On Tempelhof, my airline, Central Air do Brasil, flies two longhaul flights with a 787 - Belo Horizonte International (CNF) to Washington Dulles (IAD) and CNF to Johannesburg (JNB). Both IAD-CNF and JNB-CNF are completely full and carry 100-220 connecting passengers through my banked CNF hub. On the other hand, CNF-IAD and CNF-JNB carry no connecting traffic at all despite fitting into the same wave as the arriving flights, leaving plenty of time to connect in excess of MCT and showing up at the top the ORS results. As a result, yields on JNB-CNF are terrible (luckily, there’s enough local IAD-CNF traffic to fill the plane).
How can a plane be completely full in one direction but only 10% full the other way? And why aren’t passengers connecting to my CNF-departing flights?
It could be that the demand is only one way, means that the demand, through connection or not, from CNF-JNB the demand may be high, but the return flight is a totally different story, you can interline with more local operators, it’ll cost you some money, but the possibility of getting that plane fill is much higher, and you don’t have to send more time building more connections to gamble on where they’'ll want to fly to.
Because demand is asymmetrical in time, and passengers and units of cargo have specific destinations; all the connections in the world mean nothing if none of them can get them where they want to go.
Think of this situation; I’ll draw it out because for some reason this is a concept that people never seem able to grasp:
100 passengers want to go from X to Z; 100 passengers want to go from Z to X. They can only do this by going through Y.
100 passengers want to go from A to Z; 100 passengers want to go from Z to A. They can do this either by going through Y or by going through B.
In this metaphor your flights to Joburg and Washington are the X to Y section.
Direction #1:
No problems here - the passengers going from A to Z prefer to go through B (for whatever reason; time, price, service etc) so everyone who wants to get to Z can get a seat on the X-Y flight, which is consequently full.
Direction #2:
Problems. The connection from Z to A through B is mostly booked up - perhaps the Z-B section has been filled up with connections from other airports already. This is where the time assymetry comes in - airports calculate their demand at different times of day, so it’s possible for one direction of a route to fill up at some intermediate point before the reverse can happen. Ok - all the remaining passengers trying to get from Z to A have to go through Y, taking 90% of the seats in the process. Where does that leave the Z-X traffic? There are still 100 passengers trying to make the trip - but only 10 seats between Z and Y, even though there are still 100 between Y and X. End result - low load on that sector.
This is a very small example, but in aggregate it’s how the entire system works, and it’s illustrative of why it’s to your advantage to provide alternative routings - a single flight to a single destination is fine, but it’s difficult to cover all contingencies. I imagine your difficulties in Washington stem from the practical impossibility of getting passengers onto any of the small number of flights to New York, Chicago, Detroit, Boston etc - they’re all likely to be full when your passengers try to make the trip, but for the return they’ll trickle in quite happily as the external airports calculate their traffic. Joburg I haven’t checked, but a likely possibility is that passengers from a key connection can get onto your JNB-CNF flight but not vice-versa due to timing issues.
If your plane to Johannesburg is full with connecting passengers, try to figure out where they come from. Adding an extra flight to Rio, Brasilia or Sao Paolo when your flight from Johannesburg comes back might do the trick.
If you have 100 connections to different countries… forget about guessing
My problem isn’t in Washington or Johannesburg. It’s in my hub at Belo Horizonte. For whatever reason, people don’t want to connect through CNF to get to Washington or Johannesburg but are willing to connect through CNF when going from those cities. In other words, when demand is calculated for Rio de Janiero, virtually nobody wants to connect at CNF to go to Washington or Johannesburg despite showing up as one of the first options on the ORS. But when Washington or Johannesburg’s demand (or some other US or African city covered by one of my interline partners) is calculated, passengers will happily connect to Rio. So at least part of the bottleneck has to be at CNF.
Are connections to long-haul destinations calculated last? Perhaps the XXX-CNF planes are getting full with non-stop and other connecting traffic before any long-haul passengers could get on the long-haul flights. I think the Meigs update increased the percentage of nonstop traffic - at minimum, I know my connecting percentage is lower after the update.
There should be no timing issues with my flights. They arrive and depart in very defined waves which fully take into account CNF’s 0:45 MCT.
Edited to add: CNF-IAD seems to be picking up connecting passengers now. CNF-JNB is still picking up just 1 connecting passenger per flight. JNB-CNF averages 200+ connectors.
please forgive my poor brain. My previous reply explained what to do if your incoming flight from Joburg and Washington are empty
But the principle remains the same. Your flight from Johannesburg is filled with passengers who take a connecting flight in your hub. If these passengers take an internal connection, they perhaps take a flight to one of the bigger airports in your domestic/regional network. At least, I don’t think many South Africans travel to some small airport in the Amazone jungle
If your outbound flight to JNB is empty, it probably means passengers from airport this/these bigger airports cannot fly to your hub to take the connecting flight to JNB. Increasing the capacity from Rio, Sao Paolo or Brasilia to your hub before your flight to JNB leaves, may do the trick.
I don’t know in what order demand is checked. I would think every airport has a (alphabetical?) list of destinations, and the program just goes down the list. If the list is indeed alphabetical, your incoming flights may fill up with passengers for Belo Horizonte before passengers for Johannesburg and Washington check for the best way to reach their destination. But also in this case it means you don’t offer enough capacity from this airport to your hub.
Or passengers from other airports may fly to Rio and take a connecting flight to Belo Horizonte. If these passengers fill up your incoming plane before Rio calculates it’s demand, passengers from Rio won’t be able to fly to JNB via your hub.
I would add flights on your busiest routes. Those that fill up 3 days ahead…
Thanks for the analysis, I believe this does go a long way explain the effect, unfortunately it underlines the fact that the way in which Airlinesim works is unrealistic - in the real world you simply do not get 100 pax all booking one-way flights to a single destination at the same instant.
