I think this changes the default game too drastically to be always-on. A standard 1950 start now begins with 30 years of inflation already applied.

If it's not always on, then there is absolutely no point except providing a third incompatible variant for NewGRF authors to gripe about.

While i generally think this is a step in the right direction, i think it falls short on so many levels.

Inflation is meant as a factor that drives the difficulty forward as the game goes on, so that the player is forced to adapt to new technologies and optimize the network to stay profitable.

But forcing this divergence of income vs costs onto a newly started game makes it unnecessarily difficult.

A better approach IMHO would be to always apply the same inflation to costs, but make the difficulty setting about the income divergence. the divergence then would start being applied after game start only, while the price inflation would be consistently applied at the base game year, perhaps even further back, if we introduce fractional currency?

Also, i think applying the divergence should stop at the moment no new vehicles get introduced (the point where the reliability curve gets frozen), not at the "original" end year. maybe even a more dynamic inflation divergence based on the vehicle progression

Would it be possible to separate the inflation rate setting into as much as 7 different settings? If so, here are my proposed settings:

1. Loan interest rate: 0% to 255%.
2. Inflation rate for company expenses: 0% to 255%.
3. Inflation rate for cargo income 0% to 255%.
4. Initial inflation factor for company expenses: 1/16 to 1024 in steps of 1/16.
5. Initial inflation factor for cargo income: 1/16 to 1024 in steps of 1/16.
6. Maximum inflation factor for company expenses: 1 to 1024 in steps of 1/16.
7. Maximum inflation factor for cargo income: 1 to 1024 in steps of 1/16.

This would separate loan interest from inflation and provide customization about the income divergence and initial inflation, on top of the existing inflation rate customizability. This would allow you to make the game easier or harder at the start with finer cost adjustments then what even the Basecosts NewGRF can provide, as well as allow you to simulate deflation (such as during recessions), high inflation (like in the USA during the 1970's), or even mild hyperinflation by changing the inflation rates and therefore financial difficulty curve.

Note that splitting the settings like this would make the inflation toggle obsolete, as instead the inflation rate could be set to 0 to disable inflation (or loan interest to 0 to disable that, if you will).

When converting old savegames, the settings should be set as follows.

1. Loan interest rate: Same as before (same as the old loan interest setting)
2. Inflation rate (company expenses): Same as before.
3. Inflation rate (cargo income): Same as before, then subtract 1 percentage point.
   Note that if inflation is off, then both values will be zeroed on conversion.
4. Initial inflation factor (both sides): Always equal to 1.
5. Maximum inflation factor: Calculated as (inflation rate in percent / 100 + 1)^170, rounded to the nearest 16th and capped at 1024.

Note that you could still increase the maximum value for the initial/maximum inflation factor another 32 times, up to 32767. This "inflation cap" value is defined in

Even though the last comment was made some time ago: The whole point of this PR is to provide a consistent baseline to balance against.

As such, more settings: Just No.

Modest proposal: remove inflation.

Inflation:

• makes assumptions about span of gameplay (start / end year) that don't hold in OpenTTD with variable start / end years
• makes the game borderline unplayable when inflation has reached a high level
• causes bug reports to newgrf authors about unplayable costs (I've been on both ends of this, filing useless bug reports on newgrfs due to inflation, and receiving them myself)

If we want increasing vehicle run costs over time, newgrf authors can take care of that themselves.

To avoid derailing this PR further general inflation discussion is moved to #8429 (thanks)

If inflation as a whole is removed completely, then would it be possible to extend the precision of each of the BaseCosts values so as to allow greater precision in changing the base costs (via BaseCosts NewGRF or a similar setup)? For example:

• Current domain (from BaseCosts NewGRF parameters): Powers of 2 from 2^-8 to 2^16
• Proposed domain: All numbers from 2^-8 to 2^16 (or 2^15 or 2^23) that are multiples of 2^8

Please create a discussion to discuss if Inflation should be removed or not. Let's not try to derail PRs too much with completely opposite views of the PR :D Many people might miss it, and might want to pitch in too!

Tnx :)

To avoid derailing this PR further general inflation discussion is moved to #8429 (thanks)