Suggestions for long standing game issues.

Got an idea for OpenTTD? Post it here!

Moderator: OpenTTD Developers

User avatar
kamnet
Moderator
Moderator
Posts: 8548
Joined: 28 Sep 2009 17:15
Location: Eastern KY
Contact:

Re: Suggestions for long standing game issues.

Post by kamnet »

And with that said, let's keep this discussion focused on the suggestion and not the individuals posting, keep your passions in check.
LaChupacabra
Route Supervisor
Route Supervisor
Posts: 385
Joined: 08 Nov 2019 23:54

Re: Suggestions for long standing game issues.

Post by LaChupacabra »

uzurpator wrote: 15 Mar 2023 16:32 1. Industry rating algorithm
As I know this applies to Station rating, not industry and contrary to appearances, this is not as illogical as it might seem. It's not perfect, but the impact of speed on the rating makes sense. I admit that I also thought about a similar solution, where the speed would have some impact but with taking into account the maximum speed in the current period, but... it would cause other problems. The basic one is transport efficiency in the early years - it is low. In that case, if production levels could be easily increased in the early years, you would have to use a huge number of vehicles - that would look bad and spoil the fun, because instead of developing more connections, the player would focus on one or only few, where production is growing so fast that the entire focuses only on that. In my opinion, in the case of the base game, this element should rather not be changed, and if so, the change should be more comprehensively thought out. This is not a question of dogma, but of specific reservations.

Being on the topic of the station rating, a change I'd like to see is to add the effect of the number of transport directions when cargodist is enabled - if the player offers more destinations, the station rating would be correspondingly higher (+10%? +20%?). This would make the construction of a network of connections economically sensible, which currently, unfortunately, it is not at all.
uzurpator wrote: 15 Mar 2023 16:32 2. Can trucks get some love in the base game?
I don't consider the proportionally higher cost of road vehicles a disadvantage. I believe this is appropriate because building roads and sending vehicles over them is much simpler than building railroads. If you changed this balance, building railroads would no longer be look as profitable, and the game would lose more than it would gain. But indeed, perhaps some adjustment in favor of road vehicles would not be a bad thing. Perhaps it would be worth considering adding something like an "Improved default vehicles costs balance" option. Such an option could introduce changes not only in the case of road vehicles, but also other vehicles, and above all aircraft, whose balance is currently definitely broken. With this option will be possible to fix the poor original balance, but without replacing it which I believe is the main argument against any change on this issue.
I am sorry for may English. I know is bed.
Eddi
Tycoon
Tycoon
Posts: 8257
Joined: 17 Jan 2007 00:14

Re: Suggestions for long standing game issues.

Post by Eddi »

LaChupacabra wrote: 24 Mar 2023 00:47 a change I'd like to see is to add the effect of the number of transport directions when cargodist is enabled
the general problem here is, how do you differentiate cargos with very few destinations (say, wood to saw mill) vs. cargo with lots of destinations (say, goods)

counting the potential destinations may be a computationally expensive task.
something like an "Improved default vehicles costs balance" option.
it's currently generally viewed that we have too many options
User avatar
odisseus
Director
Director
Posts: 552
Joined: 01 Nov 2017 21:19

Re: Suggestions for long standing game issues.

Post by odisseus »

LaChupacabra wrote: 24 Mar 2023 00:47 Being on the topic of the station rating, a change I'd like to see is to add the effect of the number of transport directions when cargodist is enabled - if the player offers more destinations, the station rating would be correspondingly higher (+10%? +20%?). This would make the construction of a network of connections economically sensible, which currently, unfortunately, it is not at all.
Eddi wrote: 24 Mar 2023 10:01 counting the potential destinations may be a computationally expensive task.
I think what's being suggested here is to take into account the number of destination stations that are already identified as such by the Cargodist algorithm. (For cargo types that aren't managed by Cargodist, this boost shouldn't apply either.) I don't see why it would impose additional computation costs.
Eddi wrote: 24 Mar 2023 10:01 the general problem here is, how do you differentiate cargos with very few destinations (say, wood to saw mill) vs. cargo with lots of destinations (say, goods)
I don't see the problem here. The stations already have separate ratings for different cargo types; transporting one type to multiple destinations would boost the station rating for that cargo only, without affecting the other types.
LaChupacabra wrote: 24 Mar 2023 00:47 I don't consider the proportionally higher cost of road vehicles a disadvantage. I believe this is appropriate because building roads and sending vehicles over them is much simpler than building railroads. If you changed this balance, building railroads would no longer be look as profitable, and the game would lose more than it would gain. But indeed, perhaps some adjustment in favor of road vehicles would not be a bad thing. Perhaps it would be worth considering adding something like an "Improved default vehicles costs balance" option. Such an option could introduce changes not only in the case of road vehicles, but also other vehicles, and above all aircraft, whose balance is currently definitely broken. With this option will be possible to fix the poor original balance, but without replacing it which I believe is the main argument against any change on this issue.
I totally agree with this point of view. However, changing the balance of default vehicles doesn't require changes to the core game; it can be implemented as an add-on.
Eddi
Tycoon
Tycoon
Posts: 8257
Joined: 17 Jan 2007 00:14

