Yet another frustrating, scenario-ending AI crash

[OldProf]

It's been a while since I tried a scenario on the Newcastle to York route, so I decided to give "Cathedral Special" a whirl. This scenario begins with some shunting to get a static locomotive out of the way of the driver's loco. Then, the player begins a long dash to pick up passengers before a fixed arrival time. At first, the task seems impossible, but as the scenario progresses, the loco slowly but surely makes up time and at about five miles from the arrival platform ...



Well over an hour of driving ... ... wasted ... ... thanks to some RSC scenario writer's careless attitude toward AI traffic. This scenario was created, unless I'm mistaken, several versions of RW ago and it is possible that "improvements" resulted in this crash. Even if this is true, however, once again no one cared enough about players to test-run the scenarios that could be affected by changes in a route.

And RSC is certainly not alone in this: I've encountered numerous pay-ware scenarios that fail because whoever wrote them did not bother to test them thoroughly before putting them on the market.

[NDORFN]

If you want a job done properly...

[simer4]

I just had one of these. On the WCML North in the class 325 scenario, "Have You Been Good?" It angered me.

[OldProf]

Instead of flashing that teeny-tiny message and snatching it away just as I've found my magnifying glass and am ready to write down the coordinates, it would be great to take me to that location and ask whether I want to edit the scenario.

Even when I do have time to note the coordinates, "going" to them in the editor doesn't do much good since the scenario's time has been reset to its beginning and the cause of the crash or derailment or unscheduled lunch break or whatever doesn't exist yet, so playing detective is practically impossible.

[SMMDigital]

I just had the same thing happen earlier today in a Scenario i'm writing. It was a simple problem, and I wonder how many other AI collides happen in the same way. I had put together two consists that were moving from the other end of the route towards my train, and the second one ran into the rear of the first one, even though they were spaced two-blocks and fifteen minutes apart. The problem turned out that the last six cars of the first AI train were not connected when I set up the consist, so they were being left behind for the second AI to plow into fifteen minutes after the first train departed. It was a careless mistake, but a learning experience nonetheless.

BTW, I thought the AI recognized static consists, such as cars left behind by careless Scenario writers?

[robbit]

Instead of flashing that teeny-tiny message and snatching it away just as I've found my magnifying glass and am ready to write down the coordinates, it would be great to take me to that location and ask whether I want to edit the scenario.

Even when I do have time to note the coordinates, "going" to them in the editor doesn't do much good since the scenario's time has been reset to its beginning and the cause of the crash or derailment or unscheduled lunch break or whatever doesn't exist yet, so playing detective is practically impossible.

when you get that tiny screen with the info on it just hit f12 and grab a screenshot of the message. it might even give you the consists involved. That is what I do anyhow

[Machinist]

Even when I do have time to note the coordinates, "going" to them in the editor doesn't do much good since the scenario's time has been reset to its beginning and the cause of the crash or derailment or unscheduled lunch break or whatever doesn't exist yet, so playing detective is practically impossible.

After you got the screenshot with F12 as robbit mentioned (or F10 using FRAPS), go to the coordinates (in Scenario Editor), and simulate playing the scenario using the 32X speed up value. If Player train is not interacting directly with AI Trains, is fully possible to know what is going wrong. Even if interacting is usually possible as well...

[OldProf]

I just had the same thing happen earlier today in a Scenario i'm writing. It was a simple problem, and I wonder how many other AI collides happen in the same way. I had put together two consists that were moving from the other end of the route towards my train, and the second one ran into the rear of the first one, even though they were spaced two-blocks and fifteen minutes apart. The problem turned out that the last six cars of the first AI train were not connected when I set up the consist, so they were being left behind for the second AI to plow into fifteen minutes after the first train departed. It was a careless mistake, but a learning experience nonetheless.

BTW, I thought the AI recognized static consists, such as cars left behind by careless Scenario writers?

That certainly sounds familiar! After similar experiences too embarrassingly numerous to mention, I've learned to always check every AI train for consist-ency before testing the scenario in question. It's so easy to do, but even easier to forget. As to consist blindness, my guess (more of a SWAG, actually) I think it exists for all consists not mentioned in the scenario's instructions. Thus, in the situation you described, the second AI train would "see" the part of the first AI train included in its instructions, but not the unintentionally detached part. Another kind of consist blindness may be what keeps any cars not coupled to the player engine from appearing in the F4 HUD's ribbon?

Perhaps someday this kind of information will appear in an updated version of the Timetable Editor Manual, or at least in the by now per-historic WIKI.

[OldProf]

Thanks to Robbit and Machinist for the tips!