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

[Hawk]

Incidentally that brings up another point. Testing this stuff with diesels and electrics would be a waste of time from my perspective, and there's already enough to do. Anyone who wants a better version for diesels and electrics would have to be prepared to do their own testing.

I'm glad you mentioned that. Now I know I can safely ignore all these discussions. Thanks!

[Gumboots]

Well some other people may be prepared to test with diesels. I could handle doing a little bit of diesel testing. RoR said he could do some.

[RulerofRails]

Hawk, my view on Diesels and Electrics is that they are pretty decent to use the way they are. The majority have better fuel economy and reliability settings and less weight than steamers. These are the main performance factors that are crippling some steamers making them almost unusable.

---

Simple calculation for Breakdown Chance.
Each maintenance level has a base value. Gumboots made a nice chart for most values here. We can make a simple equation where any one of these values is x.

The initial massive jump in breakdown chance in the first two years is worth roughly 3.6. Or 3.6x. (This increases roughly 0.1 per year thereafter.)

8 default freight cars (D-era, 1950-onwards) having an approx. tripling effect so we get 3(3.6x)=10.8x.

A train will likely spend most of it's life with less than 75% Oil. At that Oil level the multiplier is 1.25, giving 1.25(10.8x)=13.5x

So, during its third running year: this train's breakdown chance =13.5x


Comparing a Above Average setting: 13.5(3.0%)=40.5%

To a Good one: 13.5(1.9%)=25.7%

So a Good setting is 57% of an Above Average one in this case. This isn't pretty.

---

Fuel Costs.
Diesels and especially Electrics benefit greatly from better fuel cost ratings. All the fuel levels are numbered from 1 to 10, but the effects of a better level are heavily skewed towards the good (smaller #s) side. Where z is the current level, the saving from moving to a better level can be represented as 1/z.

Extremely Poor (#9) to Very Poor: 1/9= 11%

Average (#5) to Above Average: 1/5= 20%

Good (#3) to Very Good: 1/3= 33%

To add to all this, Diesels and Electrics not only don't have a consist weight, but are on average a fair bit lighter than the majority of steamers. The QJ and Kriegslok are exceptions and this is why you see players using them even though they are snails. I haven't done an accurate calculation, but would hazard a conservative 30%, but in some cases the figure is much higher and likely the figure is closer to 50%.

8-car default freight consists after 1950 weigh 320 tons. If engine and tender weight of a steamer averages 80% of this, or around 260 tons, the effect would be (30%*80%)=24% saving on fuel costs due to lighter engine weight. This saving will actually be greater between 1900 and 1950 because consist weight is a smaller proportion of train weight, but it doesn't matter as much since total train weight is lower so the saving is applied to a smaller starting sum.

Combining better fuel settings with lighter weight it's not unusual to see fuel costs of 50% less than some average Steamers.

---

The effect of the 1950 car weight increase is designed in a way to starve out most steam, mainly since the Average fuel cost setting which is common for Steamers in 1.05 becomes a bleeding river with the heavier cars. The only light steamers just happen to be the only efficient steam options after 1950. This was done by design and indeed Lirio's settings have fixed some of this.

My view is that I want a train that I can make a decent profit with, meaning low fuel costs, reliability is more just a headache that I will try to manage as best I can. Only in rare cases will I consider buying a loco (after 1900) with a worse fuel rating on the grounds of reliability.


I am looking forward to making some of the Diesels into true freight haulers. That's definitely worth doing. I need to wait until the steamers are sorted though, to make sure that any changes that need to be made on a larger scale to keep overall balance can be incorporated before nutting out the finer details. I am quite confident that Diesels and Electrics will react to engine setting changes the same way Steamers do. "Shouldn't be that hard." - Famous last words. . . . .

[Gumboots]

Just ran a check on the custom freight car weights. On average they work out to 41 tons each, but some are the standard 40 tons and others are up to 50 tons (corn, grain, coal and iron). Given that I always run a caboose on Kriegsloks in the D era and that currently weighs 53 tons, an average freight consist would be 341 tons and a maximum freight consist could be up to 393 tons. That's basically halfway between default D era 8 car consist and the proposed D+50% maximum, which is encouraging since it means I must have frequently run consists up to that weight without noticing significant problems.

