcomu]

Btw @azum4roll I made a counterproposal to yours which only difference is the industry scaler, so nothing different code-wise.
Would you be willing to sponsor it too ?

 

[axatin]

This table should summarizes the effects of the proposals. A base cost of 1000 Gold will be reduced as follows:

Modifiers	current implementation	azum4roll's proposal	Anarcomu's proposal	axatin's proposal
Forbidden Palace	850	830	830	850
Industry (all policies)	700	700	620	700
Stock Exchange	800	800	800	800
Rialto (Capital)	850	800	800	850
Rialto (other city)	900	860	860	850
FP + Stock Exchange	680	680	680	680
Industry + Stock Exchange	560	590	540	560
Rialto (Cap.) + SE	670	660	660	680
FB + Ind. + SE	440	530	480	470
Rialto (Cap.) + Ind. + SE	450	520	470	470
Rialto (other) + Ind. + SE	480	540	500	470

 

[azum4roll]

Huh, how's @Anarcomu's costs more expensive than mine, with a bigger Industry bonus?

 

[axatin]

oh right, it's 10% and not 5% in that proposal. 5% was your original value. Table is corrected.

 

[usadefcon1]

I see, so rather than stacking bonuses on top of each other they all get added together into the denominator and then calculated once.

 

[azum4roll]

Which is how "modifiers" are supposed to work.

 

[pineappledan]

from the table, it really seems like this is all a nothingburger. The only one that seems to move the needle is Azum’s change, but that can be attributed to his industry nerf, rather than a significant difference in the math.

 

[azum4roll]

Because it happens to have missed the Forbidden Palace + Industry combo, which currently stack additively.

Modifiers	current implementation	azum4roll's proposal	Anarcomu's proposal	axatin's proposal
Forbidden Palace	850	830	830	850
Industry (all policies)	700	700	620	700
Stock Exchange	800	800	800	800
Rialto (Capital)	850	800	800	850
Rialto (other city)	900	860	860	850
FP + Industry	550	610	550	560
FP + Stock Exchange	680	680	680	680
Industry + Stock Exchange	560	590	540	560
Rialto (Cap.) + SE	670	660	660	680
FB + Ind. + SE	440	530	480	470
Rialto (Cap.) + Ind. + SE	450	520	470	470
Rialto (Cap.) + FP + Ind. + SE	330	470	430	400
Rialto (other) + Ind. + SE	480	540	500	470
Rialto (other) + FP + Ind. + SE	360	490	450	400

 

[Recursive]

Timestamp post to arrange all the threads in a neat order.