War Effort Campaign

[RayofSunshine]

I was surprised that there wasn't any thread for this scenario, considering that there are a number of different ways to reach the goals.

With that said, I don't see any 1 strategy being better than another, and it is a fun type of scenario, basically if you just like to watch the trains transportng required around the map. But it does take a little "thought" for a strategy for the goals.

This last attempt was considerably different that my first attempt years ago. Maybe because or more experience and strategies over the years of playing all types of challenges.

The reason I mention a difference, as the 1st attempt had me barely getting the last commodity to the port within the time limit of 6am Dec 16th. This time the goals or 10 each commodity were completed by Dec 14th by noon.

With time to spare, I redirected the 3 incorporated engines 2 of which were steam, to the farthest depots for the remaining spare commodities. The game ended, but I kept playing, and was able to complete W-10 A-15 F-20 Cz - 20 CL- 20 and M- 12 by noon on the 16th.

So there is a difference on the way a person sets up their itenaries. Now I may have also played using electric, which is also an option, but do not remember the result.

There are so many new creations all the time by user players, but I like to go back to some of the earlier scenarios to just enjoy playing them for fun.

[sleepy]

This was by far the most stressful and hardest campaign I've played so far. The other maps I finished well ahead of time. My experience was greatly improved with this greatest American marches compilation playing in the background, though.

I ended up following the basic strategy laid out in gameFAQs, because I really did not want to play this map five or more times over. This means I used the EP-2 Bipolar, Pacific 4-6-2, and the GG1. I had all the weapons, ammo, and fuel delivered first, then I shipped the clothing, cheese, and meat. The routing is almost entirely sending one type of good to Norfolk or New York and having the train return empty to a new area. I think this is inefficient and could be improved by dropping off goods at intermediary stops. However this scheme can't be used entirely and sometimes I have to ship multiple types of goods in one train.

Even though the Pacific 4-6-2 broke down once during the game I still managed to deliver all the weapons, ammo, and fuel by 2:20 AM, December 14th, and to finish by 3:45 AM on the last day. This is not as good of a win as I'd like but I really do not want to play this map again for a while (several mistakes were made, such as accidentally sending 1 ammo and 5 weapons instead of 1 ammo and 6 weapons )

Interestingly enough, since there's nothing dynamic about this map (other than maybe a train breaking down), the routes can be planned entirely ahead of time on paper. The first three items are sent to New York and the last three are sent to Norfolk. I might have been able to cut down on distance using full cars instead of half full ones. On the other hand, the train speed scaling for number of cars is not linear and a train with four cars is almost as fast as an empty train. I tried to use cabooses everywhere I could, but breakdowns did not take as long as I expected.

This is essentially a direct application of graph theory. For companies like Amazon, with millions of products and different schedules, the challenge is essentially the opposite: ship items from centralized locations outward. I'd love to see someone with the technical know-how produce a fully or nearly optimal solution to this challenge. Sadly RayofSunshine was incredibly vague in his methods but I'd also love to see the other strategies people have come up with (has anyone chosen the last option to make the trains faster? I can't imagine it being that much faster).

[Gumboots]

This is not as good of a win as I'd like but I really do not want to play this map again for a while.

I know what you mean. I played it once or twice, but didn't find it very appealing (and didn't get a Gold either).

On the other hand, the train speed scaling for number of cars is not linear and a train with four cars is almost as fast as an empty train. I tried to use cabooses everywhere I could, but breakdowns did not take as long as I expected.

Speed scaling for number of cars is exponential. Each added car reduces speed by the same percentage. Mind you, this is modified by the loco's "pulling power" and "free weight" values. There's a spreadsheet I made attached somewhere around here, if I can remember where it is.

And don't trust the speed stats that are displayed in the game. They are known for a fact to be wrong.

Edit: Found spreadsheet. It's here - Speed _on_grades_RC2.zip. This thing is dead accurate to within 1 mph at top speeds of up to 1,000 mph.

[sleepy]

Yes, I started playing again so I can make lots of money laying track and buying trains as a tycoon, not route three trains. It doesn't help that I'm slow with the UI.

The spreadsheet is quite amazing - I don't know what kind of magic went into finding those formulas.

