Creation Sharing New Timetable For Dresden - Riesa

Yeah, the physics on the 112 are completely wrong. It'll look nice as AI though.

 

I still haven't done the last RE15/18 & RB31 services - I got sidetracked a bit. With interesting results though. Welcome to the Nünchritz Trials:



Let's take a step back. If you have the G6, you're probably aware of a bug where on some shunting services you get kicked out of the game for a supposed SPAD when you were in fact pushing wagons through a clear signal. I'd assumed that to be a bug with the signalling on the RRO route, especially since I had already done a whole bunch of weird stuff in Free Roam on DRA. Well, until I decided I wanted to try creating a shunting service at Nünchritz Chemical Works yesterday and the simulation failed saying that the AI had SPADed! I tried playing in Editor, and sure enough, it wouldn't let me push the train into the Chemical Works...

Thing is, I had pushed stuff in there in Free Roam quite a few times before, and after a bit of thinking I noticed that on all my Free Roam attempts I'd used the Zacns tankers while my service had Sggmrss container flatcars* too. So then I set up the scientific trial seen above to test my hypothesis. 3 363s, one with Eanos open wagons as the control group, one with an Eanos in front of some Sggmrss to see if Sggmrss only don't work as the first car or not at all, and then one with Sggmrss only, just to confirm that Free Roam and a "real" service behave the same way.



Long story short, the first two trains made it in while the third one gave me the SPAD error.

That means we've confirmed two things:

1. There's something wrong with the Sggmrss wagons that makes the game think you SPAD when you push them through clear signals
2. It only happens if the first car of the train is a Sggmrss - if you put any other vehicle in front of them, they're fine

Gotta apologize to whoever did the signalling on RRO, I'd always just assumed the problems with the G6 services were their fault. Anyway, once I've rearranged my formations to make my shunting service work, I'll finally get the Elbe-Elster trains done - I just felt like doing something else over the last few days.

---

* After all, there is a siding with a container crane in there...

 

I am just now adding my first sggmrss push services at Aschaffenburg Hafen, so thank you a lot for bringing this to our attention just in time

 

The time of testing timings in an empty rail network is finally over! Yesterday's 363 shenanigans were a welcome palate cleanser, and I played and edited the last two Elbe-Elster trains today.

Next, I'll assemble a two hour "slice" of finished timetable from all the services I've made, and get everything to interact properly.

 

Great news! I'm really looking forward to this one

 

After a fun morning of copying and pasting services, I just simulated my two-hour slice of full (-ish) timetable for the first time, and...



...honestly, that's much better than I had expected. Nothing gets stuck forever, which is a huge win already. There are 20 to 30 services active at any time (and remember, that's passenger trains only!) and Dresden itself is pretty complex, so I'm super happy with that.



I did also go grab a quick screenshot, just to whet your appetite a bit. On the upper level, a EuroCity from Praha hl. n. is pulling into platform 12. On the lower level, from left to right, an S3 from Tharandt is arriving while a Trilex, RE50 and S8 are waiting for their next services. Not seen here are the RE3 and RE15 that left the station just before the EC came in.

Seeing Dresden this busy is really exciting for me - now I'll "just" need to make it work properly...

 

Is the coupling for the RE1/2 and RB60/61 also included? Because as far as I know the 642 cant uncouple

 

Not yet. The point of this is to get the mainline runs timed right. Proper diagrams, shunting, ECS runs etc will happen later.

I believe you can work around it by turning on the automatic coupling/uncoupling feature.

 

Having started debugging, I'm very satisfied with my new workflow. Getting all issues, for all lines as they interact with each other, is so much nicer than sequentially building full layers. I just go through the log file and fix the first stop signal issue before resimulating. Doing any more than that in one go doesn't make much sense as any later issues might well be caused by the first one, or the first problem might hide other ones that would occur later if trains were on time.

Now you'd think that doing it this way would take absolute ages given that there were more than 1600 stop signal complaints originally. I really won't need to resimulate the timetable anywhere close to 1600 times though, as the very first change I made brought the number down to just over 550. When one train grabs its route between Dresden Hbf and Neustadt too early, it really just snowballs into huge pile-ups sometimes. That means one tiny fix - often just a "don't dispatch beyond" setting in the right place - can work absolute magic.

 

I hope it will stay this easy to fix for you, as simulating time will increase a lot (even exponentially) once you reach a certain amount of time and services. 'just resimulating' becomes a real bortleneck then. Hopefully you can wrinkle out enough issues before copy-pasting.

I do think your approach is creative, efficient and effective though, especially for a complicated map like DRA.

 

Well, that's the nice thing about this approach, isn't it? The amount of time and services is staying exactly the same until I've fixed all the issues I can find. But yeah, once I'm done with this step of the work, it'll get much slower very quickly...