The 341 ton average would increase breakdown chance by 4.4% of whatever it was before for a 320 ton consist and an Above Average rating, and the maximum 393 tons consist would increase breakdown chance by 15.5% of whatever it was before. Given that with a caboose the breakdown bar is usually around 25%, depending on age and oil level, the 341 ton consist would up that to 26.1% and the 393 ton consist would up it to 28.9%. So not huge changes, but noticeable.

For fuel usage, a Kriegslok and its tender weigh 115 tons. Add a default 320 ton consist and you have 435 tons. A 341 ton consist will add 4.8% to that, and therefore to fuel consumption. A 393 ton consist would add 16.8% to fuel consumption. So in this particular case the increase in fuel consumption is fairly close to the increase in breakdown chance.

I haven't checked all the diesels, but I did check the Deltic last night since it's a pretty average diesel of the period. It weighs 14 tons less than the Kriegslok/tender combo and is one notch up the fuel economy scale, so with the same maximum 393 ton consist it'd only have 73% of the fuel costs. It's also rated one notch higher for reliability, at Good rather than Above Average, so would have about 5/8 the breakdown chance for a given consist. Say around 15.8% instead of 25% for the same consist if both are running a caboose. Without a caboose the Deltic would therefore be around 32%, which is still usable.

So making late era steamers more competitive with diesels is easy. Just give them a Good rating for reliability (a lot of steamers really did have good reliability, and some really were outstanding) and a Good rating for fuel. Alternatively, given that tenders weren't always full, leave the fuel rating alone and just rate tenders at half their loaded weight. Diesels have to carry fuel somewhere, but aren't charged any weight for it in RT3.

[Gumboots]

The game has no good use for the Poor, Extremely Poor, and Atrocious levels with only 5 locos using them. The base levels here are quite high. Going along with the theme we would need to have 2.5 ton freight cars to have a decent Poor level (Era-A passenger cars are already 3 tons and could be kicked on the lowest "1" perch)...

...There is always the possibility to consider of trying to reclaim at least the Poor level and a couple of Fuel levels with 3 ton freight cars and 1 ton Express.

Was just thinking about this. If I'm cutting express weights anyway, why not go for 1 ton A era instead of 2 tons? It does make sense. The loco stats are going to need rebalancing anyway, so makes it no difference what the absolute values are for express weights. We can just halve pulling power and free weight if we want to.

Halving the weight of all consists, and presumably of locos as well, would definitely give us far more usable scope for variations in relevant ratings. As you rightly pointed out, we get a finer gradation at the lower ends of the scale, and we should easily be able to reclaim one or even two levels of fuel economy. We may be able to reclaim one level of reliability too (not so sure about that, given the exponential nature of that scale, but worth looking at).

Hmm. Actually, just off the top of my head, this is sounding good. Default A era freight is 5 tons, while express is 3 tons. If express is cut to 1 ton, then lighter freights could be 3 tons and the heaviest freights could be 5 tons. One ton express should make even the Firefly somewhat tolerable with its current reliability rating, and we already know we can take the fuel scale past 10 if we want to (the visible text staying at Atrocious is fine IMO). The heaviest freights being 5 tons would mean that current ratings for A era freight locos would still work just as they do now.

[RulerofRails]

Sounds good. I mentioned this before, but I tested and confirmed that decimal weights can be used so there's still the possibility for variation on the low end of the scale.

I agree that reliability is the least forgiving. Potentially, we will use more levels of progression and freight could still end up near the default weights by the end of the game. So, in say 1980, freight weights could potential hit the current default 40 tons. If it fits with the reliability settings, the default eras could be split in two. Possibly do a simple 3t-5-8-12-17-23-30-38 progression (ETA: Reliability settings may mean that progression should be more like exponential), which could split the current eras in such a way that half-way through the default "era" weights become comparable to what they presently are. So in 1925 we get freight weights at 23tons. In 1980 we get freight weights of 38 tons.This may help loco stats to still appear reasonably sane compared to actual performance, at least for freight.

