Just bought my first payware addon

[LocoPower]

Hi guys, I lurk here quite a bit but rarely have anything to post about. However I do want to mention that after occasionally playing with RW over the past year, I finally pulled the plug on a payware addon. I bought the Class 101 DMU pack, and found it to be a much higher quality than the default stock, nicely modelled, and the physics feel "right" (obviously I don't know what "right" actually is, but some of the default stock feels like it goes and stops like a car).

So why did I buy an addon?

Because I finally had a route that I knew would give me some playability just tinkering about. Although I downloaded the Port Road when it first became available through Steam, I didn't touch it until recently. What a marvellous route. Lovely detail, a nice combination of double track and single line running, from decent sized stations to single track halts. Most of all, for the first time, I've been able to create some rudimentary scenarios with a bit of traffic, courtesy of good signalling and portals everywhere. As I was playing around with it, I was thinking to myself, "I need a vintage DMU in BR Green to run around this 1960s route". So I hummed and harred over the cost of the Class 101, but decided to go for it. I'm glad I did - I'll be able to use it on Port Road, Newcastle - York, and the FGW route. My only real criticism of it is that the cabview excludes all the dials and controls! You only need to angle down slightly to get most of them in view, but RW doesn't remember your viewpoint - as soon as you use another view then jump back to the cab, you're looking out at the default angle again, not where you last left it.

So although the Port Road was free on Steam (and I gather has been available here for some time), it's netted an extra £15 of sales for RW that I would never have previously dreamed of spending.

But what I really want is proper AI, dispatching, routing logic, and signalling. To me this is even more important than environmental upgrades such as superelevation or graphical enhancements or bizarre point-scoring systems. I don't want to drive around effectively on my own with moveable bits of train-shaped scenery going past on completely separate tracks, I want to feel part of a railway network where I interact with other trains, where what I do affects them, and what they do affects me.

As it is I'll carry on enjoying my Class 101s and finding good uses for them in south west Scotland. But give me the AI and signal/dispatching logic, and I'll gladly purchase more route and stock addons if I can do the kind of stuff I want...

[pjt1974]

Hi
The 101 DMU is a fine piece of kit.
There are sceanrio's out there where you interact with the AI.
Have you tried Hertsbob Leafy Suburb route from UKTrainsim library? Especially, ironically enough, a scenario he made for the DMU 101.
Others who drive the scenario's might be better to tell you more as I'm more of a route builder than a driver.

hope you find more of what you want
Phil

[StPancras]

But what I really want is proper AI, dispatching, routing logic, and signalling. To me this is even more important than environmental upgrades such as superelevation or graphical enhancements or bizarre point-scoring systems. I don't want to drive around effectively on my own with moveable bits of train-shaped scenery going past on completely separate tracks, I want to feel part of a railway network where I interact with other trains, where what I do affects them, and what they do affects me.

I'm a bit of a lurker myself. I have to agree with you, I am rather bewildered by AI? trains waiting for green signals, and then steaming through red ones. From what I have read here, Rail Sim don't seem that interested. However, I think most of the competition disagree, and RS will have to do something with it eventually, just to keep up.

[paulz6]

But what I really want is proper AI, dispatching, routing logic, and signalling. To me this is even more important than environmental upgrades such as superelevation or graphical enhancements or bizarre point-scoring systems. I don't want to drive around effectively on my own with moveable bits of train-shaped scenery going past on completely separate tracks, I want to feel part of a railway network where I interact with other trains, where what I do affects them, and what they do affects me.

I'm a bit of a lurker myself. I have to agree with you, I am rather bewildered by AI? trains waiting for green signals, and then steaming through red ones. From what I have read here, Rail Sim don't seem that interested. However, I think most of the competition disagree, and RS will have to do something with it eventually, just to keep up.

The AI decides its own view of signal block states and doesn't read the signal scripted block states. That basically means a signal scriptwriter can create signals that conform to operational logic. However, if the AI doesn't agree it will do things like plough through red signals and wait at green ones. It's a disconnect between the customisable script logic and the non-customisable AI logic.

It's probably not the case that RS.com are not interested in resolving these things, I would think it is the case that they have financial business constraints.

Their business view is that the RS career system will bring the dollars in, rather than fixing things like stutter and inadequate AI operations. Whilst I found the career scenarios a little bit of fun to start with, I really couldn't care that much about them now. To get high points you have to drive irresponsibly and someone always seems to find a way to get 400 points more than everyone else.

[bigvern]

I'm pretty certain that RSAdam posted a month or two ago, that the AI/signalling we have now is pretty much as good as it will get. Whether that's because there's no money in the budget or technically too difficult to undo what they inherited from Kuju isn't clear. So I'm afraid, for all it's good points, we are left with a "simulator" that is becoming more of a game where the AI trains have to be regarded as carefully placed moving scenery rather than integral to a network operation as found in (say) Zusi.

No doubt I will get flamed/censored for writing the above but many of these items have been on the shopping/wish list since Day 1 of Rail Simulator but here we are nearly three years in and the longer it goes, the less likelihood it seems of any major fixes.

[phat2003uk]

The AI/signalling just isn't that bad once you get to know its quirks but I suspect a lot of people just down tools before getting to know it well which is fair enough!

[bigvern]

The AI/signalling just isn't that bad once you get to know its quirks but I suspect a lot of people just down tools before getting to know it well which is fair enough!

I don't know how many you will find actually agree with that, Richard.

