Inflation

[Koshling]

It has come to light that there is a bug in the inflation calculation, that effectively gives you double-compounded inflation effects. I'll be fixing that, but once I started thinking about the modelling of inflation I relaised it's just plain wrong anyway!

Firstly a history lesson.

Prior to about release 20 inflation used to calculate your costs based on the current inflation rate and the number of turns progressed you were - i.e. - it calculated what your costs would be on the assumption that your current inflation rate is the one that had always been in effect. Thus changing the inflation modifiers INSTANTLY changed your costs, instead of just changing the rate at which they increase (which is ACTUALLY what inflation does)

Around release 20 or so, this was changed to have the inflation rate just effect the rate of increase. The bug just discovered is that it got that calculation wrong and actually increased it by more than it should have.

HOWEVER, in the real world inflation doesn't just change the cost of goods - it also changes the money supply - i.e. -the income side of the equation. In game terms that means that inflation SHOULD apply to income as well as costs. Its only effect then (which models the real world) is to erode the value of your savings (aka treasury).

As things stand now early inflation leaves you screwed in the late game, because the inflation baked in early on has escalated costs but not income. It SHOULD escalate both (and thus devalue your treasury, which is why inflation is a bad thing), which doesn't cripple your day-to-day balance of payments.

I'd like to change this, but I don't think the freeze period is the time to do it, so I propose to fix the double-compound bug for V25, and then change the inflation model to also effect income once V25 is out.

What do people think?

 

[ls612]

Oh, that was the issue, not the precents, whew. That would explain why people have so much gold early game and so little gold later on, causing many fights over too much/little gold. I'm fine with your plan for fixing this.

 

[Koshling]

Oh, that was the issue, not the precents, whew. That would explain why people have so much gold early game and so little gold later on, causing many fights over too much/little gold. I'm fine with your plan for fixing this.

Actually there is an issue.

Effectively, because trade is mediated in gold, everyone is sort of on a gold standard, so if we apply inflation to players indivdually like this it will distort trade, because a player who has historically had high inflation will have high costs and income, but for trading his gold is the same as anyone else's gold, so it actually means the high inflation players would have a (large) advantage in trading. Thus I think the model ALSO has to apply conversion factors on trade, so that the gold given by one is adjusted for the relative inflation between the two parties and what is received is therefore less or more appropriately (i.e. - a virtual currency conversion even though both sides call what they have 'gold'). Basically this means all quantities of gold on the other player's side of the table in trade negotiations have to be converted to your 'local currency' gold when displayed/received.

I don't think this is especially hard, but it does make the code changes required a bit wider-reaching.

 

[ls612]

Actually there is an issue.

Effectively, because trade is mediated in gold, everyone is sort of on a gold standard, so if we apply inflation to players indivdually like this it will distort trade, because a player who has historically had high inflation will have high costs and income, but for trading his gold is the same as anyone else's gold, so it actually means the high inflation players would have a (large) advantage in trading. Thus I think the model ALSO has to apply conversion factors on trade, so that the gold given by one is adjusted for the relative inflation between the two parties and what is received is therefore less or more appropriately (i.e. - a virtual currency conversion even though both sides call what they have 'gold'). Basically this means all quantities of gold on the other player's side of the table in trade negotiations have to be converted to your 'local currency' gold when displayed/received.

I don't think this is especially hard, but it does make the code changes required a bit wider-reaching.

Eww, that sounds complicated. Would it not just be simpler to remove inflation entirely and be done with it?

 

[Koshling]

Eww, that sounds complicated. Would it not just be simpler to remove inflation entirely and be done with it?

That would be a shame I think.

 

[AIAndy]

That would be a shame I think.

Well, wouldn't just reducing the treasury by a percent each turn do something similar without all the issues? In reality you can't save anything anyway as an economy.

Game design wise in base Civ4 inflation was mainly a measure to take gold out of the game with time.

