I've scrapped the old post about the last customised pax appeal scale I tested. After sleeping on it and trying some things, I came up with a better plan. This new scheme doesn't require changing the cargo base price, so getting a good game balance is easier. It works, but still needs more testing to work out optimum ratings for a range of locomotives.
So far I have tested with 6 Berkshires and 3 Schools class, all trains running 7 cars auto consist plus a caboose, on my 1953 Royal Tour map. The 3 Schools class were all benefiting from a 20% speed record bonus (ie: stars) but no trains were using dining cars.
Each test ran for one year from the same starting point. Six tests were made for the default pax appeal settings: Berkshire rated Acceptable (hex code 01), and Schools rated Looks Sharp (hex code 02). The annual express revenues for the six years were then averaged. This average was used as the baseline figure for comparisons.
I then reset the language file and the locos so the Berkshire was rated Even Lower (hex code 0B) and the Schools was rated Very Cool (hex code 11). This gave annual express revenue 7% less than the baseline. Average freight revenue wasn't increased to compensate, due to the way the game engine handles pax appeal (see notes further down).
The Berkshires were then set down to Almost None (hex code 20) with the Schools class not being changed. This gave annual express revenue 15% less than the baseline. I then set the Berkshires down to None At All (hex code 21). This gave annual express revenue 27% less than the baseline. So far it's all feeling about right for a workable system. The proposed new ratings look like this:
Code: Select all
;@gumbootz: Customised passenger (pax) appeal ratings begin here.
; A complete list of custom pax appeal ratings follows.
; -----------------------------------------------------
; #ID Rating Name Revenue Hex value
; -----------------------------------------------------
; @2703 - Ultra Cool 140% 12 00 00 00
; @2702 - Very Cool 130% 11 00 00 00
; @2688 - Looks Sharp 120% 03 00 00 00
; @2687 - Looks Nice 110% 02 00 00 00
; @2686 - Acceptable 100% 01 00 00 00
; @3173 - Tolerable 90% 00 00 00 00
; @2697 - Quite Low 80% 0C 00 00 00
; @2696 - Even Lower 70% 0B 00 00 00
; @2695 - Very Low 60% 0A 00 00 00
; @2716 - Ultra Low 50% 1F 00 00 00
; @2717 - Almost None 20% 20 00 00 00
; @2718 - None At All 0% 21 00 00 00
In this system the Ultra Cool level is boosted by 20%. This makes it the same as the old Ultra Cool with a speed record or dining car bonus. The scale could be extended to even higher levels (150% revenue, etc) if people think that is necessary.
Very Cool gives the same revenue as the old Looks Sharp with a speed record bonus, or dining car bonus. Looks Sharp is now at the same revenue as the old Ultra Cool. Looks Nice replaces the old Looks Sharp, at the same revenue. Acceptable is what it always was. Tolerable replaces the old Ugly, at the same revenue. Four new levels follow below Tolerable, at even 10% revenue reductions.
The Almost None level then takes a jump 30% down in revenue to 20%, because there are no available hex values that would allow revenues of 40% or 30% (these hex values are taken up by necessary building names).
There are no available hex values for 10% revenue, so the last level is down to 0%. Locos rated None At All would trash profit if run with auto consists on main lines. Main line usage would require a freight-only consist. OTOH, if you were hauling freight on a branch line but occasionally got a load of pax you wanted to feed down to your main line, using a None At All loco on auto consist could work.
When testing all of this I found something surprising. If the basic delivery price for pax on a route is $25k/load, a locomotive using an auto consist
will load pax cars on the basis of that $25k/load price regardless of its passenger appeal rating.
If the station is full of pax at a nominal $25k/load and freight at $23k/load, a train rated Almost None, or a None At All, will only haul pax if it is using an auto consist, despite the fact that once the pax appeal is applied the consist will earn little or no revenue.
This applies to all passenger appeal ratings, as far as I can tell. Even on the higher ratings, which boost pax revenue by 20% or more, cars are added to the consist on the basis of their nominal delivery price
without regard to the effect of the passenger appeal rating.
This is a bug, but we're stuck with it. It does partly explain the odd behaviour of passengers in RT3. A load you might expect to be hauled may not be hauled, because the effect the appeal rating has on delivery price is not considered when loading.
I've attached a zipped version of the custom RT3.lng file for 1.05. Note that I haven't looked at the 1.06 file yet, and it probably contains extra language strings. These won't affect the new pax ratings scale, but the extra strings would be required for 1.06 to run properly. If you want to try the new pax appeal scale in 1.06, applying the same edits to the 1.06 RT3.lng file should work.
For 1.05, just back up your existing file and throw this new one into Data/Language. You can then play with the pax appeal ratings in your .lco files and see how it all works in practice.
There is one small catch to this. Extending the pax appeal scale requires hijacking a few of the text strings that are used for freezing time when creating a new sandbox. Most of them remain as they were, and IMO the interface is still usable. Screenshot attached.
I just realised something else. When I first went to change the hex values I looked at what RoR had posted and thought ok, so he used a value of 18 in the hex, so 18 converted to hexadecimal is 12, so I put 12 in the .lco file and saved it. It didn't show properly in the game.
