C

To understand how Roller Coaster Tycoon calculates park value, the first thing you need to understand is that when it talks about park value, it's not talking about actual game dollars. It's not talking about the concrete dollar value of the things you've built or how much you can get from selling them or how much revenue you've earned. No, it's talking about theoretical "value to all the consumers in the park right now" dollars.
Before we can talk about park value, we need to talk about ride value. So here we go:
Ride Value
For Roller Coaster Tycoon 2, as soon as a ride receives an excitement/intensity/nausea rating, it is assigned a "value". You can think of it as a "dollars of happiness" value. (Before the ride receives those ratings, its value is treated as 0)
(Note to fellow source-divers; while facilities like bathrooms and shops are considered to be 'rides' by the code, they never receive excitement/intensity/nausea ratings, and so never receive a 'value'.)
Different types of ride have different 'inherent' valuations for those excitement/intensity/nausea ratings. For example, Monorails award $7 of per-user 'value' for each point of 'Excitement' rating, but only $0.60 for each point of 'Intensity', and -$1.00 for each point of 'Nausea'. By comparison, an Inverted Hairpin Coaster (a much more intense ride) awards $5 per point of 'Excitement', $3 per point of 'Intensity', and $3 per point of 'Nausea'. These modifier values are mostly pretty intuitive; stereotypically-intense rides like roller coasters will typically offer bigger valuations overall, and will typically give much higher rewards for points of intensity and nausea, than do more stereotypically-sedate rides.
As a point of interest, the single-highest valuation in the game is Go Karts, which give a whopping $12 per point of 'Excitement'. Second highest is Log Flumes, Splash Boats, and Lift, which award $8 per point of 'Excitement'.
Anyhow, here we go, calculating Ride Value:
Each of the ride's three evaluation ratings (Excitement/Intensity/Nausea) are multiplied by the ride type's valuation for that rating, and then summed together. This makes the ride's total base value. (Again, this is a theoretical "value to each customer" value, not the cost or the sale price!) [* see footnote below, for an exciting and fun note about this math]
If the ride is no more than 12 months old, it gets an extra $1 added to its value.
If the ride is no more than than 4 months old, it gets a further $2 added to its value (making a $3 bonus in total, including the one for being no more than 12 months old).
If the ride is at least 40 months old, it loses a quarter of its value.
If the ride is at least 64 months old, it loses a quarter of its remaining value (about 44% loss of value in total, including the "at least 40 months old" loss above).
If the ride is less than 200 months old, then it receives more compounding 25% penalties at 88, 104, 120, and 128 months of age. These penalties are NOT imposed after the ride is 200 months old; presumably the game considers the ride to be so old that it's now 'historic'; only the 40 and 64-month penalties will be applied for the remainder of the ride's lifetime.
If there are any other rides of the same type in the park (another Monorail, for example), then we receive another 25% penalty to whatever value remains after all other adjustments above.
Finally, if the ride value calculated above results in a total value below zero, it gets raised to zero.
That's it; that's how the game calculates "ride value".
Let's take a moment to point out some things which don't affect ride value. Scenery does not affect ride value (except inasmuch as it can increase the ride's excitement rating). Color choice doesn't affect ride value. Cleanliness does not affect ride value, and neither does the ride's maintenance track record. It doesn't even matter whether or not the ride is currently turned on; the value is strictly concerned with those three ratings, how effectively they compliment the game's preset rating valuations for the ride type, how new (or old) the ride is, and whether there are any other rides of the same type in the park.
SO! Now that you understand what affects ride value, let's move on and talk about park value.
Park Value
Calculating the park value works like this:
For each ride:
For each customer who has ridden the ride during the previous 9600 game ticks (about five minutes), add the ride's per-customer "dollars of value" value, as calculated above, in the Ride Value section.
Every ride type has an "expected number of customers" value. This is a hard-coded value per ride type; it's high for typically-high-intensity rides (for example, inverted coasters, which have a value of 400), and low on low-intensity rides (spiral slide, which has a value of 160). For each "expected customer", add the ride's value again.
This final total (the total of "value from customers in the past five minutes" and "value from expected customers in the future") is the ride's current total dollar-value contribution to park value.
Finally, add $7 per guest currently in the park.
And that's everything. That's the whole "park value" calculation.
Suggestions to increase park value
So with the above knowledge, we can make some guesses about effective ways to increase park value.
Here are some notable things: Shops don't contribute to park value at all. Bathrooms and other facilities don't contribute at all. Ride length doesn't matter. All that matters is how many people you have in the park, how highly valued your rides are, and how many people have been customers of those rides within the last five minutes. Any time before the last five minutes doesn't matter at all.
Now, I haven't actually tried this (I've been doing a deep dive into the code as I wrote all this), but it occurs to me that if you're trying to improve park value, it'd be good to make your rides shorter, and add more cars or trains, so you can get more people on the ride at the same time, and get them off the ride faster so that the next group of customers can get on; more customers means more park value! Similarly, the amount of revenue doesn't matter at all; what's important is the number of people who have ridden the ride, and the perceived value-to-customers of the ride. If dropping prices results in more people riding the rides, your park value will go up, even though you may be making less money!
Other important points: Duplicated ride types is an absolute killer. That 25% penalty to ride value is huge, and it affects more than just your park valuation; ride value is also used to determine whether people even want to come to your park at all, whether they want to spend money to go on the ride, and gets used in other spots as well. If you have two different log flume rides in your park, destroy one and the value of the remaining one will increase the next time the ride recalculates its value. And that increase gets further multiplied by the number of customers who ride it.
What you really want is a lot of different types of rides, and lots of customers on each, since your park value is increased by the ride's value for every ride customer from the last five minutes.
Although amusingly, an unused ride still makes the "expected number of customers" contribution to park value even if it's in the far corner of your park, broken down, turned off, and with no paths leading to it. As long as it has an excitement/intensity/nausea rating, its hardcoded "expected customers" do count toward your park value, even though it's never actually had a customer and never will have one.
So if you have money to burn and want to increase your park value quickly, you could just build a bunch of (different!) rides off in corners of your park, test them long enough to get their ratings, and then turn them off again and have them bump up your park value based upon their expected number of guests (intense-type roller coasters will work best for this, since they have the most "expected" guests). That feels a little silly and exploitative, and you'd probably do better financially by letting folks actually go on the rides. But it might be fun to play with sometime, to see whether you could make a highly-valued park where the guests aren't actually allowed on the rides.
Source: All of the above comes from an analysis of reverse-engineered Roller Coaster Tycoon 2 source code published by the OpenRCT2 project.
[*] Note that the math in this bit of the code has been optimised for speed over accuracy, and will actually produce a value about 2% lower than the math I've presented here. Instead of dividing by 32,000 in the final step of their calculation (as they convert between different sets of units), they're right-shifting by 15, which is equivalent to dividing by 32,768. It's close, and it's much much faster on the hardware which was common back when the game was new, which is why they've done it. It took me about a half hour to figure out what the math was supposed to be doing, and then to work out why it wasn't doing it. Yay for math kludges!