Re: Suggestions for long standing game issues.

Post by Eddi »

odisseus wrote: 24 Mar 2023 11:11 I think what's being suggested here is to take into account the number of destination stations that are already identified as such by the Cargodist algorithm.
that makes absolutely no sense, neither from a realism nor from a gamebalance point of view.
User avatar
uzurpator
Transport Empire Moderator
Transport Empire Moderator
Posts: 2178
Joined: 10 Jan 2003 12:21
Location: Katowice, Poland

Re: Suggestions for long standing game issues.

Post by uzurpator »

LaChupacabra wrote: 24 Mar 2023 00:47 It's not perfect, but the impact of speed on the rating makes sense. I admit that I also thought about a similar solution, where the speed would have some impact but with taking into account the maximum speed in the current period, but... it would cause other problems. The basic one is transport efficiency in the early years - it is low. In that case, if production levels could be easily increased in the early years, you would have to use a huge number of vehicles - that would look bad and spoil the fun, because instead of developing more connections, the player would focus on one or only few, where production is growing so fast that the entire focuses only on that. In my opinion, in the case of the base game, this element should rather not be changed, and if so, the change should be more comprehensively thought out. This is not a question of dogma, but of specific reservations.
You know. I kind of agree. This is an issue in Transport Fever - where there are no ratings - and you can get swamped by production in 1850-1920 period with no means to really handle the traffic. The solution I am considering there ( provided I'll return to modding it ) is to make the production grow with time. Currently - source industries produce 400 flat rate ( provided no mods ), while I think that starting at 100 and growing towards 400 in later game would work better.

However. That is Transport Fever.

Note, that I am omitting graphical set weirdness, assume base set only.

There are two specific issues in OTTD in regards to this.

First of all, the game is all too reliant on speed as a determining factor for just about anything. Transit time determines payment. Rated speed determines ratings, ratings determine production rate increase. It is one dimensional. In order to reach the 80+ rating at an industry you need both ~160km/h trains and a statue nearby, or 250km/h trains. If you won't get there, then the game is a slow, slow burn until you get to monorails. Then flood gates open and if you play the game right, all industries produce a TON of stuff and grow rapidly and then what you describe does happen - you need to babysit every line because the smooth industries increase all the time by a bit. In a way - the old style 2x production was, in fact, better in that regard.

Second of all - I'm not an AI. On a large enough map, or dense enough, when one develops one part of a map, the undeveloped part tends to wither away and it is a major annoyance to bring it up to full production.

I gave it a thunk over the last few days, in a way, because what odisseus said earlier.

I think that if industry _growth_ was in fixed amount, say 10-25 tons per increase, but decrease was at a percentage, than it would create a nice hysteresis where zombie industries would be easy to rescue, but getting to that umpteen hundred production would take several consecutive production increases. On the other hand, effing up and allowing for a reduction will cause immediate 10-20% lopping off and multi year loss of service.

Napkin calculations incoming:
An industry producing 30 with base of 120 requires 10 increases ( 15% each ) to get to base production.
base 66/33 chances of increase/decrease, that turns to 30 production changes on average and 75 to get to 1000 production
excellent service 83/17 chances and that turns to 'merely' 12 production changes to get to base and about 25 to 1000 production

The difference is staggering, to be honest. Which also explains why it is almost impossible to get zombie industries to full production without super fast trains and lots of luck.

In 'new' system it would require merely seven increases ( 15 tonnes each ) and:
base 66/33 chances would make it into about 10-12 production changes to get there
excellent 83/17 make it into 9 or so.