 

[BlueGenie]

@ls612:I stand with corrected numbers for Epic through Eternal for the iPercent. That is:
10/6/4/2
Normal speed should probably stay at 27 after having checked the total amount of turns in that setting.
If the numbers seem small you have to consider that they act that much longer too. 12505 turns (total - offset) of inflation on Eternal with 2% of a percent increase/turn is still 250% total (more with inflation increasing Civics).
27% of a percent/turn with 947 (1047-100 offset) turns on Normal is 255% total. Having Eternal on 4, like it is now, means 500% on Eternal total. With the speeds changed inflation needs to be lowered.
The numbers above set a total inflation barring increases/decreases due to Civics to ~250%. Marathon is a bit off though, 6.5% would be optimal but that might not be a possibility.

I agree with Offset not mattering as much as with the post-v20 version of inflation you'd not get any inflation until you got out of Barter Civic. That's why I said that the old numbers I posted would work but they are needed as Slavery has an inflation rate and comes a lot earlier than it did before Technology Tree changes.
I would like to see offset offsetting inflation until at least Trade/Mining has been researched but before Currency is researched. I think the numbers I posted, i.e. -350/-550/-900/-1800, should do that trick nearly. Getting it perfect for every game is rather impossible though.

@Koshling:
I don't see why the high inflation player would gain in any trades, would it not be the other way round? If high inflation player A with an inflation of arbitrary percentage of 100 trades with player B with no inflation a trade where B pays A money for a resource/technology A will always think the gold B offers is way too little and will not get the resource/technology unless B is willing to overpay by twice the optimal value.
If A offers to pay B money for a resource/technology he will always offer more than B would think the value of the resource/technology is worth and with his lack of inflation B will get a huge amount of gold for his good.

One thing I'm wondering is if two different Civics change the inflation rate do they multiply or add up or act on the rest? 15% and 10% would be a whooping 25% inflation increase if added, which is a rather high extra increase in inflation. Or it would be 16.5% increase if multiplied, which is a lot better but maybe not enough. Or 23.5% if acted on the rest, also a lot.
If there's a possibility maybe +50% inflation (not inflation itself but the additional rate at which it grows due to Civics) should be set and additional Civics influencing inflation rate acting on the rest of those 50%? Which would make my example of 15 and 10% inflation Civics 18.5% though then the higher numbers would always have to be counted first or it would skewer the calculation (doing 10% first and then the 15% would be 16% total) as the latter numbers would be counted as less half if it was based on 100.
This part is important if the Advanced Civics come into play as they have Inflation variances in more than one Civic category. Or Inflation is forced to only one Civic category and we won't need to worry about accumulative effects.

While I'm at it, if the Adv Civics do come into play would a No Inflation together with a 15% Inflation Civic have any effect or would inflation be halted/stopped anyway?

Mmm, Cheers, and Sorry for waking this beast...

 

[Koshling]

@Koshling:
I don't see why the high inflation player would gain in any trades, would it not be the other way round? If high inflation player A with an inflation of arbitrary percentage of 100 trades with player B with no inflation a trade where B pays A money for a resource/technology A will always think the gold B offers is way too little and will not get the resource/technology unless B is willing to overpay by twice the optimal value.
If A offers to pay B money for a resource/technology he will always offer more than B would think the value of the resource/technology is worth and with his lack of inflation B will get a huge amount of gold for his good.

Player A opens trade with Player B, and says 'what do you want for this?'. Player B responds with what is (to him) a large amount fo gold, but is nothign to player A. Player A (whose income per turn is inflated) gets to buy techs for small multiples (or even fractions) of what he kams every turn.

One thing I'm wondering is if two different Civics change the inflation rate do they multiply or add up or act on the rest? 15% and 10% would be a whooping 25% inflation increase if added, which is a rather high extra increase in inflation. Or it would be 16.5% increase if multiplied, which is a lot better but maybe not enough. Or 23.5% if acted on the rest, also a lot.

Currently they add, so they are percentages of the base amount (so 10% of the base amount + 15% of the base amount = 25%).

Well, wouldn't just reducing the treasury by a percent each turn do something similar without all the issues? In reality you can't save anything anyway as an economy.

Game design wise in base Civ4 inflation was mainly a measure to take gold out of the game with time.

Excellent idea. I like this MUCH better. It's very simple and it does what inflation should do.

 

[strategyonly]

Well, wouldn't just reducing the treasury by a percent each turn do something similar without all the issues? In reality you can't save anything anyway as an economy.