I've been doing research (google) into routing heuristics for getting close to the optimal solution to this map. I don't expect to actually get anything useful but it'd be fun to see how far someone can get on this simple map. I think the challenge is modeled closest by the vehicle routing problem. There are lots of papers in this area and it is well studied. Obviously things are complicated by service and maintenance stations, trains breaking down, etc. The best hope for someone who had the knowledge and motivation to solve this map I think would be jsprit. I'm moving on to the next campaign, though.

[Gumboots]

Good idea to use jsprit if anyone was keen on an optimal solution. I'm not that enthused about this map myself. I prefer more open-ended and varied scenarios.

The formulae in the spreadsheet were found by reverse engineering the whole thing, using tricks such as setting top speed to 1,000 mph (reduces rounding error) and setting "pulling power" and "free weight" (two variables in the .lco file) and loco weight (obvious) to simple values like 0 and 1 to enable teasing out what was happening. There's a fairly good stream of consciousness blurb about it all in the relevant thread, including false trails: Speed adjustment considerations

If you're curious, it's probably best to start at this post so you skip a fair bit of crud.

[RulerofRails]

This one is limited in strategy options, but I don't mind it as a quick logistics challenge. I have played it a number of times, but it doesn't rank among my favorites. I decided to give it a little whirl. Took maybe an hour and got it finished at 1:09 am on Dec 14. Started with the EP-2 Bipolar, then the Pacific and finally the GG1. Not near the limit of possibility. I'm not enthused enough about the map to try to get an optimization. A big part of this is my lack of any knowledge about any of that or any other such tools. I guess that the grades would have to be mapped and all the legs calculated for distance accurately.?.

---

I took a look in the editor after playing. Cue release of spoilers.

Turns out that the total number of loads hauled to the ports territory (both Norfolk and New York count) is triggering the new engine/increased speed option at 15 and 35 loads hauled respectively.

Takeaway: as soon as you get two full loads to the ports you will get a new engine.

The map has a 50% reduction in engine speed applied at the start (this time I realized this without looking in the editor, but looking confirmed). In the option for a third engine OR to increase engine speed by 75%, the 75% increase is a false statement. Poptop fell over themselves here. It is really 2.5x (1.25/0.5) locomotive speed. Probably worth going for. I haven't in the past.

Let me explain:
As most of the folks here know, the RT3 editor doesn't have true percentage modifications. Instead, its "percentages" only ever act upon the ORIGINAL value. This is a problem when applying the second, third .... etc. percentage modification. We need to apply the percentage we want manually to the CURRENT value. Then put this calculation into the event.

For example, a first event of "-50%" gives a halved value, works correctly because it acts upon the original value. To then increase by 20%, if we do "+20%" the result is skewed because the game will make it -30% (-50%+20%=-30%) of the original value.

To do the 20% increase correctly, we calculate the CURRENT value compared to the original which here is 50% (100%+(-50%)). Then we apply the percentage manually. So 50%*20%=10%. We are increasing so we choose the effect to add ten percent (+10%). Once applied, the editor will tell us that we have -40%. This is what the game sees. We are now at 60% (100%+(-40%)) of the original value.

On this map, we realize that we are at 50% of the original value because of the -50% event already in effect. For the correct 75% increase we calculate 50%*75%=37.5%. So we make the event for +38% (decimals aren't recognized as far as I know). Which, when applied, will be -50%+38%=-12%. The game would see -12%. We are at 88% (100%+(-12%)) of the original value.

[RulerofRails]

I might have discovered an issue with the map. After choosing the Pacific near Charlottsville I realized that I cannot get it to travel on the track from Charlottsville to Dale City. It will always go the long way around down to Richmond. I tried waypoints etc.. If you care, check if you can use that track.

Because the Pacific isn't powerful enough to go up through the tunnel towards Harrisonburg at more than 1mph, that leaves only the south clear for it and I'm almost sure this is a nail in the coffin for steam's chance to come close to electric on this map.

[RulerofRails]

I did another start choosing the EP-2 Bipolar. I took the faster engines choice over the GG1. Result after two breakdowns was completion at 9:35pm on Dec 13. The medal fired at 10:00pm. Each breakdown loses 1 hour. The last one definitely lost me an hour on time to complete.