Rigging up a progression with the default cars will likely be good for testing, but then we could make a nice smooth curve and adjust WP&Ps cars to fit that curve (depending on type of cargo to be a little higher or lower) according to their intro date. I don't know how this project will end up, but if it's a pack of some sort it could include WP&Ps cars. If so, there's no reason to do all the default ones unless there is a specific request.

[Gumboots]

I think we can get away without using decimals for cargo car weights, but nice to know just in case.

Default 40 tons max in later years might be about right if we're starting by halving things, although I've already confirmed that 60 tons is acceptable even with Above Average ratings, and would be comfortable with Good ratings, so there's plenty of scope at the upper end. Can try a few things out and see how it's looking.

I was thinking of using a maximum of seven eras, effectively dropping an extra one between each of the defaults. This would mainly apply to cargoes that are available from 1830 to infinity. Cargoes that come in later would have fewer eras, but I'd still list them as starting from A if using letters to denote eras.

WP&P did it in a very messy and annoying way IMO, at least if you have to look into the files. Doesn't matter for people who never look at the actual files, but he named his extra eras W, X, Y and Z, so his progression goes A, W, B, X, C, Y, D, Z. It's consistent and comprehensible, but still a PITA because it's not intuitive, so you're always scrambling around a bit. Also, when you open all files in a code editor they get listed alphabetically, which means they are all screwed up for chronological order. Come to think of it, this applies even if you just open the parent folder in a standard window. It really does end up being quite a nuisance.

Also, when he had a cargo that came in later he'd sometimes start the naming from whichever era was generally, roughly, at about that time, instead of always starting from A, which means you're often doing a double take wondering if you've missed some files. Me, I'd always start from A and go alphabetically from there. It means renaming a few more things, but at least the finished product is a no-brainer if anyone has to look at it later (maintainable code rocks).

Or even better, but which I haven't tested yet but which should work, is to drop the letters entirely and use start years: Coal_1830, Coal_1845 etc. The good thing about this is that it'll still be clear when using uneven introduction dates, which WP&P used because he thought (and I agree) that having different cargoes increase in weight at different times could add more depth if done carefully.

So I wouldn't necessarily be going for a totally smooth curve everywhere. A bit lumpy could be good for extra interest, and could tie in with historic changes in consists, which often proceeded in more or less jumps. We'd just have to decide on a reasonable amount of lumpiness to get the best overall value for the game.

WP&P's cars often use default mesh anyway, with the main differences being skins and other files. Then mesh these days is easy to tweak too (I've been cleaning it up while doing the doubles and triples). So really we can pretty much grab anything we want to use, without it adding significantly to the difficulty of putting the whole pack together.

[RulerofRails]

So I wouldn't necessarily be going for a totally smooth curve everywhere. A bit lumpy could be good for extra interest, and could tie in with historic changes in consists, which often proceeded in more or less jumps. We'd just have to decide on a reasonable amount of lumpiness to get the best overall value for the game.

I'm open on this, I would simply ask that any such jumps be simple enough and well-explained so the game can be played strategically without constant reference to a spreadsheet.

Currently your latest version of WP&Ps list has a 50% increase in average freight weights in the years 1900 to 1915 which I believe is sufficient enough to reflect the introduction of the superheater. 14t average in 1900 and 22t average in 1915. In terms of acceptable reliability this means the loss of the Below Average level. For fuel costs it means the loss of the Poor level when using a 65 ton engine.

[Gumboots]

Yeah I agree it has to be fairly straight forward. It still needs knocking into shape some more. I was thinking it'd probably make sense to group cargoes by types (mining stuff, agriculture stuff, military stuff, whatever) and have all cargoes in a category getting their weight boosts at about the same time, and by similar amounts. So that way you'd just have to remember mining stuff gets heavier about 1910 or whatever and you might want to check any locos that are hauling from mines. WP&P wanted to create maximum mayhem with introduction times, but I think it should be toned down a bit.

