[9.8 bug] Inconsistent journey time calculations

[Carl]

Opening an existing map in 9.8, I notice there are problems with journey time calculations in this new version. It's not a uniform problem -- some journey times remain the same. But many others have become much shorter than they should be given the distances and average speeds involved. This seems to occur exclusively on long-distance lines, though it may be that the problem is just more noticeable there.

I've checked that this isn't related to changes in average speed: the occurs for several lines whose average speeds remains largely unchanged.

I'm also certain that error has been introduced here rather than removed: in one case, there's a train which can apparently travel 370km in 163 mins despite its average speed being only 90kph. That's impossible -- especially given that the 370km value appears to be 'as the crow flies' rather than the actual length of the route!

Would you like to see a save-game to illustrate this?



I also found that revenues leapt by a factor of ten. I'm not sure if that's related -- or whether it's by design -- but I thought it also worth reporting.

[jamespetts]

Carl,

yes, indeed, I shall need a saved game to debug this. Thank you for the report.

[Carl]

Here's a savegame and some pointers on where to look to see the best examples of this bug in action. Thanks for looking into this!

Save-game: http://simutrans-germany.com/files/upload/UK-9,8bugs.sve
Pak folder: http://simutrans-germany.com/files/upload/pakfor9,8.rar


First, here are two-distance routes which are subject to the bug:

York to Plymouth (430km as the crow flies)
Line average speed: 90kph
Journey time in 9.3*: 332 mins
Journey time in 9.8*: 197 mins

Bournemouth to Manchester (274km)
Speed: 88kph
9.3: 253 mins
9.8: 142 mins

Many other lines exhibit this behaviour -- but it's easiest to see on these long routes. (See also: Paddington to Penzance, etc.)

However, not all routes show a significant change in journey time. Here are two long-distance routes which are not subject to the bug:

Norwich to Liverpool (316km)
Speed: 75kph
9.3: 316 mins
9.8: 306 mins

Milford Haven to Manchester (280km)
Speed: 70kph
9.3: 330 mins
9.8: 327 mins

I realise that, since the distances are point-to-point, journey times will not directly correlate with them. I've included them to highlight the absurdity of some of the new journey times.

* These times are from opening the very same file in both 9.3 (which I currently use) and 9.8.

[jamespetts]

Thank you for uploading that. With some help from Knightly, I have fixed this issue on my 9.x branch, which fix should be implemented in the next release. Thank you for your report!

[sdog]

Do crows fly straight?

If yes, how do we know so? (i've seen the phrase in Tolkien's Hobbit, a little before puting gps+logger backpacks on birds.) Sorry for being almost off topic.

[Carl]

That's great -- thanks jamespetts!

I'm afraid my knowledge of crows is next to nothing. My ability to parrot conventionally used phrases, on the other hand, is second to none

[jamespetts]

What is the flight speed of an unladen swallow?

[Dwachs]

African or European?

[jamespetts]

Now fixed in the released 9.9.