Game design wise in base Civ4 inflation was mainly a measure to take gold out of the game with time.

Well isn't that really what is supposed to happen in the first place?? Sounds like a better approach, with less stress on doing some large coding.

Look at today, all the "small" business owner(s) are going if not already gone, infact i just read in the MSN times that a 225 year old business is finally closing its doors, because of inflation/economy

 

[JosEPh_II]

In my current game in the Classical Era on Epic, inflation accounts for over 25% of my eXpenses. I'll post a screenie when I get back to the game. My Income is 315 gold and my eXpenses are 420+. So I'm bleeding gold at a rate of 100+ a turn on 60% research. My once 2400 treasury is now down to 700+ and I'm facing 3 war like AI at the moment. (One tried to invade but I was able to rebuff and eliminate his invasion force. Unit cost is 96 gold/turn. I'd like to reduce the number of units I have but I can't as they have stacks of 30+ units sitting on their respective borders with me. My stacks run about 15 units.)

If inflation can be made easier I'm all for it.

JosEPh

Edit: screenie attached. Actually inflation is closer to 33% of my expenses. City maintenance I think is too high also. I've reduce my bleeding gold deficit to -30+/turn But have to do more as I'm about to found another strategic city.

 

[Cyrusfan]

HOWEVER, in the real world inflation doesn't just change the cost of goods - it also changes the money supply - i.e. -the income side of the equation. In game terms that means that inflation SHOULD apply to income as well as costs. Its only effect then (which models the real world) is to erode the value of your savings (aka treasury).

That's almost right. Inflation doesn't increase the money supply, it's the process by which people's evaluation of money relative to goods corrects when the two are out of balance (with the previous evaluation of money being incorrectly high). That can be caused by excessive production of currency/money (since unbacked credit can cause it as well), or inadequate (decreasing) production of goods. The first is possibly implied by civic choices, but the second would have to look at some national total of and. MOO3 was supposed to model relative strength of currency between players, but got dropped because it was too complex.

The idea that seems prominent in the thread of decreasing the treasury turn to turn seems like the more feasible option. The problem I see is that this doesn't really punish you at all, unless you've been saving (playing thrifty/responsible) for a long time and then suddenly switch to an inflation heavy civic. In the real world, high (as in detectable) inflation tends to cause general unhappiness and instability, so the penalties for those civics may need reexamining.

 

[Dancing Hoskuld]

...In the real world, high (as in detectable) inflation tends to cause general unhappiness and instability, ...

For some of the population - high inflation is seen as good by others; a minor group is retirees and others who live of interest.

Having said that inflation comes and goes. How do you model the affect of runaway inflation in Roman times on the modern times. What about the effect of governments that "revalue" or change their monetary system. Egs "One new dollar buys 1,000 old dollars" or the move to decimal currency.

 

[Koshling]

That's almost right. Inflation doesn't increase the money supply, it's the process by which people's evaluation of money relative to goods corrects when the two are out of balance (with the previous evaluation of money being incorrectly high). That can be caused by excessive production of currency/money (since unbacked credit can cause it as well), or inadequate (decreasing) production of goods. The first is possibly implied by civic choices, but the second would have to look at some national total of and. MOO3 was supposed to model relative strength of currency between players, but got dropped because it was too complex.

The idea that seems prominent in the thread of decreasing the treasury turn to turn seems like the more feasible option. The problem I see is that this doesn't really punish you at all, unless you've been saving (playing thrifty/responsible) for a long time and then suddenly switch to an inflation heavy civic. In the real world, high (as in detectable) inflation tends to cause general unhappiness and instability, so the penalties for those civics may need reexamining.

Yeh, you're right of course. I plan ro change it to just erode treasury as per aiandy's suggestion, as being the coarse simple model. I agree however, that looking at potential other effects on hig inflation civics might be a good idea in the longer term, but not right before the v25 release.

 

[Cyrusfan]

For some of the population - high inflation is seen as good by others; a minor group is retirees and others who live of interest.