* There is not a single connecting passenger on any of the flights (this has been checked for a 3-day period, both ways)
* My prices, services etc. and hence ORS ratings are identical for both ways.
* Competition's prices and ORS ratings is also symmetrical (I am first in the ORS anyways).
How?
Edit: on a different route in the exact same situation, checking out C - D in the ORS shows most of the flights of the operators (about 5 airlines or so) fully booked, while for the D - C route it shows most of the flights bookable (for everyone)!
Airport demand? What about it? If this is what you wanted to say, demand is supposedly symmetrical both ways, i.e. demand (A - B) = demand (B - A)
Network supply? Ground network connections is the only one that could provide some explanation, although not to this extent, and hardly to believe for all airlines on the route.
Interlining supply?I clearly mentioned ZERO connecting passengers.
Competition on this HUB? I clearly mentioned ZERO connecting passengers.
Now to think of it, ground connection pax can be the only explanation.
Assume the set up:
C....A----------B
Where A - B is the route in question and C....A is ground connected.
Assume there's also a direct flight C - B.
When demand is calculated at C, flight C - B is already fully booked by connections, hence the pax on this route must go via ground to A and then fly to B.
For the return flight, when demand at B is calculated, flight B - C is not already booked and they can use it. Thus, flight B - A is much poorer in terms of load.
Demand is not symmetrical. Even if you have the exact same ratings between the two airports. There are some one-way routes that have more demand than their counterparts (A -> B having more demand than B -> A). There are real world datapoints that AS uses in determining demand on certain routes.
Demand is not symmetrical. Even if you have the exact same ratings between the two airports. There are some one-way routes that have more demand than their counterparts (A -> B having more demand than B -> A). There are real world datapoints that AS uses in determining demand on certain routes.
I doubt that there are scheduled civil air transportation routes where dem(A - B) = 100% and dem(B-A) = 50-60% (as I mentioned in my first post), so that's why I think this argument is out of place here.
Yes to have asymmetric demand would be highly unrealistic. Usually, passengers book a roundtrip and even if they don't they will go back someway or an other (on an other airline, on a not yet known date etc.). Ofcourse there are some people moving from one country to an other, but usually this is also symmetric and even if its not if would be a very low percentage(less than 1%, propably even less than 0,5%) of all passengers. I think morefocus explanation is the best.
:) The tricky part is to figure out the airport in the ground connection that is generating this effect (which is not too hard, considering the expected positioning of the demand calculation time... alternatively you could of course refresh aicraft or flight page around the demand calculation times of ground-connected airports, but that might be too hardcore) and then intervene in the corresponding route...
Why are you assuming it is a ground connection that is causing the difference?
When you are first in the ORS you don't get a 100% share of the passengers until your flight is full (unless you are the only flight available and your proposition is good enough). The other flights will be getting a distribution of the passengers too - exactly how much is not disclosed in the game.
You say there are other airlines competing in the route. Now you have to factor in the impact of those flights. If the outward leg of your competitor has capacity then it will likely gain some direct bookings. If their return leg is already full, then of course it cannot. That would increase the distribution to the other flights including yours.
So whilst interlining is not affecting your bookings directly, the interlining of your competitors' flights may well be indirectly affecting your bookings.
If you want to test this accurately, then try a route where there is no competition at all (ie no results in the ORS). That would make it much easier to study.
Why are you assuming it is a ground connection that is causing the difference?
From my OP: Edit: on a different route in the exact same situation, checking out C - D in the ORS shows most of the flights of the operators (about 5 airlines or so) fully booked, while for the D - C route it shows most of the flights bookable (for everyone)!
Thanks I read that before answering, I was offering you a different explanation to show there are more factors than just the ground network.
We don't know the connecting flight situation for competing flights, and there are so many combinations possible it would be ridiculous to guess. The competitors could have their flights filled by connecting flights on the outbound but not the return. If your flight (C-D) were the only one with availability when the demand calculation was made then you would receive 100% of the passenger allocation. If on the return (D-C) there were 6 flights with availability, then you would not receive 100%. There would be a distribution across all of the flights and it could be that none of them fly full.
Jan has already mentioned that demand is symmetrical in the game. You can prove this by using a clean example with no competition. Asymmetry happens as soon as we create different schedules and route possibilities. I agree it looks odd at times, and it can be one of the more challenging parts of the game.
At the beginning of Aspern I had a route (Lagos, Nigeria to Bamako, Mali) where flights were usually completely empty on the way to Bamako while the return flights had a 100% load, with no connecting passengers whatsoever. At this time Bamako had only two daily flights, the other being to Casablanca (i.e. opposite direction), from where there was no direct flight to Lagos anyway. Bamako also has a very modest ground network, i.e. none. I can pretty much rule out competition as well as problems with connecting passengers. So much for Demand being symmetrical.
I also have some routes where I suspect that the demand in Business class is not symmetrical. As far as I know the percentage of business pax flying from larger airports (e.g. 10 bars) is higher than from medium airports (e.g. 6 bars). But is the size of the destination airport also considered? Currently there are much more business passengers flying from LHR to LOS than the other way round. True, LHR does have a fairly large ground network and a tad more competition than Bamako, but still...
I also have some routes where I suspect that the demand in Business class is not symmetrical. As far as I know the percentage of business pax flying from larger airports (e.g. 10 bars) is higher than from medium airports (e.g. 6 bars). But is the size of the destination airport also considered? Currently there are much more business passengers flying from LHR to LOS than the other way round.
So what that there are more business pax starting their journey from LHR to LOS (then the other way around)?
You realize that they do come back to London, right? Otherwise in short time Lagos would become the financial capital of the world...