Also in 'new' system industries would hover about 200 production with normal service and 400 production with excellent service, as at that level one percentage decrease equals same number of increases.

To make it smoother:
normal 60% rating service is 2:1 increase/decrease - no change
add above average rating ( 70% ) makes it into 3:1
and then excellent rating ( 80% ) makes it into 4:1 - no change
I don't consider the proportionally higher cost of road vehicles a disadvantage. I believe this is appropriate because building roads and sending vehicles over them is much simpler than building railroads. If you changed this balance, building railroads would no longer be look as profitable, and the game would lose more than it would gain. But indeed, perhaps some adjustment in favor of road vehicles would not be a bad thing. Perhaps it would be worth considering adding something like an "Improved default vehicles costs balance" option. Such an option could introduce changes not only in the case of road vehicles, but also other vehicles, and above all aircraft, whose balance is currently definitely broken. With this option will be possible to fix the poor original balance, but without replacing it which I believe is the main argument against any change on this issue.
I strongly disagree. The current situation is that of RVs becoming 'worse' as technology progresses and they start as worse in the first place. I also don't get why adding another viable option would make people forget about trains in the first place. On top of that - RVs have their own disadvantages - like dealing with traffic and congestion which can't easily be remedied. Try a game with some of the aggresive AI ( like AAhogx ), force them to use RVs only and see chaos happen.
All art and vehicle stats I authored for TT and derivatives are as of now PUBLIC DOMAIN! Use as you see fit
Just say NO to the TT fan-art sprite licensing madness. Public domain your art as well.
Eddi
Tycoon
Tycoon
Posts: 8257
Joined: 17 Jan 2007 00:14

Re: Suggestions for long standing game issues.

Post by Eddi »

uzurpator wrote: 24 Mar 2023 21:11 However. That is Transport Fever.
why'd you bring that up? i have about a dozen ideas about this, but it would be highly off topic.
the smooth industries increase all the time by a bit. In a way - the old style 2x production was, in fact, better in that regard.
i think the smooth economy is paid very little attention because most industry NewGRFs will disable this mechanic. And since this is a mechanic introduced by OpenTTD, it won't be subject of the same "don't mess with the base set" mindset as fiddling with the default vehicles. So if you have suggestions here you might actually get anywhere.
Second of all - I'm not an AI. On a large enough map, or dense enough, when one develops one part of a map, the undeveloped part tends to wither away and it is a major annoyance to bring it up to full production.
i agree, my suggestion in this area is that industries that belong to towns without any company rating will enter a "frozen" state. currently, industries have a 5 year protection period from game start (or industry creation), this countdown should only start if the associated town has had any company interacting with it.
User avatar
jfs
Tycoon
Tycoon
Posts: 1749
Joined: 08 Jan 2003 23:09
Location: Denmark

Re: Suggestions for long standing game issues.

Post by jfs »

Eddi wrote: 24 Mar 2023 21:46 i agree, my suggestion in this area is that industries that belong to towns without any company rating will enter a "frozen" state. currently, industries have a 5 year protection period from game start (or industry creation), this countdown should only start if the associated town has had any company interacting with it.
One potential problem with this approach is that it gives less incentives to suddenly move to another corner of the map. A random industry somewhere you've never been increasing production can often be a trigger for making that place interesting to start building. That's part of what makes the game fun IMHO.

On the other hand I do agree from a prebuilt scenario situation, where the scenario designer may have placed industries in specific positions to offer specific challenges. Those industries changing productions, closing, and new ones opening can ruin those designed challenges, and was the trigger for me developing the Industry Control flags feature for GS, so that kind of rules could be implemented.
User avatar
uzurpator
Transport Empire Moderator
Transport Empire Moderator
Posts: 2178
Joined: 10 Jan 2003 12:21
Location: Katowice, Poland

Re: Suggestions for long standing game issues.

Post by uzurpator »

So. Instead of discussing this vacuously, I decided to throw some good old fashioned processing power and simulate things. I've made two simulations.

One - comparing growth of an industry from an arbitrary production level, another showing chances of recovery into full health starting from an arbitrary point.

The 'old' algorithm in this is the current 'smooth economy' - with 33/66 split for declining industries, 66/33 for serviced and 83/17 for serviced excellently. 4.5% chance of production change and 3-23% change in production.

The 'suggested' algorithm works on the same principle for decreasing, but increase is done always in 10-30 units per change. So - a production 50 industry may bump to 80 in a single pop. I also tested 12/88 and 75/25 thresholds for recovery.

