Primary industries disappear in time

[pickpacket]

Primary industries that don't receive service are likely to shut down. I get that, but; shouldn't they eventually be replaced by new instances being generated?

I've started a game where I just set up a passenger train line to not go bankrupt, and then I'm running the game on fast forward to see what happens. I haven't kept track of all industries, I just focus on coal and tea tree farms (of course I always use my NewGRF for tea). Coal mines have gone from 142 in 1950 to 55 in 2010, and the tea tree farms have gone from 179 to 58 in that same time. Meanwhile banks and oil rigs have been popping up all over the place. Of course oil wells have all disappeared, but the oil rigs just get more and more numerous.

This doesn't really help either: https://newgrf-specs.tt-wiki.net/wiki/I ... ce_Chances because appearance chances have no correlation to closures, but are rather relative to appearance chances of other industries. The result is that primary industries which have not yet received any service close down over time, while tertiary industries increase (I might be wrong about the tertiary industries because I don't know how the number of power plants have changed, but teahouses inherit their appearance chances and conditions from the temperate banks and are thus increasingly common the longer the game goes on).

This is not a bug; it's entirely as designed: https://wiki.openttd.org/en/Manual/Game ... nomy-rules

The question is if it should be designed that way. I'll give my own current opinion in a following post.

[pickpacket]

The easiest solution to this is already available. You can turn off industry closure and creation. I find that to be a crutch rather than a solution, though. It means you can look at a coal mine in 1950 and plan to extend your service to it 60 years later and in a single player game that plan will not be challenged at all.

The closure of badly serviced primary industries is a good thing, imho. The problem is that new primary industries don't appear at roughly the same rate.

I would like the probability of primary industry creation to change from being relative to other industries (and comparably very small probability) to being relative to the number of industries in existence compared to the starting number. I'll come back when I have figured out how to express what I mean in mathematical terms...

[pickpacket]

So. I would propose the following:

x = number of instances of the industry type at map creation
y = number of instances of the industry at this time
p = the probability of a new instance appearing at this tick

if ( y < 0.9x ) then p = 1
else if ( y > 1.1x ) then p = 0
else p = 0.5 + 5 ( y / x - 1 )

This would mean that the number of instances of an industry fluctuates between 90% and 110% of the initial number.

It's not a perfect solution. It solves the problem of the number of primary industries decreasing over time, however the *reason* the industries disappear in the first place is that their production keeps falling as there's a much higher probability of that happening than the reverse. This means that hardly any primary industries disappear at first, but then as more and more of them reach the precipice whole swaths of them disappear in short order. I'm not sure what a good solution to that would be.

[LaChupacabra]

You're not the first to notice this problem. Here is a similar topic: viewtopic.php?t=90314 And here you will even find a proposed, ready-made solution: https://github.com/OpenTTD/OpenTTD/pull/10134

There you can find also an alternative proposal: https://github.com/OpenTTD/OpenTTD/pull ... 1304583775

In the example above, "Very Low" + "Very Low" means that the game would try to keep a similar number of specific enterprises. But you could also set that this amount could decrease or increase over time. Now, with the addition of the "Custom" option, a slightly different approach would be needed, but the overall effect would be the same.

[pickpacket]

Any of those would be an improvement to the current!