News:

Simutrans Wiki Manual
The official on-line manual for Simutrans. Read and contribute.

Making of an Extended pakset

Started by 1993matias, January 10, 2022, 11:34:21 AM

0 Members and 1 Guest are viewing this topic.

1993matias

Hello,

Currently, I am trying to help pak192Extended with vehicle data. I want to get it right the first time, so before I start entering the vehicle capacity, I would like to ask the experts: How do you best figure out the definition of a class? E.g. is modern 2nd class low or medium class?
Also comfort, how do you find what the comfort rating of a vehicle actually is without guessing? It's obvious that a sleeper is more comfortable than an IC-coach which is more comfortable than a Tube train. But exactly how much is the difference? That is the difficult part...

Do you have some kind of guidelines you follow when you add a new vehicle? Could that be something you would like to share? Since you have already done so much work with Extended and all the cool features, then it would make little sense for the team behind pak192Extended to make the same mistakes again
Thank you!

jamespetts

It is lovely to see Pak192 have Extended data added. Balancing a pakset is complex and requires a mixture of detailed historical research, mathematical calibration and in-game trial and error. Even balancing Pak128.Britain-Ex is still a work in progress, especially as balance critical features are awaited. I deal with the two specific queries below.

Class

Pak128.Britain-Ex has five classes, but the number of classes in a pakset can be set by the pakset author. Bear in mind that having too many classes may cause the path explorer to run too slowly on larger maps.

In Pak128.Britain-Ex, the idea is for classes roughly to correspond to a number of historical circumstances. It was originally intended that Pak128.Britain-Ex have only three classes, corresponding to first, second and third class rail travel in the 19th century, but it turned out that 5 were really needed, as (1) we need something below third class to represent the fourth class that was on some trains in the 1850s, passenger canal travel in the mid 19th century that was cheaper but slower than third class rail travel, and, more importantly, the tram and later 'bus fares that were cheaper than third class rail fares; and (2) we need something above first class to represent Pullman cars on the railways (for which first class passengers would pay extra) and some very specialist forms of transport, such as the early jet airliners, Concorde and the 1830s luxury bed carriages.

For other instances, the process is simply to find an approximate equivalent. For example, 'bus travel in the days of horse 'buses was too expensive for many people because of the economics, but wealthier people had their own carriages - so the wealth level for the early horse omnibuses was estimated as "medium". As better horse drawn vehicles with higher capacities and lower resistances were introduced, the price came down somewhat, so these are set to "low". Trams could be cheaper than 'buses because fewer horses could pull more people, so this is set to "very low". It is all a matter of finding analogies and extrapolating from them.

Comfort

For Pak128.Britain-Ex, there is a comfort calibration spreadsheet here. What you need to work out is the maximum comfortable journey that one could sensibly make on a vehicle, and then calibrate the comfort according to this and the equivalent settings in the pakset. . Small adjustments can be made simply by ranking vehicles against each other: for example, one might say that an average 'bus has a comfort level of 40 based on the maximum comfortable journey time of 46:15 - then one might say that one particular 'bus has slightly more comfortable seats and so add one or two points, or that another 'bus is rather draughty and reduce a few points, always keeping a check in a spreadsheet to make sure that the ranking of comfort among different types of vehicles is coherent and makes sense.

**

Aside from those two particular things, the general starting point is to try to use real world data where possible, as real world data are already calibrated for you.

Very best wishes with this!