And I was pretty much thinking that required reliability (for the heaviest freights) would be Above Average post-1900, and Good post-1955. That's pretty realistic IMO. I think reclaiming of lower grades would be mainly for getting greater flexibility for mixed and express. Express, in particular, should be able to run quite low ratings well into the 20th century.

[RulerofRails]

I am having trouble adjusting an express engine to be good on grades while at the same time bad with freight with the 1 and 3 ton era. From my fooling so far, with 0 free weight, at 28tons combined engine and tender weights, the maximum on flat ground for 8x 3t freight is approxiametly equal to the speed of the Express with 8x 1t consist on a 2% grade. To get a big drop off with heavier cars on the flat seems to occur the side-effect of a similarily large drop off on grades with express. Any ideas on engine settings to get a small drop off on the grade scale accompanied by a larger one if more cars are added? I haven't given up on the idea 1t and 3t may be workable yet.

[Gumboots]

Well if we're also using reliability and fuel economy ratings that should automatically penalise express locos that try hauling freight, so it won't just be a matter of pulling power and free weight. Also, 3 tons is the minimum freight weight, included for fast-rotting cargoes. The maximum freight weight is 5 tons in the spreadsheet you have, and I was thinking of increasing it to 6 tons.

I gave it some more thought, and figured it'd make sense to put a 48 ton upper limit on the late 20th century. This basically gives the same relative range in all car weights as WP&P's original sheet, just with lower absolute values that will allow us more flexibility. It will mean Above Average ratings for fuel and reliability will be workable with maximum freight consists at any time, although for reliability you'd really want a caboose, especially if you were relying on random stops at inline facilities. A Very Good reliability rating should keep you out of trouble without needing a caboose, even in the late 20th century, which feels about right.

Loco and tender weights should probably be cut more too. I think as a rough rule of thumb, based on what we've found when using existing stats, no locomotive should weigh more than the consist it's supposed to haul. That should be enough to get something like the Big Boy being usable without the stats looking ridiculous. It'll mean dedicated express locos can have fuel economy set to give reasonable fuel costs with total weights (loco/tender/consist) varying from 16 tons in 1830 to 128 tons in 2000. Kreigsloks with full default D era freight are about 435 tons all-up weight, and are fine with a fuel rating of 4. So late 20th century express with 8 ton cars could easily run an Atrocious fuel rating and still be looking pretty decent on actual costs.

Hmm. Come to think of it, this may be a worry. We may have to compensate with extra locomotive weight for express. That should solve the problem, since we can tweak it to get whatever fuel bill looks right and loco/tender weight doesn't appear to have any effect on speed.

[RulerofRails]

We may have to compensate with extra locomotive weight for express. That should solve the problem, since we can tweak it to get whatever fuel bill looks right and loco/tender weight doesn't appear to have any effect on speed.

I was thinking the same thing. While the concept of paying fuel costs to run an empty engine isn't top notch, it makes sense that a passenger train will have to pay a decent amount of fuel costs with a certain consist even if nobody actually boards the train. In the game, we can't keep a constant consist whether it is running at 30%, 50%, 80% etc.. As a simulation though I think this concept is good.

Loco and tender weight does have some effect on speed. It's a lot less than the same weight of consist for sure. It is noticeable a lot sooner on graded speeds than flat ground. However, adjustments to free-weight will complete remove/adjust/invert this as neccesary. Because reliability is unaffected engine tender weight that is cancelled out by equivalent free weight is all about fuel cost for the engine itself. So far I haven't seen it doing anything else.

PS.
Other things you suggested sound good. I'm finally looking into the engine stats. Hopefully I can crack the formula so we can make our own calculations for custom consist weights.

[Gumboots]

Just thinking some more, by drastically cutting baseline express weights and bringing down maximum freight weights we've effectively solved any reliability problems. The bugbear now is going to be fuel consumption.