Only trick I use is to add the caboose onto the train after all 8 cars have been assigned for the run but are still loading. This way I can have 8 cars per train AND a caboose. One could cheat the crashes by re-loading a saved game, but if one is to cheat, may as well go into the editor and flick the breakdown switch.

Because this scenario is pretty simple, I made an outline of my strategy. I'm not suggesting that this is the best strategy as some factors are beyond my brainpower and limited time to consider. Here goes for whatever it's worth.

Start game:
Choose engine EP-2 Bipolar
Philadelphia (6 meat, 2 weapons) -> New York
Baltimore (6 weapons, 2 cheese) -> New York
(unlock Pacific or Northern, I picked the Pacific see below for it's itinerary)
Dover (4 ammo) - Allentown + (4 ammo) -> New York . . . .breakdown on this run
Salisbury (6 diesel) - Philadelphia + (2 weapons) -> New York
Winchester (1 weapons) - Dale City + (3 diesel) - Baltimore + (4 cheese) -> New York


Pacific
Richmond (4 weapons) - Petersburg + (4 ammo) -> Norfolk
Lynchburg (8 clothing) -> Norfolk ADD the Lynchburg service tower as a stop or the train will miss it (too close to the junction)
(unlock GG1 or increased speed, I picked increased speed)
Roanoke Junction (4 cheese, 4 meat) -> Norfolk
Roanoke Junction (2 clothing) - Danville + (6 diesel) -> Norfolk
Washington D.C. (3 ammo) -> Norfolk . . . .breakdown on this run

[sleepy]

It is really 2.5x (1.25/0.5) locomotive speed. Probably worth going for. I haven't in the past.

Let me explain:

Wow, that is a ton faster It's strange the game designers chose to use such a convoluted system though. It really would be as simple as multiplying by a decimal once.

I believe the slopes slow the train down too much. In fact I think the game would be more fun if the trains traveled faster but more goods had to be delivered. And, the main reason I didn't enjoy this map much is explained by this John D. Rockefeller quote:

Do you know the only thing that gives me pleasure? It's to see my dividends coming in.

RulerofRails, your routing abilities are incredible. I see you chose the full train method, which was a good idea because what I used wasted a lot of time going back and forth between cities. I had no idea about the caboose trick. I only had one breakdown from the Pacific I believe (the EP-2 Bipolar is pretty reliable). My schedule also always delivered the weapons, ammunition, and fuel first before everything else. This is at the obvious cost of efficiency. What's strange is that I assumed all the goods of one type had to be delivered all to one destination city. This is what's stated in the gameFAQs strategy, although I've only played five campaigns and I already notice some errors (Central Pacific starts in 1855 not 1850, and the train speeds for this campaign are listed at twice their actual value?). Maybe it was written for a previous version of the game (I'm playing 1.05).

[RulerofRails]

If you delve into the workings of the game and hang around the forums you will find quite a few bugs and things that could have been done better by the programmers.

The speeds on slopes are accurate when using the correct loco stats that Gumboots linked to. The ones we see in the game are too slow which makes it appear that trains are generally rubbish on hills.

A shortcut method I use to get a better idea of real loco performance without using a spreadsheet is to select the "Express" car weights from the in-game display. This still reads slightly slower than reality (because of the weight that the express cars are listed at) but it's close enough to get a reasonable idea for freight at a glance. The real car weights for freight progress over time, in the default game it's 5 tons before 1850, 10 tons 1850-1900, 20 tons 1900-1950, and finally 40 tons 1950 onward. Yep, the notices of doubling in car weight that the game displays are wrong.

If you weren't talking about that, but instead the reduction of speed on this map; I agree that the locos are slow if you sit and watch them. I just turn the game up higher to Fast or Very Fast setting which works well enough that it doesn't bother me. If the map was re-done, sure it could do with some more things to add interest and more difficult tasks.

The gameFAQs strategy mistakes are likely true mistakes. Some of the early stuff for the game isn't well-polished.

[Gumboots]

Some of the late stuff for the game isn't well-polished either.