46 replies to this topic

[AYFlier]

I am currently playing in AE4, and I think that the pax numbers don't add up.

There are airports in there that are having more seats than their real numbers by the order of a magnitude. Ie most European cities seem to have direct flights to places like Honolulu, Dallas Love field etc when in reality this does not happen. If you multiply they daily pax of an airport by 365, it is very often 10 times the real life numbers.

Would it make more sense to implement the demand as it exists in real life ?

Ie there is a certain number of people, say 10 000, than need to travel between London and Tokyo each day. Then when you set up the flights, if it's LHR-NRT it is from this same pool, and if you set up LGW-HND it is also from this same pool.

And another twist, if you have a hub in say ARN and you have flights LHR-ARN and ARN-NRT you would be able to tap into that same pool of 10 000 pax. Ie you could also set up routes that are based on existing point to point connections and price them separately. Of course if a direct flight is priced the same as a connecting option, it would be more viable. But if you priced LHR-ARN-NRT at 70%-80% to that of LHR-NRT direct, you would be able to attract customers from that pool of 10 000 pax.

[BritAbroad]

The reason for this is the way that AE calculates route demand. When we add an airport, we input a single figure for the total number of passengers travelling through the real world airport each year (rather than the number of people flying on each individual route). The game then automatically calculates demand between that airport and every other in the game, based upon a single formula. This takes into account the number of real world passengers at the origin airport and at the destination airport, the distance apart and a number of other factors.
Naturally, this leads to discrepancies - in AE, there's no demand for high frequency island-hopping flights, yet you can fill a flight from London to outer Siberia - but to put it simply, its the easiest way of doing things.
The demand script can and will be tinkered with, but this is how we calculate things for now.

I highly doubt that data exists for the numbers of people wishing to travel between a given city and every other city in the world, and even if it did, putting it into AE would be a mammoth task. This is why route data is calculated by passengers at a single airport and a single mathematical equation.



However, I do hear what you are saying, and you have a valid point. I'd like to see "city pools" - for instance, an airline at LHR competes with one at LCY, etc.
How that is going to be calculated is yet to be considered.

[AYFlier]

But what about the fact, that the current formula produces unrealistic amounts of traffic ? There are AirPorts in AE that have ten times more traffic than they do in real life.

Would it be plausible to introduce some general guidelines into the formula ie. There is certain amount of traffic between europe and US. Certain amount between US and Australia etc.

[BritAbroad]

But what about the fact, that the current formula produces unrealistic amounts of traffic ? There are AirPorts in AE that have ten times more traffic than they do in real life.

Would it be plausible to introduce some general guidelines into the formula ie. There is certain amount of traffic between europe and US. Certain amount between US and Australia etc.

As I said, there's a single calculation, so it produces average results. For example, lets say that real life airport X has huge demand to airport Y, but next to nothing to airport A, B, C.... etc. When we add airport X, we will input the total number of passengers (including all those flying to Y, A, B, C.... etc), with the result that the AE routes will have less than real-life demand on route X to Y, but higher that real-life on route X to A.

In real life, there is huge demand on LHR-JFK (BA has sometihng like 10 daily 744 flights), but not much demand on LHR-URC, if any. Because of the way that AE works things out, it kind of levels things out. LHR-JFK in AE can barely support 3x 744s, whereas LHR-URC can probably support 1x 744.
So when you say there are airports with ten times more traffic than in real life, it may be only on certain routes. *


There are modifiers involved (for example, our database links certain countries to introduce greater demand between them), but it takes a lot of work to make it work, and it will never be perfectly realistic.


I don't mean to completely talk down your suggestion, but I just want to explain why things are the way they are. Of course, everything is in development and what you are suggesting is something that is known and will be considered. The coding that Yuxi has to do to create the demand script is pretty tricky, I understand, so it will take time to perfect.


* [Unless I'm misunderstanding you, and you mean that the passenger number we have put into the in the game is incorrect. If you look at the airport detail page (for example: http://ae4.airline-e...fo.php?city=LCY ) at the top under the airport name, the "Passengers per year" field should show the real-world passenger number for that airport. Sometimes data here can be out of date or incorrect and this will effect routes. I can, however, easily correct that if pointed towards the incorrect airports]

[AYFlier]

Hi,

By ten times I mean the following: take Tallinna for example (in ae4). The AirPort has 100 000+ pax per day in the game, but a mere 3million a year in real life. 365 * 100000 = 36.5 million pax per year from Tallinn in the game. Completely unrealistic, so what I'm saying is that there is something wrong in the formula because it allows that.

In reality TLL has a few destinations within Europe + a couple of charters. In the game it is a major hub with flights to all corners of the world.

[BritAbroad]

I see what you are saying now. I've taken a look at a selection of routes from TLL in AE4 and it seems overserved - the airline based there is flying routes with about up to twice the number of seats available than there is demand. This means that the airline is actually either not being very efficient (lower frequencies would yield more profit), or he has a huge amount of connecting traffic. He may be doing as Emirates did in the real world and effectively turned TLL into a new DXB - I would reckon more than half of real world Emirates passengers are passing through and not actually going to Dubai.

