Hypothesis: Reliability depends on voodoo, phase of the moon, and age.

[Gumboots]

NVM. I see what you're saying. You were assuming the default B era freight consist, not the proposed new B era. I think 8x default B freight would be way overloading a Firefly.

If you want to include a caboose in the calculations add an extra 3 tons. Default A era is 5 freight, 7 caboose. Default B is 10 freight, 13 caboose.

[RulerofRails]

Yep, you figured it out. Sorry to make you puzzled. That comparison was default B-era, and yes I forgot to add 3 extra tons for the Caboose. The point of that was mainly to get an overview of a situation that can be easily agreed on as off limits for acceptable range of reliability.

Maybe it could be regarded as adding spice/variation to gameplay, but balancing an engine such that it has a 3 year lifespan isn't ideal IMO. I think of it this way, apart from anyone who read this thread carefully, I doubt anyone else replaces their engines that often. Anyone else who does for a different reason or always used this strategy, please speak up.


A better explanation for my suggested limits of 40% breakdown chance for the early era, and 30% in the 20th century is to consider that train speeds are faster in the 20th century, and since breakdown chance is per 1,000 miles, a faster train incurs a greater breakdown chance within any specific period of time (for example in a year). This is painfully evident with the "futuristic" E-88. The only map where I used them extensively was the Dutch-Atlantis map. Way back in the day I sort of got a premonition of the 3 year grace period, because I tried to run them without a Caboose and they would do ok just for a little while and then break. For a mixed default consist (294 tons, in the spirit of extracting economic potential engines, I know that even if such a train existed it would be a pax only affair) the sheet tells me that they should last 6 years before reaching 30% breakdown chance. At the time I was puzzling why they were dropping like flies after just a few years. Now I know better, but I was encouraged in my behavior because I saw quite clearly that running with an 8th paying car was that much better for my bottom line (IIRC, worth 100+k extra per year in profit from the train).

Adding a caboose is an economic penalty on a train (maximum penalty is capped at 7/8 less revenue). A higher reliability chance is an economic penalty on a train. The effects of both penalties on a player are hard to define in an actual game. If all other things are equal, surely the player will chose the engine with better reliability. In the example of the Firefly from my previous post, one can see that the 3 year replacement cycle will put costs for the Firefly 25% above the Optimum. (Not that it's representative of what we will do, but reliability is costing dearly in this case.)

Some analogy could be made in how much advantage maximum speed when loaded, hill climbing ability, and acceleration give a train. Once again hard to define and arguably more so than reliability depending on usage. Assuming all other things again being equal, surely the player will choose the engine with better performance.

When balancing fuel cost and by extension Average Yearly Cost some allowance could be made for these penalties/advantages. However, whenever a player sense that he faces a dis-advantage there are steps that he may take to limit his downside (depending on how much weight he gives it compared to ever-reaching goal of company growth). To reduce the penalty on reliability, there's over-servicing, shortening consist length, etc.. A deficiency in performance can be partially made up with better track laying practices for lower grades especially in situations like bridge ramps, routing for less congestion when acceleration is poorer, using double track in preference to single etc.. Generally such measures pay off better in situations of higher cargo flow.

I know there are different classes and sub-classes of loco so our selection at a certain point in time may be limited enough that it makes sense to have significantly different optimum Average Yearly Cost between locos. But with the 1.06 locos and at least in the time period of 1920-1950 there will be enough selection to mean that selection between locos will come down to economics.

---

I took a look at the default locos in D-era that use a reliability ratings of #6 and #7 after 1970 or so (still available not the ones just introduced after 1970). And what do you know, bar the Class QJ which remains in service until 2000, those engines are the engines which were used mainly for express in real life.

Makes sense , although I don't remember a game set in North America in the late 20th century with enough express volume to actually make dedicated express engines with the diesels like the FP45. In Europe, the Electrics are generally a much better proposition to give high speed express service. Still not sure there's a scenario I used them as express-only purely for economic gain either.

[Gumboots]

Maybe it could be regarded as adding spice/variation to gameplay, but balancing an engine such that it has a 3 year lifespan isn't ideal IMO. I think of it this way, apart from anyone who read this thread carefully, I doubt anyone else replaces their engines that often.

I get the impression that most people just grumble about them breaking down. I'm unsure of how many booby traps we should build in.

[RulerofRails]

What about keeping things simple with the design, but planning so that an event which decreases reliability by one level will make more careful servicing and train management a mid-level necessity? A note in a read-me could suggest an event to do this if such challenge was wanted by the map maker.

A calculation of the ratio of difference between one reliability and the next.