Say we take the Big Boy hauling late 20th century freight as an example of an upper limit, since everyone agrees it should become useful. That'll be a maximum consist weight of 384 tons for the heaviest freights if we go with the 48 ton limit I'm currently thinking of. We already know it burns far too much fuel with loco+tender .car weight of 620 tons and an 8 car consist .car weight of 320 tons, with fuel economy set to Average (x5 in the hex). So 940 tons in total at a fuel rating of 5 makes a locomotive utterly useless. Then you asked me if I'd ever found it useful in the C era when consists are only 160 tons or less, and I said no. Wolverine also found the same, which got him asking about the Big Boy on the Canyonlands map.

Ok, so we know that even with a total weight of 780 tons, a fuel rating of 5 will still result in a useless locomotive. If the fuel rating was dropped to 4 would it become useful? Or do we need to drop it to 3? Or do we need to also drop loco weight some more to keep all-up weight under (for example) 600 tons? We should start by nailing down what a useful Big Boy would look like, bearing in mind that it should still be a heavy grades and heavy freight specialist rather than a general purpose fallback.

That will give us one end of the scale. The other end is obviously going to be dedicated express. The more difference we have between the two the better for differentiation of locomotive classes, but we only have a limited range of ratings to play with. Do we want all express locos having Atrocious fuel economy in the 1990's? Strictly speaking it makes no difference to game strategy, but it could make people think something was wrong. So we also need to nail down what we think would be acceptable for late 20th century express locos. If we get decisions on both of these factors, that will tell us what range we need to fit everything else into.

[Gumboots]

While I think of it, there is some scope for renaming some ratings. We're stuck with acceleration, fuel economy and reliability sharing language strings, but it may be worth considering renaming them to use a more neutral-sounding scale. Just off the top of my head, if we used numerical ratings instead of words like Atrocious that may give a better result, just in terms of things looking more sensible without limiting what we use.

[RulerofRails]

The Fuel Cost Calculator sheet works well for determining fuel costs. At 600 tons a fuel rating of 5 is looking questionable (starts at 109k for 532 miles). Slightly less than the current unloaded Big Boy's fuel cost of 111k on that run. Setting 4 (starts at 87k) may work, but cost will rise to 100k by year 7.

My feeling is that fuel costs when new should be in 55-75k range for this period. The Big Boy should likely be on the upside of that range. So suppose we choose 75k. To get there would require 690 tons at setting 3 or 520 tons on setting 4. That's why I made the spreadsheet. Thanks to your refinements it's even better. Toss in a few values and play with it.

[Gumboots]

Ok, I see what you mean. Maybe we should assume a fuel rating of 4. At a rating of 5 there's hardly any weight left at all for the loco and tender, which seems a bit over the top. OTOH, I'm not sure we should give a Big Boy a rating of 3 for fuel. So 136 tons is almost exactly 1/4 the weight of the real thing, with tender fully loaded. I wonder if that'll work as a guideline for loco/tender weights.

Just ran a quick check on a Kriegslok. Actual weight of the thing IRL is 140 tons stocked water and coal. 1/4 of that is 35 tons, plus consist of 384 tons, equals 419 tons. Plug that into the sheet with a fuel rating of 4 and you get a year 1 cost of $61k. Seems to be in the ballpark for that example at least. It'd be good to have a fairly consistent guideline for the various categories of locos.

---

Edit: Just checked a few UK express locos. It appears that if we use 100% of the actual loco+tender+load weight for those, it will work with a Below Average fuel rating. Gives a fuel cost in the mid 50's on your sheet.

Doesn't work for the N&W J class though, which is far too big and heavy to use 100% of the actual weight. For something that big we'd need to use about half its actual weight if wanting it to be dedicated express. Maybe we will end up with a sliding scale for express.

Checked a default Adler too. Weight from .car files is 22 tons. Default A era 50/50 consist is 32 tons for 8 cars. Put 54 tons total into the sheet and fuel cost comes in at 20k for a rating of 10. So obviously the early locos need to have lower fuel costs to work.