Also with this example, I would suggest that we see a result of the "averaging out" that I mentioned earlier. I imagine that Tallinn has a lot of real world flights to Eastern Europe, Scandinavia and the other Baltic states. Now in AE, the demand for these routes will probably be lower than in real life, but the demand on routes to south america for example will be higher as a result. I suppose that this is necessary to give players a bit more freedom, rather than restricting us to real life routes.


Again, I appreciate and agree with what you are saying, but it is simply the only way we can do things.

However, if you are able to find data for where everybody in each city of the world would like to go to (and note that will not be the same as real world routes), we would then use that to create a representation of real world route demands. However, I highly doubt that such data exists for even the major cities, let alone every single one in the world.

[AYFlier]

I was thinking about this last night, and I came up with a following idea:

Typically, if an airport has a low number of annual pax (say <= 1 mil) it means that there are maybe 2-3 routes that it serves. Real life example would be OUL, like 90 % of the traffic there is OUL-HEL. The explanation for this makes sense. It is a small/regional airport that does needs a service, but only to a hub as it is not able to support direct flights all over the place. Now we could tweak the formula in a way that when the world starts and there is a void, whoever is the first one to tap into such an airport would be likely to get good demand to where ever he is flying according to the rules in place now. However once this 1 mil a year (+ some extra for making the game a bit more interesting)is used up, there is no demand left anywhere. So if the first airline to start service to OUL would flood the airport with 1 mil seats a year, the next one would find no demand left. Unless of course he starts to compete with price, which is easiest if you start flights to a closer hence cheaper hub. So as an example if the first one to setup flights to OUL would be CPH-OUL and he takes up all the demand, this would still leave a possibility that someone else starts a route OUL-HEL (shorter hence cheaper tickets) he would be able to replace the original CPH-OUL.

Then if the airport is a medium sized airport (say <=20 mil a year) there would probably be something like 100 - 300 destinations served. Again once the annual number of pax is (again, + some extra) used up there is no more demand left after that.

So basically what I'm saying, is one addition to the formula as it is now. When calculating demand, it would check if the overall number of pax/year is already hitting the roof either at destination or at origin (which is related real numbers + some game extra) there would be no demand for the route.

This would then mean, that you cannot fill 20 B747s from JFK to TLL in the game, as TLL would get saturated. But there would still be outgoing demand left at JFK that would need to find a destination that also has similar demand left.

Am I making any sense ;-)

Also one way to limit overrealistic traffic would be factoring the available runways into the formula. Ie. TLL has one ~10 000 ft runway, that can only accomodate so many flights. Assuming a runway can have round the clock an operation every 2 minutes (leave room for maintenance etc may also be necessary) it would mean 15 departures and 15 arrivals an hour, ie. max 360 departures a day. Though that number is not exceeded by the TLL example, yet ;-)

[Adaptice]

I have no problem with this present setting

[Sheepy]

I was thinking about this last night, and I came up with a following idea:

Typically, if an airport has a low number of annual pax (say <= 1 mil) it means that there are maybe 2-3 routes that it serves. Real life example would be OUL, like 90 % of the traffic there is OUL-HEL. The explanation for this makes sense. It is a small/regional airport that does needs a service, but only to a hub as it is not able to support direct flights all over the place. Now we could tweak the formula in a way that when the world starts and there is a void, whoever is the first one to tap into such an airport would be likely to get good demand to where ever he is flying according to the rules in place now. However once this 1 mil a year (+ some extra for making the game a bit more interesting)is used up, there is no demand left anywhere. So if the first airline to start service to OUL would flood the airport with 1 mil seats a year, the next one would find no demand left. Unless of course he starts to compete with price, which is easiest if you start flights to a closer hence cheaper hub. So as an example if the first one to setup flights to OUL would be CPH-OUL and he takes up all the demand, this would still leave a possibility that someone else starts a route OUL-HEL (shorter hence cheaper tickets) he would be able to replace the original CPH-OUL.

Then if the airport is a medium sized airport (say <=20 mil a year) there would probably be something like 100 - 300 destinations served. Again once the annual number of pax is (again, + some extra) used up there is no more demand left after that.

So basically what I'm saying, is one addition to the formula as it is now. When calculating demand, it would check if the overall number of pax/year is already hitting the roof either at destination or at origin (which is related real numbers + some game extra) there would be no demand for the route.

This would then mean, that you cannot fill 20 B747s from JFK to TLL in the game, as TLL would get saturated. But there would still be outgoing demand left at JFK that would need to find a destination that also has similar demand left.