Now for a doodle to illustrate the maximum consist weights that the reliability system will allow.

Reliability Doodle, 30% chance tonnage.jpg (20.96 KiB) Viewed 4437 times

Not suggesting that we change the proposed weights, just thought to put some of this math to work (for fun ).

Observations: Using the extreme end of the scale would mean that everything designed for that time period is restricted to one or two levels. In terms of the fuel scale, the 1900-2000 ton mark would be manageable with a rating of #1 (Outstanding). But, same problem: would end up with no variety in loco stats in the end game. Also, locos post-WWII need to overlap weight changes more often. IMO, should be viable for one full 25 year cycle. For example the Class QJ is available for 44 years (1956-2000). Currently, the first loco that is available into the future (end date: 9999) is the EF-66 in 1966.

[Gumboots]

Although it hadn't occurred to me before, scripting a scenario to tie reliability and other stats to game difficulty level has potential, if any scenario author is feeling evil. Perhaps the easiest way of doing it would just to be increase oil consumption. That would effectively allow fine tuning of reliability. The ratings in the pop-up would stay the same, and you wouldn't necessarily have to warn the player.

Anyway, evil trickery aside...

I have thought about the weight range, but I got the impression you weren't keen on pushing things higher than what I've done (which is currently 480 tons for 8 of the heaviest freight cars in the last era). That would allow tolerable lifespan and usage for Very Good without a caboose, or quite satisfactory lifespan and usage for Good with a caboose. Which is about where I thought it would be.

It could be bumped a bit, if we want it bumped. But then we have to consider the balance with fuel consumption/running costs, etc. The short version is that if you think there is benefit in altering the absolute values of the weights, while keeping a similar progression between eras, I could probably be talked into that. It only requires altering four bytes in each .car file, which isn't that big a deal.

The other possibility is adding a ninth era for the 21st century, and bump heavy freights up to around 80 tons for that.

[Just Crazy Jim]

I am all for a 9th era - but honestly, that's were we need to hire in an artist of the Syd Mead school

[RulerofRails]

Although it hadn't occurred to me before, scripting a scenario to tie reliability and other stats to game difficulty level has potential, if any scenario author is feeling evil. Perhaps the easiest way of doing it would just to be increase oil consumption. That would effectively allow fine tuning of reliability. The ratings in the pop-up would stay the same, and you wouldn't necessarily have to warn the player.

That's a good idea. Do this at the start of the game.

What I don't recommend is increasing Oil/Water consumption without warning mid-game. One must plan route distance between servicing stops when setting up a route initially. Maybe less true with spur service, but even then it would be frustrating to go through the routing for all your current locos and re-do service frequency.

Something I need to remember, in the reliability scale smaller is better in the lco file, but in-game when adjusted via event it's setup so that + a level increases reliability.

---

I hope you didn't think this was a suggestion to tweak with the weight scale. Just further exploration of the effect of consist weight on reliability. Maybe it gives more usable info, but I wasn't recommending to increase the final weight. IMO, it's the progression that matters, not the actual weights.

I have thought about the weight range, but I got the impression you weren't keen on pushing things higher than what I've done (which is currently 480 tons for 8 of the heaviest freight cars in the last era).

Yes, it's true. And the main reason is that the fuel scale is much more adjustable on the lower end. Basically a linear scale for fuel ends up with the difference between levels as a power curve, whereas, a power function as the scale for reliability actually ends up with the difference between levels being much more even as a progression. Illustration:

My thought previously was that lower weights might enable reclaiming a poorer reliability level or two for practical use in the early game. Now that I realized that any engine (regardless of actual weight) is counting as roughly 150 tons by the formula, the steam escaped from this idea. Smaller weights in the initial time-frame will increase reliability, but it isn't worth that much because it is always being compared against the 150 tons of the loco. For example, with a consist weight of 10 tons, the consist portion only makes up 10/(150+10) or 1/16th of effective breakdown chance. For a consist weight of 5 tons, we get the math 5/(150+5) or 1/31st of effective breakdown chance. Not what I expected.

But, IMO the weights we settled on are still the way to go for the fuel scale especially to fine-tune the fuel bill for express. From our test runs at adjusting physical engine performance it seems like it's workable on that front as well.

In regards to a 9th era, I have no objection per se, but I don't see lots of realistic scope to bring in new models. From my limited perspective, IRL the car types in use for freight seem pretty static lately.

[Gumboots]

Yeah I had thought much the same regarding a 9th era. It's something that would be fairly easy to add later if there seems to be a good use for it, but I'd prefer to stick with 8 for now.