That's possible but dicey. The higher interest rates tend to follow the inflation, so if they feedback correctly to crush the inflation, and you can lock in an investment at the higher rates, you do quite well, but if you were already locked in at lower rates, you get hosed pretty badly. As a matter of modeling it, I think it's more bad than good.

That said, I think banks favor inflation to deflation because more money sloshing around tends to make defaults less likely which makes bank failures less likely, and the inflation itself serves to motivate folks to money from the mattress to the system.

Having said that inflation comes and goes. How do you model the affect of runaway inflation in Roman times on the modern times. What about the effect of governments that "revalue" or change their monetary system. Egs "One new dollar buys 1,000 old dollars" or the move to decimal currency.

Truly runaway inflation (beyond the capacity of interest rates to mitigate) probably should be in the apocalypse thread. I think something could be done with the value of foreign trade routes since exports tend to be the main motivation for nations to devalue their currency (with 'soft' defaulting on national debt a close second in hard times-but we don't have a model for that unless someone resurrects the SMAC ability to loan money to other players). I don't have any examples, but I would suspect 'new' currency doesn't really do much to help confidence as it's an admission of how bad things are-maybe this could be an event of some kind?

 

[Terxpahseyton]

Eh, I am a little unsettled by this thread. ^^ Inflation as implemented in Civ4 has literally nothing to do with with monetary real-world inflation. It's just a modifier to give "inflated" - that means big - empires a penalty.
If you guys want to add monetary inflation as a feature - well that's an interesting thought. But maybe keep in mind what the current "inflation"-mechanic actually is before going to replace it with something entirely different bearing the only similarity in carrying the same label, "inflation". Just a thought

 

[AIAndy]

For some of the population - high inflation is seen as good by others; a minor group is retirees and others who live of interest.

Having said that inflation comes and goes. How do you model the affect of runaway inflation in Roman times on the modern times. What about the effect of governments that "revalue" or change their monetary system. Egs "One new dollar buys 1,000 old dollars" or the move to decimal currency.

A simple "revalue" does nothing beyond messing with people's knowledge of prices.
Hyperinflation usually begins in economies that run at their limit of production when at the same time there is high additional demand and a surplus of currency units in the hand of those with demand. Once started, it feeds itself because people spend their money as fast as they can before it devalues and the production drops as the currency is no more suited as well for business relations.

 

[AIAndy]

Eh, I am a little unsettled by this thread. ^^ Inflation as implemented in Civ4 has literally nothing to do with with monetary real-world inflation. It's just a modifier to give "inflated" - that means big - empires a penalty.
If you guys want to add monetary inflation as a feature - well that's an interesting thought. But maybe keep in mind what the current "inflation"-mechanic actually is before going to replace it with something entirely different bearing the only similarity in carrying the same label, "inflation". Just a thought

I don't think the current mechanic penalizes big empires any more than smaller ones. It just increases gold costs by a multiplier that increases with time. That is an important factor in late game gold balance though and I am not sure if a serious change to that mechanic should be done at the end of a development cycle (unless it is just adding the saving reduction as an additional effect).

 

[Koshling]

I don't think the current mechanic penalizes big empires any more than smaller ones. It just increases gold costs by a multiplier that increases with time. That is an important factor in late game gold balance though and I am not sure if a serious change to that mechanic should be done at the end of a development cycle (unless it is just adding the saving reduction as an additional effect).

The reason it should be done is that the change made back in V20 or so, which effectively bakes in early civic effects later in the game, is turning out to have consequences that are game-breaking.

You're right in that it has no differential impact on civilizations of different sizes however.

 

[AIAndy]

The reason it should be done is that the change made back in V20 or so, which effectively bakes in early civic effects later in the game, is turning out to have consequences that are game-breaking.

You're right in that it has no differential impact on civilizations of different sizes however.

I agree that it is a bad mechanic and needs changing but I wonder what effect it will have on late game gold proliferation.

 

[Koshling]

I agree that it is a bad mechanic and needs changing but I wonder what effect it will have on late game gold proliferation.

I'd rather have a (uniform) late game gold proliferation problem (which can be tweaked in various well understood ways), than a terminal, but non-unifiorm (depends on your civic history and so on, and thus only effects some play styles) late game gold crunch.