If we go with the 1 ton A cars for express, and a 4 ton average freight, using 150% of the actual loco+tender weight would make an Adler have about the same fuel cost as it does by default at the same fuel rating of 10, when hauling an 8 car mixed consist.

Ok, so as a first stab at a guideline it seems that if we take a straight line going from y=1.6 at x=0 tons, and going down to y=0.4 at x=600 tons, this will give us a multiplier for the locomotive's actual weight, which when applied to get a figure for the .car file will work quite well for express locos. It works for the Adler and it works for the N&W J, and it works for average UK locos, given some tweaking of fuel rating. Obviously fuel rating can, if necessary, go past 10 for early locos. Needs more checking, but is a good start.

[RulerofRails]

Been doing some more digging into this.

I think I identified some formulas which are now in use in the Lifetime Engine Analysis spreadsheet. But more testing would be in order. If anyone wants to give it a stab. . . .

My tests involved a brand-new engine taking the breakdown chance with a specific consist weight, subtracting the component for the engine (checkout this graph for how they look), then dividing this by the tonnage in the consist. This number turned out to be about 1/150 of the engine component. It's not exact but it's fairly close.

B = Brand-new breakdown chance (varies only with reliability level)
C = Consist weight

So the breakdown chance the consist adds to a train of Oil level 100% (no caboose) can be approximated by
(B/150)*C or BC/150

For the entire train add this to the engine component (that's B)
B + (BC/150)

Increase due to a lower Oil level is a simple inverse multiplier
1 - O
(where O = Oil level)

To simplify application so after application we have (original + increase) we add 1 to give
2 - O

Combining with the weight component
(B + (BC/150)) * (2 - O)

The multiplier for age is represent by a linear function whenever Engine Age is 2 or more. (At Engine Age 1 value is approx. 2.059)

Where A = Displayed Engine Age
(A+30)/9.

Combining again we get:
(B + (BC/150)) * (2 - O) * (A+30)/9

If there is a Caboose, halve the result.

I assumed some rounding in the Brand-new breakdown figures so I tried to find more exact values by solving for the consist multiplier with high consist weights. I used those in the spreadsheet. But this simple formula should be right in the ballpark.

Here's the graph of the multiplier for age:

[Gumboots]

I just typed up a PM, but then thought of stuff to add, then thought hey why not just post it for the public record. So here goes.

---

A bit more information for you. I just ran a quick test. The longest run on the UK map is London to Kingston-upon-Hull. I usually run the Firefly with 6 cars + caboose, mixed traffic. This is using WP&P custom cars, but for A era you can assume the weights are the same as default, so that's a 31 ton consist on average.

So starting at London, fully serviced, it gets to K-u-H with 68% oil at the point where it starts unloading. AFAIK that's the minimum I've ever seen when running Fireflies, and gives an average of 84% oil level over the whole run. They're reliable used that way, as long as I replace them once the three year "grace period" is up (IOW, I usually run them for years 0, 1 and 2 only).

On some other maps, with some other locos, I would have run lower oil levels than that. London to K-u-H leaves 14% water available, and since consumption of oil and water is just linear and distance-based that would mean 0% water equates to just under 63% oil. So occasionally running locos out of water just for the last bit of a long trip, I'd say my locos usually never run less than 60% oil.

The "usually" is because I will sometimes, not often, run a few locos on inline facilities that aren't actually scheduled.

To keep the Firefly usable with the new scale I'm thinking it should haul 8 pax cars at 2 tons each, for a 16 ton consist, with no caboose, and without exceeding the default Firefly's breakdown chance with 31 tons and a caboose.

Edit: Just remembered something. Back here I measured actual changes in breakdown chance with different consists. This was back in the days when we thought the .cgo value was used, so consist weights in that post are 3x the .car value.

So the C era of 480 tons is an actual 160 tons, and as you can see it results in breakdown chance increasing by just over 100%. Which implies your value of 150 is pretty close. We just didn't notice it before.

