Four updates rolled into one. I held them together because they turned out to be the same job: the simulation already knew a great deal, and the game was doing a bad job of telling you any of it.
I want to be straight about something first. A fair number of the entries below are fixes for things I got wrong, and most of those are here because a player sat down and wrote up exactly what had happened. That is not a small thing to do. Thank you.
The demand, pricing and cost models have not moved. The only real change to the simulation itself is in 0.19, and it is a bug fix.
Routes, then Feasibility study. You pick two airports, pay a consultancy somewhere between $150k and $300k depending on how big they are, and you get a report back. Passengers a day each way, how many of them would be connecting over the rest of your network and to where, and what to fly it with: how often, how many seats, which of your aircraft fit, roughly what times the departures should leave and why, and how many aeroplanes that ties up.
Each study also collects what the places at either end think about it. Residents who would use it every month, a chamber of commerce that has turned down work it could not reach, someone pointing out that you could drive it in a morning, and at a small airfield under a big aeroplane, the objection this genre is named after. None of it changes how the route performs. It is there because it is the same set of facts the numbers come from, said out loud.
A study is a model rather than the world, and it is deliberately a little wrong, more so on connecting traffic, which depends on a timetable that does not exist yet. That is why it gives you a range as well as a single number. The same route always studies the same way, so if the flying comes in under the study, that is something you can learn from rather than noise.
It will also report on a route you currently cannot fly. If nothing in your fleet has the range, it still tells you what the market is worth and says plainly that the aeroplane is the problem.
Rather than dots. Runway, elevation, passengers a year, the weather and what it is doing to operations, what you fly from there, which rivals are based there, your slot position hour by hour, and everywhere anyone flies to from it. If a city is missing from that last list, nobody serves it.
The part I am most pleased with is fees and charges. Landing and ground handling are billed on arriving flights, so a big hub costs you more to serve inbound, and until now the only way to discover that was to fly the route and read your ledger a week later. Every airport now shows its rates, what your service posture does to them, and a calculator: pick one of your aircraft and a load factor and see what a landing there actually costs, per landing, per passenger, and across a week of your arrivals.
One caveat, because the comparison table on that page can mislead. Handling is a flat rate per head, so it cancels out, and what the ranking really measures is weight per seat. That is why a turboprop can come out on top. Airport charges are not the whole cost of an aeroplane. Fuel, crew and maintenance are what pay for a bigger airframe, and Route P&L is still where a route's real economics live.
There are also arrivals and departures boards at every airport, for the next 12 hours, the next 24, or the whole week. Your own flights carry their real status once they have operated. It is the quickest way to read a market: six departures to Chicago, mostly A320s, and nothing at all to Denver.
This was a real bug and it had been there a long time. If a flight was cancelled, the aeroplane's next service still departed from an airport it had never reached. It does not any more. A cancellation strands the aircraft, and the next flight either gets a positioning leg or does not operate. Cancelling the outbound of an out and back takes the return with it, which is what happens in reality.
A positioning flight is a real flight. Fuel, crew, maintenance and landing fees, nobody on board, and it eats into the day. They are listed with their cost under Network, then Assignment.
Which of the two you get is your call rather than mine. On the Airline page, Delay before cancellation sets how much accumulated delay an aeroplane can carry before the flight is cancelled and it starts clean, anywhere from 15 minutes to never. A tolerant airline flies the empty leg. A strict one cancels and leaves the aircraft where it is. Schedule flexibility sets how far past its published time a departure can be moved to find an aeroplane, and Exact means it flies at the minute you chose or is reported as not flown.
Over a simulated year on a tolerant airline this came out at 81 empty legs, with the cancellation rate unchanged at about 3%. It changes the shape of a disruption rather than how often one happens.
A player running a twelve a day Chicago to New York shuttle with eight 757s reported that their Sundays were not flying. They were right, and the way it failed was worse than the missing flights.
Twelve daily departures is 84 round trips a week. There was a ceiling of 70 per route, so it kept the first 70 in clock order, flew twelve Monday to Friday, ten on Saturday, and nothing at all on Sunday. And because I counted the request after applying the ceiling, the page told them it had flown 84 out of 84. A number on screen that is simply untrue is the worst kind of bug in a game like this, and I am sorry it was there. The ceiling is now well above anything the editor will let you ask for, and if you ever reach it the shortfall is reported properly.
The same report turned up something else. Departures were wandering. The assigner used to pick whichever time in its flexibility window added least to the airport's traffic, so 05:00, 05:40, 06:20 and 07:00 came back as 05:00, 06:00, 07:00 and 07:20, at an airport with slots to spare and eight idle aeroplanes. A departure time you chose is a decision, not a suggestion. It now flies at your minute unless that hour is genuinely full or nothing is free, and when it does have to move one, Assignment tells you where it went and which of the two reasons applied.
It is still a greedy first fit underneath. It places one departure at a time and reports the ones it cannot place. It does not re-solve your network, and I have not tried to hide that. What has changed is that you can see the squeeze while you are planning: how many block hours your fleet is flying against what it could, how many aeroplanes are idle, and in the plan editor, how many aircraft a plan needs at a given base against how many are actually there. When it does not fit, the other routes competing for those aeroplanes are named, instead of the shortfall being blamed on whatever route you happened to have open.
Two smaller ones: up to 24 departures a day on a route rather than 12, and you can base aircraft at any airport you hold a licence to rather than only at a hub.
0.20 was me going through every screen from the main menu to the weekly report, writing down everything that was unclear. It came to 97 things. None of them were about the model underneath, which was humbling in its own way.
The ones worth naming:
The Finances page was understating a week's loss by about 97%. It reported an operating result of around $2k for a week that actually cost $68,797. Ownership, leases and station overheads are charged once at a week boundary, so a rolling seven day window either contained one or did not, depending on which day you happened to look. There are whole week windows now, and a block showing cash at the start, financing in and out, the result and cash at the end, so you can check the screen against your own balance instead of taking it on trust.
The weekly report was also telling you that buying an aeroplane was a loss. Week one read "net this week, minus $56.7M" next to "cash on hand, minus $50k this week". Capital spending is reported separately as Invested in growth now, and the two figures agree. The report used to arrive a day late as well, and it disagreed with the Cancellations page about how many flights had been cancelled.
Your routes are drawn in your livery colour, your bases are badged, and a new game opens looking at your airline instead of at the whole planet. There is a week view on the Schedule page. New Game has a map, so you can pick a hub by clicking it. Aircraft get registrations from your airline's country.
Settings exist in a meaningful way now: units, a 12 or 24 hour clock, a spend confirmation threshold including never, colour blind safe colours, reduced motion, page text size and a keyboard reference. A third of the interface was failing minimum contrast, which I only found by measuring it. And the tutorial used to promise you would set a departure time and a fare, then fill both in for you, which meant it taught two of the five decisions by skipping them.
Most of what is above came from people posting on the discussion boards. The assignment bug, the airport pages, the studies. The in game Roadmap shows what has shipped and what I am looking at, and it is honest that some of the latter will not happen.
This is still Early Access. Systems and balance are still moving, and a save may not survive every update.