I wrote a simple program that runs thousands of iterations of the old and new algorithm and accumulates results.

For 'growth' test I've run a 100 year simulation where each year an industry was tested 1000 times for a production change and these were averaged. The results are that in the low production levels the 'old' algorithm behaves very erratically and, because of rounding errors, might never really make an industry recover. The results I attach are for 30 units/month production. This is because an average growth rate at the 66/33 split is only about 2.5%.

In the 'suggested' algorithm behaves much more consistently and, if an industry is regularily serviced its production will reach, statistically, a certain output.
- for 12/82 split the industry will close
- for 33/66 split the industry will taper off to ~64/month
- for 66/33 split the industry will taper off at ~270/month, after 85 years
- for 75/25 split the industry will taper off to ~420/month, after 120 years
- for 83/17 split the industry will taper off to ~700/month, after 150 years
Attachments
compare_growth_old.csv
(1.94 KiB) Downloaded 32 times
compare_growth_suggested.csv
(2.08 KiB) Downloaded 33 times
All art and vehicle stats I authored for TT and derivatives are as of now PUBLIC DOMAIN! Use as you see fit
Just say NO to the TT fan-art sprite licensing madness. Public domain your art as well.
User avatar
uzurpator
Transport Empire Moderator
Transport Empire Moderator
Posts: 2178
Joined: 10 Jan 2003 12:21
Location: Katowice, Poland

Re: Suggestions for long standing game issues.

Post by uzurpator »

When it comes to recovery rate, the old algorithm also behaves weirdly, as chance of recovery, and period in which it happens, is tied to both chance split and current production level. For the sake of this test, I assume an industry of 25 production to be closed ( game recession excluded ) and I am using 66/33 split. An industry is considered 'recovered' if its production reaches above 100units / month.

In the base game, if we start with an industry on a brink of collapse - production of 32, we get about 50/50 chance to recover it ( remember 66/33 production increase chances ) but even at that, it is going to take about 56 years of game time to happen. Even if we start with an industry with 64 production, we get 90+ chance of recovery, but on average it is going to take about 26 game years to happen.

In the 'suggested' option, the same actions are going to be:
from 32 - 88% chance of success in 11 years
from 64 - 99% chance of success in 7 years
Attachments
recovery_chances_old.csv
(2.6 KiB) Downloaded 31 times
recovery_chances_suggested.csv
(1.97 KiB) Downloaded 31 times
All art and vehicle stats I authored for TT and derivatives are as of now PUBLIC DOMAIN! Use as you see fit
Just say NO to the TT fan-art sprite licensing madness. Public domain your art as well.
User avatar
uzurpator
Transport Empire Moderator
Transport Empire Moderator
Posts: 2178
Joined: 10 Jan 2003 12:21
Location: Katowice, Poland

Re: Suggestions for long standing game issues.

Post by uzurpator »

Lastly - I have reimplemented the algorithm as it is described in the wiki, not necessarily how it works in the game - so there might be discrepances, although statistical properties should remain. If you want to browse through the source - I attach it, at least I hope the forum will allow me to.

From all this, I also reckon that the 'unserviced' chances need to be lowered in the suggested algorithm to 10-15 percent for increase, otherwise zombie industries will never close. I also think that It might be sensible to have an equivalent of the "Do you want to purchase" bankrupt companies for industries. Something like "Flefinbridge sawmill is looking for an investor to reinvigorate the company. Do you want to invest $$$$$" after which the industry will increase production by some random amount. Obviously - only once per game per industry, with maybe a random chance to trigger.

Thoughts?
Attachments
main.cpp
(8.64 KiB) Downloaded 31 times
All art and vehicle stats I authored for TT and derivatives are as of now PUBLIC DOMAIN! Use as you see fit
Just say NO to the TT fan-art sprite licensing madness. Public domain your art as well.
LaChupacabra
Route Supervisor
Route Supervisor
Posts: 385
Joined: 08 Nov 2019 23:54

Re: Suggestions for long standing game issues.