Since adding 160 tons to a Firefly boosts breakdown chance (new) from 8.6 to 17.9, then a 31 ton consist should give a result of 10.4%. Then I was using a caboose with that, so cut it to 5.2%.

So that's what the new Firefly with the new scale has to match, with a 16 ton consist and no caboose. Which is obviously impossible at a rating of Very Poor. It would require an unloaded breakdown chance of 4.7% when new, which is intermediate between Average and Below Average.

The other alternative is to replace them faster, on a 2 year cycle instead of 3, which with a Very Poor rating would give the same result in terms of reliability (IOW, 3 year old hauling 31 tons with caboose is about the same as 2 year old hauling 16 tons without caboose).

Now all this is to match my preferred standard of reliability. Other people are obviously happy with less than that, since they aren't as fussy about servicing or cabooses. But, at the least, I think the Firefly should be boosted to Poor. Alternatively it could be made a bit cheaper, so that rapid turnover isn't as bad on the books.

The other thing is maybe the Planet should be dropped to Very Poor and left cheap to buy. As the earliest it should probably be the least reliable, and Very Poor would basically force replacement on a 3 year schedule. Actual real life turnover of locomotives in the very early days was that rapid. They became obsolete very quickly. So perhaps pricing them to make sense with a rapid turnover is the way to go.

[RulerofRails]

Well here's some stats of a Planet vs. Firefly comparison with the Lifetime spreadsheet. Important provisos, Firefly is technically pro-B-era loco and distance traveled is equal for both locos (399 miles). I'm yet to determine a rough flat-ground guide for miles traveled per year taking into account top speed and acceleration. With such a measurement we could then work out the Operating Cost per mile.

I think for me a guide of what's too much reliability-wise (even for the early experimental design engines) would be 40%+ Breakdown Chance at 60% Oil when run with Caboose. With the above comparison we can see that such a "rule" would limit the Firefly to it's optimum service life of 10 years of service in the default A-era (reality is that B-era arrives 7 years after it's release) and then I'm hitting this rule (44.4%) in the 3rd year of B-era.

When it comes to considering 20th century engines, my feeling is that the 40% rule is too generous. It feels wrong to get the same types of breakdown chance for a loco that was made in sufficient quantity that somebody took the time and effort to model it for RT3 (I get that in the early years there were no better alternatives available). Putting this limit at 30% would put this right in line with a 10 year lifespan on Reliability rating of #7 (Above Average) for 50-ton cars*, again at 60% Oil. That's a 350 ton weight (assuming a 0-ton Caboose/ETD).


*Will crunch more numbers later. Ex, is 60 tons per car for heavy freight too much for #7, even with a 0-ton Caboose? Comparing the last full weight Caboose vs. the introduction of the ETD. Also, it may be that the post-2000 (50+ ton average freight weight) corresponds to a current end game loco selection where there aren't many Above Average reliability ratings in use.

[Gumboots]

Default D caboose is 53 tons, and some of WP&P's freights in the same period are that weight. I know a Kriegslok (Above Average reliability) is ok with 7 cars at 50-53 tons each, if one is a caboose, and providing maintenance is scheduled.

7 cars at 60 tons each + 0 ton "caboose"? With inline facilities that aren't scheduled? I was always assuming that would require a rating of Good or better. I said so in the spreadsheet.

Consists of all heavy freights in the post-1990 period (60 ton cars) will require fuel and reliability ratings of Good, but I think this is reasonable for the late 20th century.

Edit: Hang on. Where are you getting 80 ton consists from in the pre-1850 era? 10+ ton freight cars don't come in until 1865. 7x heavy freight + caboose, pre-1865, is only 59.8 tons. Anyway, the Firefly is only intended to haul express, not heavy freight.

Edit #2: IMO anyone who attempted to run a Firefly for 9 or 10 years, with 80 ton consists, would be not exactly a rocket scientist. This is definitely a case where reliability concerns would beat fuel/maintenance concerns. Fireflies are bad enough after 3 years with 30 ton consists. 9 years and 80 ton consists would be insane.