Am I making any sense ;-)

Also one way to limit overrealistic traffic would be factoring the available runways into the formula. Ie. TLL has one ~10 000 ft runway, that can only accomodate so many flights. Assuming a runway can have round the clock an operation every 2 minutes (leave room for maintenance etc may also be necessary) it would mean 15 departures and 15 arrivals an hour, ie. max 360 departures a day. Though that number is not exceeded by the TLL example, yet ;-)

Although I see what you're getting at, I would oppose any limits on growth. I would further disagree with using pax numbers as a way of dividing airports into 'small' 'medium' and 'large'

[AYFlier]

Well couldn't the system cater for both by setting up separate worlds ? In reality there are economic laws that do limit growth, however I do realize that some people like it the way it is now. I just feel that the game would be more challenging if you couldn't just grow exponentially. Airline industry is not exactly a gold mine in the real world. Easiest way to become a millionare in airline industry is to start of as a billionare

[BritAbroad]

The demand script is by no means perfect, but we do have to compromise between the fun of the game and the element of realism.

I understand your idea, AYFlier, and it is a good one, however, I imagine the coding for that would be hideous. It would also not be a good idea to have two different demand scripts.


AE is constantly under development though, so your ideas may be taken on board later down the line.

Regarding runway slots - I'm all in favour of this, but lets say TLL became the new DXB in the real world - the chances are they'd build another runway. I think if you are to bring in restrictions like this, there should be the option for expansion.

[un1]

Regarding runway slots - I'm all in favour of this, but lets say TLL became the new DXB in the real world -

I'd say it already has.

Anyways, what Brit said pretty much is the dev's view on this. Though I think runway slots are going to be implemented with timetabling (whenever that comes).

[AYFlier]

Regarding runway slots - I'm all in favour of this, but lets say TLL became the new DXB in the real world - the chances are they'd build another runway. I think if you are to bring in restrictions like this, there should be the option for expansion.

That too would be really hard to simulate, as we would need to find some kind of balance between realism and game fun. Where would you build another runway in San Diego ? (TLL is not really that different in that respect, the airport is located practically just next to the city itself). Even places like LHR where there is room and demand are facing NIMBY issues about runway expansions.

[polarscribe]

To me, the real issue is that there are just too many airlines allowed per world. It really should be reduced down to 100 or so, maybe even less. I completely lost interest in Airline Mogul because of the ridiculous hub-camping that made it impossible to do any sort of realistic airline model.

The simulation aspect utterly breaks down at the point when every airline tries to build a hub in every city with 20x MD-87s flying between every random city on the map "just because."

If there were fewer airlines chasing pax, the demand-side issue would solve itself.

[AYFlier]

The simulation aspect utterly breaks down at the point when every airline tries to build a hub in every city with 20x MD-87s flying between every random city on the map "just because."

Maybe the easiest way would be to limit the availability of new airframes ?

In reality, when you place an order for a new frame you won't get it in two weeks.

What if we started a way that when a world starts, there is a limited availability of used B747s,B737-400s, MD80s, MD11s, B727s etc on the market (so that you could actually start the airlines). But the availability of new airframes would be based on manufacturer capacity (you would get a new frame several months from the order at a minimum). Airbus & Boeing can each produce about 500 frames a year, so that alone would control the exponential growth here.

Unlimited supply of new frames is a major unrealistic feature in AE.

[polarscribe]

That's a good idea that I think would help slow down the carnage, but at the same time, it runs into a problem of too many players allowed per world. With 500 airlines per world, every single manufacturer is going to have their orderbooks filled to capacity long before the world comes close to filling up. Anyone who tries to enter even slightly late will be hopelessly screwed - they'll be unable to buy *anything* from the new market for years to come.

The fun part of these games is building up a route structure, developing hubs, buying planes, etc. Not going through routes endlessly clicking and changing prices by $1 to fix your economics. That's just FarmVille grind crap.

[AYFlier]

That's a good idea that I think would help slow down the carnage, but at the same time, it runs into a problem of too many players allowed per world. With 500 airlines per world, every single manufacturer is going to have their orderbooks filled to capacity long before the world comes close to filling up. Anyone who tries to enter even slightly late will be hopelessly screwed - they'll be unable to buy *anything* from the new market for years to come.

The easy way to solve that problem, is to run several games in parallel.

ie you would have AE4.1, then AE4.2 starts when 4.1 is filled with a predetermined amount of airlines. When that happens to AE4.2 you start 4.3 etc.

Not knowing the system architecture of course, that might bring some capacity issues though ...

[bolli]

The easy way to solve that problem, is to run several games in parallel.

ie you would have AE4.1, then AE4.2 starts when 4.1 is filled with a predetermined amount of airlines. When that happens to AE4.2 you start 4.3 etc.

Not knowing the system architecture of course, that might bring some capacity issues though ...

Like we already do?

[Moldova96]

I see the issues surrounding the demand but there are serious issues trying to get all the data. What Im trying to do is get the gate numbers to the correct figures, this limits the amount of airlines at one airport and helps stop airlines becoming huge from and airport un-capable of doing so in real life.

[polarscribe]

Like we already do?

The limits are way too high. Should be 100 airlines per world.