Either way, my new approch already proved itself yesterday: the second issue I had yesterday was an RE18 stealing a path from the S1 at Coswig. I added a "don't dispatch beyond" on the RE18, only to discover it made everything worse: the S1 still wasn't on time as an RE50 was in its way, but now it was also blocking the RE18. Because I already had the whole timetable, I could immediately see that second issue, which I might not have done if I'd been building whole layers one after the other, and because I only have two hours' worth of services, making and then reverting the changes was very quick and painless.

 

I've now resolved a good number of issues. Quite a few trains that were previously delayed now make it into Dresden on time. This had the interesting effect of making the overall performance of the timetable much worse: it was back at 1600 stop signal complaints at one point. I can only assume that this is because, by being late, the trains I've worked on previously avoided some kind of massive pile-up in Dresden.

 

Das hört sich ja gut an, That sounds good

 

Do you use software for this graph, or do you simply keep track of it in an Excel file?

 

Just Excel.



Here's what it's looking like right now. I was studying for university pretty much all day today, and I had the timetable simulation tool running in the background. Whenever it was done simulating, I took a 5 minute break from studying, fixed one issue, and started the simulation again. This was a surprisingly productive way to handle both tasks. I now feel prepared for my exam tomorrow, and I'm also down to just 47 stop signal complaints. Not a typo. Just 47.

 

Fingers crossed for the exam!

 

I've got another day of studying and debugging ahead of me, but before that, I treated myself to a first drive of my complete passenger timetable. The chronological nature of my current workflow means that I know when the first problems crop up, so I simply picked a service that finished before that time, an InterCity from Riesa to Dresden. Not to toot my own horn too much, but I had a lot of fun. Coming into Dresden, you follow the IC from Berlin, which follows the RE50. This means lots of opportunities to get PZB'd if you drive "Hebel to the Table" as we say in Germany. I tried to go slow and steady instead, and I got lots of yellows that changed to green right before I passed them.


I was going to say Dresden Hbf was pretty busy when I arrived, but that's not actually true. It always looks like that in this timetable. So much for Rush Hour, hm?

 

Sieht sehr sehr schön aus, Looks very very nice

 

Oof, DTG set a trap for me and I stepped right in it: The track markers for Weißig "Track 1" and "Platform 1" are on opposite tracks to one another. I resimulated the timetable THREE times today until I figured out why one of my trains started insisting on going all the way from Radebeul to Nünchritz on the wrong track.

 

Lol Sorry for laughing but I'm glad I'm not the only one experiencing this kind of timeconsumeing bugs

 

All good, I totally get it.



Anyway, it's going great today

 

more reasons to ride BR 101 and Vectron and Taurus as passenger... yes from me

 

Nah mate, that spike is from a deadlock I managed to cause at Großenhain. Not the fun type of red signal. Easy to fix though.

Anyway, after a lot of work today I now have... 59 red signal complaints.

You might notice that's more than I had yesterday.

Doesn't mean I didn't achieve anything - it's just that Coswig ist really difficult to get working properly. The thing is that trains can sometimes avoid conflicts that they should actually have with another train by being late. If you're unlucky, the new problem you expose by fixing the old one is a lot bigger than the original one. Thus the numbers don't really reflect it, but I am in fact quite a lot closer to having Coswig work than I was before.

 

Progress update: making Coswig work is still a nightmare, so I'm trying a new idea I had in the shower after not working on the timetable for a few days.

The worst problems have been RE50 and S1 rush hour services disagreeing with another, so I'm now putting them on the same layer. That'll a) help me figure out what I'm doing, and b) let me use dependencies if necessary.



This is fine.

 

when you are done with this you can send your application straight to DB or OBB

 

I'm just going straight to work at the infrastructure division and tell them how to rebuild Coswig so train doesn't cross every other's path.

 

or persuade them to go berserk like OBB does with some originally small regional stations like Wampersdorf (on Pottendorfer Linie, looks like it will have relatively massive sidings/yard)... like, make Coswig great again and again until DTG cant create it anymore lol

 

Fwiw: you can use dependencies between layers. Just drag the dependency arrow to the layer tab of the depending service and you'll be able to connect them. You won't see a line between them, but their link points are filled en you can disconnect them by rightclicking. I have no idea what happens when one of the layers isn't loaded, but this is just a pointer in case you need an occasional dependency.

I do think putting them on one layer will make your work far easier though, I finally decided to go the same route with most of the MSB freight services.

 

The issue isn't not enough tracks, it's that all the paths cross over each other...