Post by LaChupacabra »

Eddi wrote: 24 Mar 2023 10:01
LaChupacabra wrote: 24 Mar 2023 00:47 a change I'd like to see is to add the effect of the number of transport directions when cargodist is enabled
the general problem here is, how do you differentiate cargos with very few destinations (say, wood to saw mill) vs. cargo with lots of destinations (say, goods)

counting the potential destinations may be a computationally expensive task.
In order not to go offtopic...
[+] Spoiler
...I will just write that I also had an idea of a deeper change in the cargodist itself, so that creating many directions of transport (especially passenger and mail) would make some economic sense. Those would be rather big changes. Adding the influence of the number of directions on the station rating is a much simpler solution, which would give a very similar effect at low cost.
This should have almost no effect on performance. The rating would be influenced by the number of available directions visible in the station window, which is rather easy to count by the game. This way of evaluating transport would only apply to passengers and mail.
Image
Adding such a feature to the cargodist for goods in its current form makes no sense.
If I finally find the time and inspiration to finish it, someday I will publish these ideas in separate topic.
uzurpator wrote: 26 Mar 2023 11:18 ... industry on a brink of collapse...
If the problem is mainly to revive an almost bankrupt industry, why do you want to change the whole mechanism instead of focusing on what you think is the problem?
Why not just increase the chance of production growth when it is low but it is served?
uzurpator wrote: 26 Mar 2023 11:23 I also think that It might be sensible to have an equivalent of the "Do you want to purchase" bankrupt companies for industries. Something like "Flefinbridge sawmill is looking for an investor to reinvigorate the company. Do you want to invest $$$$$" after which the industry will increase production by some random amount.
This was my first post and the first of ideas I presented on this forum. :)
Enterprise menagement.png
Enterprise menagement.png (135.76 KiB) Viewed 938 times
In details it could look different, but in general the possibilities would be similar to those available in the window of local authorities of the city.
Attachments
Station rating for passengers and mail.png
(79.08 KiB) Not downloaded yet
I am sorry for may English. I know is bed.
User avatar
uzurpator
Transport Empire Moderator
Transport Empire Moderator
Posts: 2178
Joined: 10 Jan 2003 12:21
Location: Katowice, Poland

Re: Suggestions for long standing game issues.

Post by uzurpator »

LaChupacabra wrote: 26 Mar 2023 17:19 If the problem is mainly to revive an almost bankrupt industry, why do you want to change the whole mechanism instead of focusing on what you think is the problem?
Why not just increase the chance of production growth when it is low but it is served?
TL:DR: Because now I understand that the problem is the algorithm itself, and it being based on compound percentages.

My initial problem always was just how much of a pain it is to get good production going on with slower vehicles is hard. In OTTD this turned to be even more of a pain for reasons I didn't fully appreciate until this thread.

The general problem is the shape of the red curve ( here 83/17 split ):
shape.JPG
shape.JPG (102.79 KiB) Viewed 926 times
The red is production growth averaged over thousands of runs. The dark blue is the suggested change. If you keep perfect service over the entire 48 year period, then starting from 30, this is what, on average will happen.

Let me repeat - for _growth_ the suggested change is to add 10-30 flat bonus to an industry, but for _reduction_ use the current algorithm. The benefits:
- the production grows stably if service is stable, even with 66/33 split
- no more sudden productivity explosion once you roll TGVs, thus get 80+ ratings
- no more waiting for several decades until zombie industries die off or get brought back to life
- zombie industries can be reliabily competed for and rescued, thus most games are not going to be littered with them
- because all sources will taper off, there will be more incentive to build complex transfer networks and such
- lastly, effing up on service will have much more severe consequences, relatively speaking, as industry grows on the blue curve, but declines on the red one
All art and vehicle stats I authored for TT and derivatives are as of now PUBLIC DOMAIN! Use as you see fit
Just say NO to the TT fan-art sprite licensing madness. Public domain your art as well.
LaChupacabra
Route Supervisor
Route Supervisor
Posts: 385
Joined: 08 Nov 2019 23:54

Re: Suggestions for long standing game issues.

Post by LaChupacabra »

uzurpator wrote: 26 Mar 2023 19:11 The dark blue is the suggested change.
As a separate economy type setting: Why not! While I think the enterprises shutdown acceleration element is a bad idea (quickly clearing the map or too often changing on the map), overall gameplay with lower levels of production can be quite interesting.

As a revolution of the smooth economy mode without the ability to play as before: I definitely don't like this idea! These changes are too radical and change the nature of the game too much.
I am sorry for may English. I know is bed.
Post Reply

Return to “OpenTTD Suggestions”

Who is online

Users browsing this forum: Bing [Bot] and 5 guests