The fact that even at the basic scenario building stage you can't get two trains to run in opposite directions on a single line without having to fine tune the timings artificially, or AI trains seizing a path long before they get there indicates the thing is basically banjaxed.

We know from Adam's earlier statement re AI that such features as dynamic routing, using alternate lines if the one specified in the path is blocked etc. is never going to happen. I mean it's so basic we can't even save a string of driver commands if we decide to change the traction on a train in the scenario editor (unless using the officially unapproved RW Tools).

Like many others I will be buying the WCML payware on Friday, but largely as a thank you to Keith for building it, not in any sense accepting RS.com's take is going to make any difference to RW's ongoing evolution as a scenery simulator with trains travelling through it (if the scenario actually loads of course!).

[phat2003uk]

To be honest, I've not had much experience with single line operation but like you've said, I suspect it requires artificially timings the trains around the signalling which is a shame.

My experience with standard 2/4 track mainlines hasn't been too bad at all though in recent months so to sum up really, I think it's fine for most situations.

I'm not sure I quite buy how much dynamic routing would really add to the game. For example, if the usual path was to somehow be blocked, surely this would be known by the scenario creator in advance so he just can route the player train around the blockage or are you talking about things such as spontaneous failures which are not planned in advance?

I agree in that we should at least be able to change a locomotive which has instructions above it with ease.

Like you, I'll be purchasing the WCML route on Friday and I'm looking forward to seeing how the signalling copes.

[paulz6]

Sadly Vern, I think you are right about any expected progress on AI/signalling. The recent assertion on the WCML thread that it was either one or the other when in comes to default/LED signalling compatibility leads me to believe that valuable coding time is severely restricted. Both signals types behave logically in exactly the same way, and I didn't think the default/GARL signal scripting was significantly different to be honest. If they can't stand the cost of standardising these two scripts, more complex programming requirements is never going to get signed off unless it adds more (potentially buggy) functionality that can be included in the sales brochure.
I'm pretty certain I can write signal scripts that cater for bi-directional line working, but the AI as it's own idea's, so it won't solve that problem I'm afraid.
If the AI is more dynamic, there would be more onus on the scenario creator to avoid traffic jams blocking up the entire route system, so that is something that has to be considered.

[RSderek]

It would be interesting to know exactly what Adam said and in what context.


LocoPower, thank you for letting us know about your thoughts, and also thank you for choosing the 101, I put loads into it as did everyone who worked on it.

best regards

Derek

[Retro]

I feel I must agree with Vern and Paulz6 here having encountered problems on my own Route with the Dispatcher and Signals. Trying to get anything slightly realistic on the Scenario front is a major struggle and involves a lot of extra Markers and even some Track alterations. I am not sure what logic the Dispatcher is using but it is very unpredictable and baulks at the first complicated Traffic movement. This usually causes Signals to do strange things or refuse to react at all. I have also seen AI ignoring the Signal State.
Kind regards James.

[LocoPower]

Ah, glad to see I'm not alone. All the nicely modelled payware rolling stock is lovely, but if I purchased it, what am I supposed to do with it?

That said I hope the vast amount of stock available is selling and putting some coffers into RW's finances, as it looks to me as though this is their way of continuing the cash flow - fair enough, all businesses need to look after their cashflow in order to sustain themselves. It's just that there seems an awful lot of focus and publicity over all the stock available (or becoming available) and precious little mention of improvements to the simulator core. How much more money is out there, willing to be spent on stock that can actually be used as part of a railway?

I saw a quote of Adam's post in a recent topic I lurked in, and tried not to read too much into it. One the one hand, it just seemed to be a simple statement of fact - this is what it does, which is what we intended it to do. However I also inferred from it that it simply hasn't been given any thought either - it could equally be read as - "Hey, you can have other trains moving around the map. What more do you expect?" Taken this way, it is not an encouraging sign.

I'm sure that redoing the AI and signalling to provide a more realistic experience is an expensive (i.e. time consuming) and extremely complex project to tackle, with no immediate return on investment, unlike the release of more rolling stock. But I would've thought that the RW devs need to take the longer term view, for the future of the product.

Here's hoping, anyway.

[RSderek]

hi, some interesting views and some not so.


We have short term, mid term and long term plans.

Sometimes, depending on circumstances they trade places making us very busy chaps.
Either way, we are only just at the start of this exciting adventure.

As for routes, signals and dispatcher just make sure your routes have:
A - track set up correctly .
B - Signals are set up correctly (using the correct signal in the correct place, even old hats get it wrong)
C - your scenarios and trains are set up correctly.

best regards

Derek

[nsupersonic]

The AI/signalling just isn't that bad once you get to know its quirks but I suspect a lot of people just down tools before getting to know it well which is fair enough!

Quite agree that the AI signalling is not that bad, once you learn it's habits. An example being: As a rule of thumb I always make sure that the marker arrows (spanning from the signal at world editor screen); are well clear of rendered points in a junction. If unsure check one of the supplied routes(without undoing the 'padlock' protection in the world editor) for an example.

regards
Charles

p.s the rendered start and finish of a junction are clealy visible in Railworks.

[USRailFan]

But then you get the problem sometimes, that changes are done to how signalling and pathing works, which breaks previously functioning scenarios... (e.g. the latest 'Career Mode' updates, which broke several scenarios from IHH and DT add-ons, and even at least one default scenario on Hagen - Siegen).