Oh, I didn't know that. Thank you! But yeah, I think having it on one layer is going to be easier, specifically because it will be easier for me to tell in which order the trains are supposed to go, which is genuinely one of the things I messed up on my last go :')

 

COSWIG NOW WORKS!

It's taken three attempts and several actual months to get to this point, so I'd like to give a quick explanation of what I've done and why it did or didn't work.

In my first attempt at making this timetable, I went at it layer by layer: I made the S1, then I made the S2, then I made the RE50, etc.. This just flat out doesn't work on DRA. You can't make an exact 1:1 copy of the real world timetable for various reasons, and any decisions you make on how/where to deviate from them are final when you work like this. You make a choice on how to do the S1, you implement it that way, and if you then find out later on that you should have done it a different way, not only are you going to have to edit a large number of services, but also you might break all the other layers you've built on top of it.

Having learned that the hard way, I went for a completely different approach: I simulated and manually test drove one of each train on an otherwise empty route, to learn their timings, then assembled a two hour "slice" of complete timetable and started debugging that. That worked pretty well, but I just couldn't make Coswig work properly. After taking a step back, I found the problem: Going through purely chronologically like I did makes a timing issue on the RB45, an AI-only train that's essentially set dressing in Riesa, look as important as one on the RE50, which the player drives for nearly an hour and that interacts with almost everything. So for attempt number three, I took what worked from each of the two previous approaches.

I'm doing approach number two for the important trains. Those are: S1, S2, RE15, RE18, RB31, and RE50. All the regional services that the player drives for longer than just between Dresden Hbf and Dresden-Neustadt. Those are the ones that the timetable just has to get right, and they're the ones that go through the difficult Coswig node. So, reusing the timings from attempt number two, I made a two-hour timetable slice with just these, and set to debugging. After just a day and a half, I've got them working! Now that I have this solid foundation, I can switch to approach number one (still in my two-hour timetable slice though) and build the other layers on top of it: first the long-distance trains, then the diesel regionals at Dresden, then the AI only stuff. That's in order of a) how complex they are, and b) how important it is for their timings to be correct. This prioritisation was exactly what was missing from my first and second attempts. I feel like I'm onto a winner with this one!

 

oh helm yeah

 

I've got good news and bad news.

Good News: I've added the long distance trains to the timetable and after a few tweaks, they're working really well.

Bad News: The Winter Core Update broke the door controls on the Vectron. I had to (temporarily, I hope!) replace the Vectron with the 101 as otherwise the EuroCity trains would get stuck while the AI tried to open the doors forever.

I'll submit a bug report and we'll have to see if that results in anything.

 

big oof, good luck

 

god damn it DTG, right when I made a Vectron Intercity formation!

 

I am very happy to announce that:


Desiros are in the timetable.


AI only trains are in the timetable.



My two-hour test timetable is complete! The AI can complete every service on time. Now I'll need to do another round of test drives to confirm that a human player following signal speed limits with PZB on can do it too.

 

Kriegt denn der 642 auch noch eine andere passendere Lackierung? Will the 642 also get another more suitable paint job?

 

English!!!

 

Ist doch in Englisch übersetzt! It's translated into English, isn't it?

 

Außerdem kann ich nun mal auch so gut wie kein Englisch, und muss es im Google Übersetzer übersetzen lassen! Besides, I can hardly speak English, and have to have it translated in Google Translate!

 

Good thing want while have.

 

Not from me, but if someone else makes one I'll happily configure substitution settings so they can be used.

 

Okay

 

Looks great! Sadly a bunch of destinations aren’t in the station directory for DRA so you can’t add them onto the trains’ PIS, but do you think you could just add some random destinations that are in the SD onto those services so they don’t have blank destination screens or say ‘Nicht Einsteigen‘? Obvs it wouldn’t be realistic but it would be less immersion breaking than seeing nothing on the PIS screens at all

 

Didn‘t emil#3734 already make one for them

 

He told me, that he added an according geographical region for my repaint.

 

Ah ok. Nice

 

I've been thinking about this for a while, and I'm genuinely not quite sure if I prefer it to an empty PIS or not. I'll continue seeing how I feel while I playtest the timetable.

Haven't actually done it yet as I'm just using Desiro formations I've copied from Maintalbahn right now, but I will do once I get to configuring them properly.

 

Ok, thanks

 

Testing, one, two...



The 10:14 RE50 from Dresden to Leipzig speeding towards a yellow Ks 2 signal, following an S-Bahn into Coswig. The westbound RE50 is an absolute nonstop PZB extravaganza. I had to hit PZB Wachsam eight times from Dresden Hbf to Coswig. You still get there on time though, at least after some small schedule tweaks I made during testing today.

 

Personally I would prefer not to see "Nicht Einsteigen" at all, but am less fussed about what replaces it.