The Great Experiment
Posted: Thu Jan 31, 2008 9:32 pm
I have created (well, modified to an extent) a map for the sole purpose of testing the efficiency of train routing in RT3.
I have been playing for a long time, and I still haven't been able to find the "perfect" route setup. I've tried almost everything, and I am looking into finding the "most efficient" scheduling.
There are several factors that contribute to how efficient scheduling is, most notably availability of stuff to send from one city to the next, the next being the ability to move said stuff.
The test map is in the rework phase, since I screwed up some major things that I didn't take into account on the first trial runs; however, the data collected so far presents a promising baseline, and a wealth of information. The map is currently 30 cities, arranged in 3 groups of 10, each group subdivided into two parallel strips of 5 cities each. No starting city has more than 3 stars for size, and each strip is arranged as 1-2-3-2-1 and 3-2-1-2-3 (star size) respectively. Each grouping of 10 cities is equidistant from the next and are arranged in short, medium, and long distance ranges, to determine if distance from each city is a factor. In the next phase, I intend to add 10 more cities, at "extra long" distances from each other; however, this may skew the results for "overall" performance. Additionally, I intend to add 10 cities that will not be connected to rails at all, as a baseline set to determine how effective city growth is with the railroad vice without.
Currently, the tests are to be run in 8 categories, trains that run 2 stations (1-2), three stations (1-2-3-2), four stations (1-2-3-4-3-2) and five stations (do you really need to see this one? ), with first batch runs as (min/max) 1/8 any cargo, followed by second runs as 1 train freight 1/8, 1 train express 1/7+DC for each route, for a total of 8 run-throughs. Each run will last 60 years, with data observation every ten years, data collection at 30 and 60 years.
Making this scarier, all 8 runs should be repeated in 30 year intervals to determine if locomotive speed is a major determining factor into a railroad's efficiency (I know it is, but I want to be thorough), starting in 1840 and ending in 1960 (for 32 runs, at about 2-2 1/2 hours each (which, in turn, makes me wish that I was getting paid to run this...)). However, this may be mitigated once the "best" schedule is determined, and just running that set up for each time period.
There is no mountains, rivers, or other hindrances to railroad development, however rivers/streams/hills may be added later to better divide the groups of 10 if cross-map-non-railroad moving of cargo is compromising data results.
So what does this mean to you? Probably not much. It is a "best case" style test map, goods can flow (as seen in already completed runs) freely without assistance of the rails. This is to test how efficient each style of scheduling is at taking care of the customers. Data may be used on other maps to a degree, however, environments native to other maps (such as hills, valleys, rivers, etc.) may make other options more viable. At the very least, there would be a starting point.
So far, using Test Map v 0.01, I've run both runs of the "two stations" trains. So far, based on data after 30/60 years, city development is faster using trains that are set to "Any Cargo" but overall profit is better using 1 freight/1 express on the same routes. Company cash was a little shorter on this setup due to having twice the number of trains. Stock value and stock book value were less using the 1f/1e schedules, however, player's stock was double that of the any schedule.
Of note, the scenario is set up with a ridiculous amount of company and player cash, no need to issue stock or bonds. No industries were purchased (didn't want to have them be a factor, it is the RR's efficiency I am testing, after all). No crashes/breakdowns allowed (again, this would be a factor in the final results). Medal win conditions not altered from the original map, but the win text was changed, as was the lose text. With the exception of the size and win conditions, no elements from the original map exist as original, but I'll still give credit to the original "writer" once I can recall who it was.
Changes for Test Map v 0.10:
-Cargo counter
-add ten cities at longer distances than cities 21-30
-add ten cities mirroring cities 11-20, but not placing rail.
-move edge of map cities away from the edge of the map, they were unable to grow (one shrank).
-remove (if able) economic conditions. Map should be forever in "normal" economy.
-rezone cities to place high demanding products further away from their demanding cities.
The following are criteria for observation, all are recorded at 30 and then 60 years:
-# of 1 Star Cities
-# of 2 Star Cities
-# of 3 Star cities
-# of 4 Star Cities
-# of 5 Star Cities
-CBV
-PNW
-Personal Cash
-Company Cash
-Player Stocks
-$/share
-Book value
-Company Profit
-Loads of Freight Hauled (currently recorded as a cash value, v. 0.10 will have it as # of loads)
-Loads of Express Hauled (same as above)
-Year of Gold (With only 30 years to get there, the test map has cut it very close so far on each run)
-# of trains
-Salary (I feel that this one is a good judge of how well the game thinks you're doing, so it is a factor)
I will continue to observe my trains and report my findings, however, due to the time to run, record, rerun, etc., I will probably report every two runs. Any input/help/comments/suggestions/questions are greatly welcomed.
Czar "Map will be available, too, when I'm done tweaking it" Mohab.
P.S. Yes, Hawk, et al, it is Barney with Captain Kirk.
I have been playing for a long time, and I still haven't been able to find the "perfect" route setup. I've tried almost everything, and I am looking into finding the "most efficient" scheduling.
There are several factors that contribute to how efficient scheduling is, most notably availability of stuff to send from one city to the next, the next being the ability to move said stuff.
The test map is in the rework phase, since I screwed up some major things that I didn't take into account on the first trial runs; however, the data collected so far presents a promising baseline, and a wealth of information. The map is currently 30 cities, arranged in 3 groups of 10, each group subdivided into two parallel strips of 5 cities each. No starting city has more than 3 stars for size, and each strip is arranged as 1-2-3-2-1 and 3-2-1-2-3 (star size) respectively. Each grouping of 10 cities is equidistant from the next and are arranged in short, medium, and long distance ranges, to determine if distance from each city is a factor. In the next phase, I intend to add 10 more cities, at "extra long" distances from each other; however, this may skew the results for "overall" performance. Additionally, I intend to add 10 cities that will not be connected to rails at all, as a baseline set to determine how effective city growth is with the railroad vice without.
Currently, the tests are to be run in 8 categories, trains that run 2 stations (1-2), three stations (1-2-3-2), four stations (1-2-3-4-3-2) and five stations (do you really need to see this one? ), with first batch runs as (min/max) 1/8 any cargo, followed by second runs as 1 train freight 1/8, 1 train express 1/7+DC for each route, for a total of 8 run-throughs. Each run will last 60 years, with data observation every ten years, data collection at 30 and 60 years.
Making this scarier, all 8 runs should be repeated in 30 year intervals to determine if locomotive speed is a major determining factor into a railroad's efficiency (I know it is, but I want to be thorough), starting in 1840 and ending in 1960 (for 32 runs, at about 2-2 1/2 hours each (which, in turn, makes me wish that I was getting paid to run this...)). However, this may be mitigated once the "best" schedule is determined, and just running that set up for each time period.
There is no mountains, rivers, or other hindrances to railroad development, however rivers/streams/hills may be added later to better divide the groups of 10 if cross-map-non-railroad moving of cargo is compromising data results.
So what does this mean to you? Probably not much. It is a "best case" style test map, goods can flow (as seen in already completed runs) freely without assistance of the rails. This is to test how efficient each style of scheduling is at taking care of the customers. Data may be used on other maps to a degree, however, environments native to other maps (such as hills, valleys, rivers, etc.) may make other options more viable. At the very least, there would be a starting point.
So far, using Test Map v 0.01, I've run both runs of the "two stations" trains. So far, based on data after 30/60 years, city development is faster using trains that are set to "Any Cargo" but overall profit is better using 1 freight/1 express on the same routes. Company cash was a little shorter on this setup due to having twice the number of trains. Stock value and stock book value were less using the 1f/1e schedules, however, player's stock was double that of the any schedule.
Of note, the scenario is set up with a ridiculous amount of company and player cash, no need to issue stock or bonds. No industries were purchased (didn't want to have them be a factor, it is the RR's efficiency I am testing, after all). No crashes/breakdowns allowed (again, this would be a factor in the final results). Medal win conditions not altered from the original map, but the win text was changed, as was the lose text. With the exception of the size and win conditions, no elements from the original map exist as original, but I'll still give credit to the original "writer" once I can recall who it was.
Changes for Test Map v 0.10:
-Cargo counter
-add ten cities at longer distances than cities 21-30
-add ten cities mirroring cities 11-20, but not placing rail.
-move edge of map cities away from the edge of the map, they were unable to grow (one shrank).
-remove (if able) economic conditions. Map should be forever in "normal" economy.
-rezone cities to place high demanding products further away from their demanding cities.
The following are criteria for observation, all are recorded at 30 and then 60 years:
-# of 1 Star Cities
-# of 2 Star Cities
-# of 3 Star cities
-# of 4 Star Cities
-# of 5 Star Cities
-CBV
-PNW
-Personal Cash
-Company Cash
-Player Stocks
-$/share
-Book value
-Company Profit
-Loads of Freight Hauled (currently recorded as a cash value, v. 0.10 will have it as # of loads)
-Loads of Express Hauled (same as above)
-Year of Gold (With only 30 years to get there, the test map has cut it very close so far on each run)
-# of trains
-Salary (I feel that this one is a good judge of how well the game thinks you're doing, so it is a factor)
I will continue to observe my trains and report my findings, however, due to the time to run, record, rerun, etc., I will probably report every two runs. Any input/help/comments/suggestions/questions are greatly welcomed.
Czar "Map will be available, too, when I'm done tweaking it" Mohab.
P.S. Yes, Hawk, et al, it is Barney with Captain Kirk.